Sparks and White Noise

So Where is Octave?

2011-11-09T23:32:00.001-06:00

I know my blog has been very quiet since April and people have been wondering, just where is Octave? Well I'm working at Oracle as a Director of Enterprise Architecture! Without a doubt this has been an amazing year for me and I'm very excited about this position at Oracle. At Oracle OpenWorld 2011, I presented at the session "Getting the Most out of Oracle VM Server for SPARC", where I talked about the new SPARC T4 servers, what's new with LDoms, and how SPARC can be used in your cloud deployments. There are soo many exciting things going on at Oracle around cloud computing, Solaris, SPARC, virtualization, etc.

There is soo much I want to talk about and I will do my best to bring new content to my blog. If there are topics you'd like to see me make posts on, leave a comment below!

Fujitsu Launches PRIMEHPC FX10 Super Computer

2011-11-09T23:24:00.002-06:00

As if Fujitsu didn't surprise the industry enough with the K Computer being the #1 super computer on the Top500 list, they have announced an even faster platform, the PRIMEHPC FX-10!

When I was at Oracle OpenWorld 2011, I had a chance to take a look at the SPARC64 VIIIfx 2.0GHz nodes that are used in the K Computer:

The really interesting thing about this server is the water cooling and of course the performance. Sadly, Fujitsu is using Linux on this server and the PRIMEHPC FX10 super computer. The PRIMEHPC FX10 super computer uses a derivative processor called the SPARC64 IXfx which runs at 1.848GHz and has 16 cores that provides 236.5 GFLOPs of performance. A fully loaded PRIMEHPC FX10 will provide 23,248 TFLOPs with 1024 nodes!

All in all, very impressive to SPARC at the top after all these years! With what Oracle has done with the SPARC T4 processor and what Fujitsu is doing on the high-end, there is a renaissance for the SPARC architecture.

Solaris 11 Released!

2011-11-09T23:01:00.000-06:00

Finally after all of these years in the works, Solaris 11 has been released! Check out the below links:

I've been waiting for this moment for a long time and I'm really happy that it has been released finally. The announcement today was full of lots of information that many in the OpenSolaris space already knew would come. However, there are many pleasant surprises for everyone. Here's a quick summary from the "What's New" document:

Better Auto-Installer
More configuration files migrated into SMF (/etc/nsswitch.conf, /etc/resolv.conf, /etc/nodename, /etc/defaultdomain, /etc/default/init, etc.)
New sysconfig command replaces the old sys-unconfig and sysidtool commands
Integrated p2v and v2v tools for migrating Solaris 10 machines to Solaris 11 Zones
NFS Server in a Zone!!
Crossbow and Zones Integrated!
RBAC based administrative controls for Zone administration
Enhanced ZFS support for Zone Boot Environments
Immutable Zones replaces Sparse Root Zones
Clean Shutdowns for Zones
zonestat tool and libraries for monitoring Zones
Root account is just a role now as it should be!
TPM Support in Trusted Extensions
Labeled IPSec reduces the number of physical network links and maintains security
More security and cryptographic enhancements to make Solaris 11 the most secure OS
X.509 certificate support for Solaris SSH
Tons of features around Crossbow
netadm and netcfg control SMF services for network configuration
Default vanity naming for NIC interfaces
dladm can make MAC address changes persistent across reboots!
Infiniband enhancements for TCP/IP to support SD and RDMA for performance
VLAN IDs are now broadcasted over VNICs to enable dynamic network reconfiguration
llpdadm command enables control of LLPD support on networks
New Socket architecture replaces the old STREAMS framework
Load balancer, bridging, and tunneling are finally included!
Link protection for VMs to prevent bad things from happening on your networks!
dlstat give you link observation stats. Plus netcat and wireshark is included for indepth analysis.
IPMP enhancements with ipadm, simpler configuration, better probing options, and SMF service
ProFTPd replaces the old WU-ftpd
tcp, udp, and ip providers for Dtrace!
ZFS Encryption included, but not for the root pool :(
ZFS deduplication included!
ZFS shadow datasets for migrating data
ZFS backups with NDMP with support for snapshots and clones!
ZFS datasets can now be mounted temporarily in alternate locations
ZFS sends can now be recursive to descendents!
ZFS diffs for showing what has changed between snapshots
NFSv4 support for alerting clients of when a file system has been migrated!
Built-int CIFS server replaces Samba, FINALLY! Integrated with ZFS for ACL support for Windows
DTrace providers for SMB and iSCSI protocols
COMSTAR framework integrated to provide FC, FCoE, iSCSI, iSER, and Infiniband SRP protocol support. Enables you to turn your Solaris server into a storage array!
DTrace probes for COMSTAR SCSI STMF and SBD
SPARC T4 support and optimizations, Surprise!
Critical Threads support in kernel to enable faster performance on newer CPUs
SR-IOV support to enhance virtualization, FINALLY!
NUMA I/O Support!
Intel AVX Support
Dynamic Intimate Shared Memory enhancements with 8x performance gains get those Oracle Databases up faster!
Suspend and Resume on RAM for certain platforms
Enhanced FMA support in x86
DTrace cpc provider for CPU profiling
Tons of open source software in the IPS repository
Xorg with Gnome 2.30.2 and stable with Firefox 6, Thunderbird 6, etc. Compviz support for fancy eyecandy!
GNU and BSDisms integrated into /usr/bin and /usr/gnu/bin
Bash is the default shell with ksh93 being the default system shell
HAL and DBus replace the old vold framework for removeable media
Open Sound System API replaces old sound framework
Enhancements for the old Man Pages with search strings and an SMF service to build your indexes! No more manual catman jobs!
Virtual Console support, FINALLY!
Time Slider tool for managing ZFS snapshots of home directories, enables easy backups and restores for users
ZFS SMF service for automating snapshots
CUPS replaces the old LPR print system, FINALLY!
LibC enhancements to support compatibility with Linux and BSD
path.h included to help developers with knowing common paths in Solaris for application development
Over 200 Locales supported, without a doubt the best cross locale support out there
Enhanced True Type fonts

And now the massive upgrades begin at home and on my laptop VMs. I'll post my experiences and thoughts as I go through this process.

Oracle Solaris Online Forum

2011-04-01T14:06:00.001-05:00

Oracle will be hosting an on-line forum about Solaris 11! I'm planning on attending this on-line session. Here is the agenda:

Thursday, April 14th, 2011
All times are listed in US Pacific.

9:00 a.m. - 9:45 a.m. PT
Oracle Solaris Strategy Overview
Bill Nesheim, VP Oracle Solaris Engineering
Live online chat available for this session
Add this event to your Outlook calendar

9:45 a.m. - 10:00 a.m. PT
An Industry Analyst's View of the Operating System Market
Gary Chen, IDC
Live online chat available for this session
Add this event to your Outlook calendar

10:00 a.m. - 10:45 a.m. PT
Manage Your Deployments With Image Packaging System and the Automated Installer
Bart Smaalders, Oracle Solaris Engineering
Dave Miner, Oracle Solaris Engineering
Glynn Foster, Oracle Solaris Product Management
Isaac Rozenfeld, Oracle Solaris Product Management
Live online chat available for this session
Add this event to your Outlook calendar

10:45 a.m. - 11:30 a.m. PT
Get More out of Your Oracle Solaris Environments With Virtualization
Dan Price, Oracle Solaris Engineering
Nicolas Droux, Oracle Solaris Engineering
Duncan Hardie, Oracle Solaris Product Management
Live online chat available for this session
Add this event to your Outlook calendar

11:30 a.m. - 12:15 p.m. PT
Learn How All New Features in Oracle Solaris 11 Raise The Bar For Operating Systems
Markus Flierl, Sr. Director Oracle Solaris Engineering
Liane Praza, Oracle Solaris Engineering
Joost Pronk, Oracle Solaris Product Management
Live online chat available for this session
Add this event to your Outlook calendar

SPARC Punches Back!

2010-12-03T22:41:00.000-06:00

Larry Ellison at Oracle made some really exciting announcements yesterday on SPARC and Solaris! It's definitely worth registering and watching the event to find all the details. But here are the high lights:

SPARC SuperCluster
SPARC ExaLogic
Oracle Beats IBM on TPC-C Benchmark at 30M Transactions per Minute!
SPARC T4 Processor
SPARC64 VII+ (M3) Processor

Honestly, I don't think I've seen Larry soo excited! So basically Oracle now has several pre-configured SPARC based appliances that can run Solaris 10 or Solaris 11. The SPARC SuperCluster can be used for databases or general computing, while the SPARC ExaLogic can be used for the Oracle Fusion middleware products. There will be three SuperCluster configurations, one based on the SPARC T3-2 servers, the second based on the SPARC T3-4 servers, and the last one based on the M5000 servers. The SPARC ExaLogic product is based on the SPARC T3-1B blades. All of the configurations leverage the same management tools, Infiniband fabric, flash storage, flash cache, and the ZFS based 7000 series arrays as the x64 ExaData and ExaLogic configurations in place today. As a result, the SPARC SuperCluster and the SPARC ExaLogic systems can scale big time!

I think the important thing to keep in mind here is that the SPARC T3 processor obviously can pack a bigger punch than HP or IBM and with a much smaller foot-print and cost. Considering that it now has the TPC-C and the SPECweb top spots, it'll be harder for the FUD spreaders to say that SPARC is slow or out-dated. Clearly betting on the whole CMT concept has paid off and forced the industry to follow.

The SPARC T4 processor was announced in this event as well. As many have speculated these processors will have 8 cores and focus on improving single-threaded performance by a factor of 3. These should use the new VT (Yellowstone) core that Sun talked about in the past that will start out with 8 cores and 8 threads. These should go beyond the 2Ghz clock rate. It's great to hear that these processors are already in the labs and will come out next year. It would appear that Oracle is ensuring that Sun engineers and TSMC get things tapped out in a reasonable amount of time. If that means that the T4 servers will come out next year, then I would expect that we'll see the T5 servers in 2012/2013 time-frame. All together, this is great news and I can't wait to get my hands on a T4 box!

The SPARC64-VII+ or SPARC M3 processor as Larry likes to call it, was also announced. This is nothing new if you attended the Oracle OpenWorld 2010 event like I did. This brings a 20% boost in performance with a higher clock rate (up to 3Ghz finally!) and 12MB cache! Of course this upgrade is available on the M4000-M9000 servers, with full backward compatibility. As pointed out by Fujitsu and Oracle, this is the 4th generation processor upgrade for the M-series servers, which is definitely a lot of investment protection.

Now for some goodies!

You can also read what the Register had to say on this here and here.

Using Auto Installer on Solaris 11 Express

2010-12-03T20:51:00.000-06:00

If you've ever had a conversation with me about Jumpstart, you know that I'm a big believer in the Jumpstart Enterprise Toolkit (JET). I've used it for over 7 years now and it's usually one of the first things I've setup for shops when I walk in the door to find that they don't have a good Solaris provisioning mechanism.

With OpenSolaris and thusly Solaris 11 Express, many changes have taken place around packaging, patching, and operating system installation. The one thing that has definitely changed things is Auto Installer, or AI as it's known. AI works with many of the same assumptions as IPS and SMF. No longer do you have pre or post installation scripts. You don't have all the options you've been use to using with things like JET or N1 SPS. However, this has been improving incrementally over the past year, but still has some way to go. I wanted to post what I've found to work.

The first step on this journey is to download the Solaris 11 Express AI images from:

http://www.oracle.com/technetwork/server-storage/solaris11/downloads/index.html

Once you've done that, you need to install the AI tools:

$ pfexec pkg install installadm
...

As you can see this service is dependent upon DNS, TFTP, and DHCP. If you don't have DHCP setup, it'll configure it for you. Or you can set it up with the following:

$ pfexec dhcpconfig -D -r SUNWfiles -p /var/tmp
Created DHCP configuration file.
Created dhcptab.
Added "Locale" macro to dhcptab.
Added server macro to dhcptab - katana.
DHCP server started.

$ pfexec dhcpconfig -N -m -t
$ pfexec dhtadm -M -m -e DNSserv=

If you have TCP Wrappers enabled, you'll have to open it up in your /etc/hosts.allow file with a "in.tftpd:ALL" line added to it.

Once that has been done, you can create what is known as an "install service" in AI:

$ pfexec installadm create-service -n s11-151a-sparc -s /export/install/isos/sol-11-exp-201011-ai-sparc.iso /export/install/aiserver/s11-151a-sparc
Setting up the target image at /export/install/aiserver/s11-151a-sparc ...
Registering the service s11-151a-sparc._OSInstall._tcp.local
Service discovery fallback mechanism set up
Creating SPARC configuration file

$ pfexec installadm create-service -n s11-151a-x86 -s /export/install/isos/sol-11-exp-201011-ai-x86.iso /export/install/aiserver/s11-151a-x86
Setting up the target image at /export/install/aiserver/s11-151a-x86 ...
Registering the service s11-151a-x86._OSInstall._tcp.local
copying boot file to /tftpboot/pxegrub.I86PC.Solaris-1
Service discovery fallback mechanism set up

What the first command will do is setup an installation service called "s11-151a-sparc" using the ISO we've downloaded and installing it into "/export/install/aiserver/s11-151a-sparc". The second command creates a service for x86 servers using similar syntax and of course the x86 ISO. This will enable SPARC clients to boot with WANBOOT and for x86 clients to boot with PXE. This will register the image with SMF install service, which you can check on below:

$ svcs -l install/server
fmri svc:/system/install/server:default
name Installadm Utility
enabled true
state online
next_state none
state_time November 27, 2010 02:03:25 AM CST
logfile /var/svc/log/system-install-server:default.log
restarter svc:/system/svc/restarter:default
contract_id 134
dependency optional_all/restart svc:/network/dns/multicast:default (online)
dependency optional_all/none svc:/network/tftp/udp6:default (online)
dependency optional_all/none svc:/network/dhcp-server:default (online)

This service will start-up Apache instances for each install service you create:

$ svcs -p install/server
STATE STIME FMRI
online Nov_27 svc:/system/install/server:default
Nov_27 1732 webserver
Nov_27 2164 webserver
Nov_27 2500 httpd
Nov_27 2571 httpd
Nov_27 2572 httpd
Nov_27 2573 httpd
Nov_27 2574 httpd
Nov_27 2575 httpd

You can see your AI service details with the installadm command:

$ installadm list
Service Name Status Arch Port Image Path
------------ ------ ---- ---- ----------
osol-b134-sparc off Sparc 46501 /export/install/aiserver/osol-b134-sparc
s11-151a-sparc on Sparc 46502 /export/install/aiserver/s11-151a-sparc
s11-151a-x86 on x86 46503 /export/install/aiserver/s11-151a-x86

At this point, your AI setup is very generic. So how do you customize it? Well you have to modify the default manifest for each service and create further manifests for any client specific customizations you want to do. Now, you can definitely read the documentation on doc.sun.com, which I would highly recommend considering this will continue to change.

However, to help you get started I've provided a sample default manifest you can use. The default manifest for each service is called default.xml, which is an XML file. The important thing about this file is that it'll be used as the base manifest and is critical to be careful with editing it, I would recreate a backup copy just in case before doing anything. So here it is:

<?xml version="1.0" encoding="UTF-8"?>  <!DOCTYPE auto_install SYSTEM "file:///usr/share/auto_install/ai.dtd"> <auto_install> <ai_instance name="default" auto_reboot="true"> <target> <target_device> <swap> <zvol action="create" name="swap"> <size val="4gb"/> </zvol> </swap> <dump> <zvol action="create" name="dump"> <size val="4gb"/> </zvol> </dump> </target_device> </target> <software> <source> <publisher name="solaris"> <origin name="http://pkg.oracle.com/solaris/release"/> </publisher> </source>  <software_data action="install" type="IPS"> <name>pkg:/entire</name> <name>pkg:/babel_install</name>  <name>pkg:/network/iscsi/initiator</name> <name>pkg:/network/iscsi/iser</name> </software_data>  <software_data action="uninstall" type="IPS"> <name>pkg:/babel_install</name> <name>pkg:/slim_install</name> </software_data> </software>  <add_drivers> <search_all/> </add_drivers> <sc_embedded_manifest name="AI">  </sc_embedded_manifest> </ai_instance> </auto_install>

This will do the following:

Make clients reboot after the installation
Install the base Solaris 11 Express image
Install any drivers your platform requires that are not in the boot image
Create the "operator" account for doing administrative tasks using RBAC with the password of "newroot"
Set the root password to "newroot"
Set the timezone
Set the host name to "solaris"
Set the console type to "vt100"
Set the keyboard layout to "US-English"
Enable the NWAM service to auto configure the networking

Now this will enable your clients to come up with some account settings to get you logged in and set the language and console type. The client will grab the IP information from DHCP and call itself "solaris". But this hardly matches up with what Jumpstart could do for us. So now lets create a manifest for a client called the "test":

/export/install/aiserver/s11-151a-sparc/auto_install$ pfexec mkdir clients /export/install/aiserver/s11-151a-sparc/auto_install/clients$ pfexec vi test.xml <?xml version="1.0" encoding="UTF-8"?>  <!DOCTYPE auto_install SYSTEM "file:///usr/share/auto_install/ai.dtd"> <auto_install> <ai_instance name="test" auto_reboot="true"> <target> <target_device> <disk> <disk_keyword key="boot_disk"/> </disk> </target_device> <target_device> <swap> <zvol action="create" name="swap"> <size val="4gb"/> </zvol> </swap> </target_device> <target_device> <dump> <zvol action="create" name="dump"> <size val="4gb"/> </zvol> </dump> </target_device> </target> <software> <source> <publisher name="solaris"> <origin name="http://pkg.oracle.com/solaris/release"/> </publisher> </source>  <software_data action="install" type="IPS"> <name>pkg:/entire</name> <name>pkg:/babel_install</name>  <name>pkg:/network/iscsi/initiator</name> <name>pkg:/network/iscsi/iser</name> </software_data>  <software_data action="uninstall" type="IPS"> <name>pkg:/babel_install</name> <name>pkg:/slim_install</name> </software_data> </software>  <add_drivers> <search_all/> </add_drivers> <sc_embedded_manifest name="AI">  </sc_embedded_manifest> </ai_instance> </auto_install>

This will set the host name and the static IP configuration, which is a bit more helpful than the default manifest. Now we must associate this manifest with the install service:

$ pfexec installadm add-manifest -m /export/install/aiserver/s11-151a-sparc/auto_install/clients/test.xml -n s11-151a-sparc -c MAC="00:14:4b:26:31:ab"
$ installadm list -m
Service Name Manifest
------------ --------
s11-151a-sparc test.xml

That will associate the manifest with the installation service and tie it to the MAC address of our client, which we haven't configured yet. You can use other ways of associating a manifest with an install service. But this is the easiest for associating it with a specific client.

Now lets create our client:

$ pfexec installadm create-client -e 0:14:4b:26:31:ab -t /export/install/aiserver/s11-151a-sparc -n s11-151a-sparc
Creating SPARC configuration file

$ installadm list
Service Name Status Arch Port Image Path
------------ ------ ---- ---- ----------
osol-b134-sparc off Sparc 46501 /export/install/aiserver/osol-b134-sparc
s11-151a-sparc on Sparc 46502 /export/install/aiserver/s11-151a-sparc
s11-151a-x86 on x86 46503 /export/install/aiserver/s11-151a-x86
$ installadm list -c
Service Name Client Address Arch Image Path
------------ -------------- ---- ----------
s11-151a-sparc 00:14:4b:26:31:AB Sparc /export/install/aiserver/s11-151a-sparc

At this point you can do an AI installation on the client:

{0} ok boot net:dhcp - install
Boot device: /pci@500/pci@0/pci@8/network@0:dhcp File and args: - install
/pci@500/pci@0/pci@8/network@0: 100 Mbps full duplex link up
wanboot info: WAN boot messages->console
wanboot info: configuring /pci@500/pci@0/pci@8/network@0:dhcp

/pci@500/pci@0/pci@8/network@0: 100 Mbps full duplex link up
wanboot info: Starting DHCP configuration
wanboot info: DHCP configuration succeeded
wanboot progress: wanbootfs: Read 366 of 366 kB (100%)
wanboot info: wanbootfs: Download complete
Tue Nov 23 03:01:22 wanboot progress: miniroot: Read 191571 of 191571 kB (100%)
Tue Nov 23 03:01:22 wanboot info: miniroot: Download complete
SunOS Release 5.11 Version snv_151a 64-bit
Copyright (c) 1983, 2010, Oracle and/or its affiliates. All rights reserved.
WARNING: Last shutdown is later than time on time-of-day chip; check date.
Hostname: solaris
Remounting root read/write
Probing for device nodes ...
Preparing automated install image for use
The AI image will be retrieved from /export/install/aiserver/s11-151a-sparc/ directory
Downloading solaris.zlib archive
--2009-11-23 02:51:18-- http://192.168.1.3:5555/export/install/aiserver/s11-151a-sparc//solaris.zlib
Connecting to 192.168.1.3:5555... connected.
HTTP request sent, awaiting response... 200 OK
Length: 101885952 (97M) [text/plain]
Saving to: `/tmp/solaris.zlib'

100%[======================================>] 101,885,952 11.2M/s in 8.7s

2009-11-23 02:51:27 (11.2 MB/s) - `/tmp/solaris.zlib' saved [101885952/101885952]

Downloading solarismisc.zlib archive
--2009-11-23 02:51:27-- http://192.168.1.3:5555/export/install/aiserver/s11-151a-sparc//solarismisc.zlib
Connecting to 192.168.1.3:5555... connected.
HTTP request sent, awaiting response... 200 OK
Length: 7248896 (6.9M) [text/plain]
Saving to: `/tmp/solarismisc.zlib'

100%[======================================>] 7,248,896 11.1M/s in 0.6s

2009-11-23 02:51:27 (11.1 MB/s) - `/tmp/solarismisc.zlib' saved [7248896/7248896]

--2009-11-23 02:51:27-- http://192.168.1.3:5555/export/install/aiserver/s11-151a-sparc//install.conf
Connecting to 192.168.1.3:5555... connected.
HTTP request sent, awaiting response... 200 OK
Length: 69 [text/plain]
Saving to: `/tmp/install.conf'

100%[======================================>] 69 --.-K/s in 0s

2009-11-23 02:51:27 (1.74 MB/s) - `/tmp/install.conf' saved [69/69]

Done mounting automated install image
Configuring devices.
Service discovery phase initiated
Service name to look up: s11-151a-sparc
Nov 23 02:52:22 solaris sendmail[766]: My unqualified host name (localhost) unknown; sleeping for retry
Service discovery finished successfully
Process of obtaining configuration manifest initiated
Using the configuration manifest obtained via service discovery

Automated Installation started
The progress of the Automated Installation can be followed by viewing
the logfile at /tmp/install_log

solaris console login: Nov 23 02:53:24 solaris sendmail[766]: unable to qualify my own domain name (localhost) -- using short name
Automated Installation finished successfully
Automated reboot enabled. The system will be rebooted now
Log files will be available in /var/sadm/system/logs/ directory after reboot
Nov 23 04:00:11 solaris reboot: initiated by root
Nov 23 04:00:18 solaris syslogd: going down on signal 15
syncing file systems... done
rebooting...
Resetting...

Boot device: /pci@400/pci@0/pci@8/scsi@0/disk@0,0:a File and args: -Z rpool/ROOT/solaris
SunOS Release 5.11 Version snv_151a 64-bit
Copyright (c) 1983, 2010, Oracle and/or its affiliates. All rights reserved.
Loading smf(5) service descriptions: 174/174
Hostname: test
Configuring devices.
Loading smf(5) service descriptions: 20/20

test console login:

The one thing that I've found that does not work in the manifest is the sizing of the swap and the dump datasets in the ZFS rpool. But this is easy to fix afterwards:

root@test:~# zfs set volsize=4g rpool/swap
root@test:~# zfs set volsize=4g rpool/dump

I'm sure I'll find more things to play with in AI. But I think you'll be able to learn from this and get started. Hopefully we'll see some plug-ins in JET and Ops Center next to support AI.

Upgrading from OpenSolaris to Solaris 11 Express

2010-12-03T20:17:00.000-06:00

Over the Thanksgiving break, I worked on upgrading all of my OpenSolaris systems to Solaris 11 Express. The journey wasn't as difficult as one might have expected and I wanted to share the steps for doing this with everyone.

The first thing that you should do is to upgrade to the last build of OpenSolaris, known as build 134b. This will enable you to then upgrade to Solaris 11 Express and is a requirement.

$ pfexec pkg set-publisher -P -O http://pkg.opensolaris.org/release/ opensolaris.org
$ pkg publisher
PUBLISHER TYPE STATUS URI
opensolaris.org (preferred) origin online http://pkg.opensolaris.org/release/

At this point your OpenSolaris server is ready for doing an image update with IPS. Now an important consideration here is if you have Containers running. You will have to update them as well and instead of going through and manually updating them, we'll allow the "-U attach" command-line option for zoneadm do the hard work for us. So for now, lets bring down any Containers that are running.

$ zoneadm list
global
takeda
miroku
$ pfexec zoneadm -z takeda halt
$ pfexec zoneadm -z takeda detach
$ pfexec zoneadm -z miroku halt
$ pfexec zoneadm -z miroku detach
$ zoneadm list -cv
ID NAME STATUS PATH BRAND IP
0 global running / ipkg shared
- takeda configured /zones/takeda ipkg shared
- miroku configured /zones/miroku ipkg shared

Now that the Containers are shutdown and detached, lets focus on the global-zone. One of the things I like to do when performing an image update is to make sure that I name the boot environment something meaningful, so lets call the new boot environment "opensolaris-b134b". You can pass this information to the pkg command, here the full out-put of an upgrade:

$ beadm list
BE Active Mountpoint Space Policy Created
-- ------ ---------- ----- ------ -------
opensolaris-b133 - - 29.63M static 2010-02-21 20:17
opensolaris-b133-1 NR / 33.67G static 2010-09-29 22:00
$ pfexec pkg image-update --be-name opensolaris-b134b
DOWNLOAD PKGS FILES XFER (MB)
Completed 1063/1063 13433/13433 327.7/327.7

PHASE ACTIONS
Removal Phase 7165/7165
Install Phase 7234/7234
Update Phase 18965/18965
PHASE ITEMS
Reading Existing Index 8/8
Indexing Packages 1063/1063
Indexing Packages 1063/1063
Optimizing Index...
PHASE ITEMS
Indexing Packages 1127/1127

A clone of opensolaris-b133-1 exists and has been updated and activated.
On the next boot the Boot Environment opensolaris-b134b will be mounted on '/'.
Reboot when ready to switch to this updated BE.

$ beadm list
BE Active Mountpoint Space Policy Created
-- ------ ---------- ----- ------ -------
opensolaris-b133 - - 29.63M static 2010-02-21 20:17
opensolaris-b133-1 N / 11.66M static 2010-09-29 22:00
opensolaris-b134b R - 35.48G static 2010-11-19 16:56

At this point we can reboot the server and move onto updating our Containers before the big move to Solaris 11 Express.

$ pfexec shutdown -y -g0 -i 6
...
$ uname -a
SunOS katana 5.11 snv_134b i86pc i386 i86pc
$ cat /etc/release
OpenSolaris 2010.05 snv_134b X86
Copyright (c) 2010, Oracle and/or its affiliates. All rights reserved.
Assembled 28 May 2010

At this point we can now update our Containers using the zoneadm command:

$ pfexec zoneadm -z takeda attach -U
Log File: /var/tmp/takeda.attach_log.CRaydi
Attaching...

Global zone version: entire@0.5.11,5.11-0.134.0.2:20100528T233058Z
Non-Global zone version: entire@0.5.11,5.11-0.134:20100302T023003Z
Publisher Check: Zone preferred publisher does not contain
entire@0.5.11,5.11-0.134.0.2:20100528T233058Z.
Publisher Reset: Copying preferred publisher from global zone.
Cache: Using /var/pkg/download.
Updating non-global zone: Output follows
DOWNLOAD PKGS FILES XFER (MB)
Completed 112/112 2168/2168 48.7/48.7

PHASE ACTIONS
Removal Phase 1913/1913
Install Phase 3146/3146
Update Phase 2708/2708
Updating non-global zone: Zone updated to entire@0.5.11,5.11-0.134.0.2:20100528T233058Z
Attach complete.

$ pfexec zoneadm -z miroku attach -U
Log File: /var/tmp/miroku.attach_log.igaWHi
Attaching...

Global zone version: entire@0.5.11,5.11-0.134.0.2:20100528T233058Z
Non-Global zone version: entire@0.5.11,5.11-0.134:20100302T023003Z
Publisher Check: Zone preferred publisher does not contain
entire@0.5.11,5.11-0.134.0.2:20100528T233058Z.
Publisher Reset: Copying preferred publisher from global zone.
Cache: Using /var/pkg/download.
Updating non-global zone: Output follows
DOWNLOAD PKGS FILES XFER (MB)
Completed 112/112 2168/2168 48.7/48.7

PHASE ACTIONS
Removal Phase 1913/1913
Install Phase 3146/3146
Update Phase 2708/2708
Updating non-global zone: Zone updated to entire@0.5.11,5.11-0.134.0.2:20100528T233058Z
Attach complete.

At this point our Containers are also upgraded. So now we can move onto the big upgrade, but lets first detach our Containers.

$ pfexec zoneadm -z takeda detach
$ pfexec zoneadm -z miroku detach

Now it's time to set the IPS publisher to the official Solaris 11 Express repository:

$ pkg publisher
PUBLISHER TYPE STATUS URI
opensolaris.org (preferred) origin online http://pkg.opensolaris.org/release/

$ pfexec pkg set-publisher --non-sticky opensolaris.org
$ pfexec pkg set-publisher -P -g http://pkg.oracle.com/solaris/release/ solaris

Before updating to the Solaris 11 Express release, we have to update our pkg tool set:

$ pfexec pkg install pkg:/package/pkg
DOWNLOAD PKGS FILES XFER (MB)
Completed 1/1 1/1 0.0/0.0

PHASE ACTIONS
Update Phase 2/2
PHASE ITEMS
Reading Existing Index 8/8
Indexing Packages 1/1

At this point we can now do the image update to Solaris 11 Express. We'll use "solaris11-151a" as the name for the new boot environment. Notice the OTN license acceptance option and how it's printed out so you don't forget.

$ pfexec pkg image-update --be-name solaris11-151a --accept
------------------------------------------------------------
Package: pkg://solaris/consolidation/osnet/osnet-incorporation@0.5.11,5.11-0.151.0.1:20101104T230646Z
License: usr/src/pkg/license_files/lic_OTN

Oracle Technology Network Developer License Agreement

Oracle Solaris, Oracle Solaris Cluster and Oracle Solaris Express

EXPORT CONTROLS
Selecting the "Accept License Agreement" button is a confirmation
of your agreement that you comply, now and during the trial term
(if applicable), with each of the following statements:

-You are not a citizen, national, or resident of, and are not under
control of, the government of Cuba, Iran, Sudan, North Korea, Syria,
or any country to which the United States has prohibited export.

-You will not download or otherwise export or re-export the Programs
(as defined below), directly or indirectly, to the above mentioned
countries or to citizens, nationals or residents of those countries.

-You are not listed on the United States Department of Treasury lists
of Specially Designated Nationals, Specially Designated Terrorists,
and Specially Designated Narcotic Traffickers, nor are you listed on
the United States Department of Commerce Table of Denial Orders.

You will not download or otherwise export or re-export the Programs,
directly or indirectly, to persons on the above mentioned lists.

You will not use the Programs for, and will not allow the Programs to
be used for, any purposes prohibited by United States law, including,
without limitation, for the development, design, manufacture or
production of nuclear, chemical or biological weapons of mass
destruction.

EXPORT RESTRICTIONS
You agree that U.S. export control laws and other applicable export
and import laws govern your use of the Programs, including technical
data; additional information can be found on Oracle's Global Trade
Compliance web site (http://www.oracle.com/products/export).

You agree that neither the Programs nor any direct product thereof
will be exported, directly, or indirectly, in violation of these
laws, or will be used for any purpose prohibited by these laws
including, without limitation, nuclear, chemical, or biological
weapons proliferation.

Oracle Employees: Under no circumstances are Oracle Employees
authorized to download software for the purpose of distributing it to
customers. Oracle products are available to employees for internal
use or demonstration purposes only. In keeping with Oracle's trade
compliance obligations under U.S. and applicable multilateral law,
failure to comply with this policy could result in disciplinary action
up to and including termination.

PLEASE READ THE FOLLOWING LICENSE AGREEMENT TERMS AND CONDITIONS
CAREFULLY BEFORE DOWNLOADING, INSTALLING OR USING THE PROGRAMS.
THESE TERMS AND CONDITIONS CONSTITUTE A LEGAL AGREEMENT BETWEEN YOU
AND ORACLE.

Oracle Technology Network Development License Agreement
"We," "us," and "our" refers to Oracle America, Inc., for and on
behalf of itself and its subsidiaries and affiliates under common
control. "You" and "your" refers to the individual or entity that
wishes to use the Programs. "Programs" refers to Oracle Solaris,
Oracle Solaris Cluster and/or Oracle Solaris Express that you download
from this site (including, any updates and/or additional software
packages that you may receive in the future from the product's
package repository) and related program documentation. "License"
refers to your right to use the Programs under the terms of this
agreement. This agreement is governed by California law, except
for that body of laws related to the conflict of laws. You agree to
submit to the exclusive jurisdiction of, and venue in, the courts of
San Francisco or Santa Clara counties in California in any dispute
arising out of or relating to this agreement.

In order to use the Programs, You must first agree to this License
Agreement by selecting the "Accept License Agreement" button below. If
You do not or cannot agree to this License Agreement, You are not
permitted to download or use the Programs.

LICENSE RIGHTS
Except for any included software package or file that is licensed to
you by Oracle under different license terms, we grant you a perpetual
(unless terminated as provided in this agreement), nonexclusive,
nontransferable, limited License to use the Programs only for the
purpose of developing, testing, prototyping and demonstrating your
applications, and not for any other purpose.

All rights not expressly granted above are hereby reserved. If you
want to use the Programs for any purpose other than as permitted
under this agreement, including but not limited to distribution of the
Programs or any use of the Programs for your internal business purposes
(other than developing, testing, prototyping and demonstrating your
applications) or for any commercial production purposes, you must
obtain a valid license permitting such use. We may audit your use of
the Programs. Program documentation, if available, may be accessed
online at http://otn.oracle.com/docs.

Third-Party Technology. The Programs may contain or be distributed
with certain third-party technology. Oracle may provide certain notices
related to such third-party technology in the program documentation,
or in readme or notice files provided with the Programs.

Third party technology will be licensed to you either under the terms
of this agreement, or, if specified in the program documentation,
readme files or otherwise in writing, under separate license
terms ("Separate Terms") and not under the terms of this agreement
("Separately Licensed Third Party Technology"). Licensee's rights to
use such Separately Licensed Third Party Technology under the Separate
Terms are not restricted or modified in any way by this Agreement.

Information Collection and Registration
Configuration Data and Registration: The Programs may communicate
configuration data to Oracle. You can register your version of
the Programs to capture this data for your use, or the data is
sent anonymously. For information about what configuration data is
communicated and how to control this facility, refer to the release
notes or www.oracle.com/goto/solarisautoreg.

Ownership and Restrictions
We retain all ownership and intellectual property rights in the
Programs. The Programs may be installed on one computer only, and
used by one person in the hardware environment identified by us. You
may make one copy of the Programs for backup purposes.

You may not:
- use the Programs for your own internal business purposes (other than
developing, testing, prototyping and demonstrating your applications)
or for any commercial or production purposes;

- remove or modify any program markings or any notice of our
proprietary rights;

- make the Programs available in any manner to any third party;

- use the Programs to provide third-party training;

- assign this agreement or give or transfer the Programs or an interest
in them to another individual or entity;

- cause or permit reverse engineering (unless required by law for
interoperability), disassembly or decompilation of the Programs;

- disclose results of any benchmark test results related to the
Programs without our prior consent.

Export
You agree that U.S. export control laws and other applicable
export and import laws govern your use of the Programs,
including technical data; additional information can be
found on Oracle's Global Trade Compliance web site located at
http://www.oracle.com/products/export/index.html?content.html. You
agree that neither the Programs nor any direct product thereof will be
exported, directly, or indirectly, in violation of these laws, or will
be used for any purpose prohibited by these laws including, without
limitation, nuclear, chemical, or biological weapons proliferation.

Disclaimer of Warranty and Exclusive Remedies
THE PROGRAMS ARE PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. WE
FURTHER DISCLAIM ALL WARRANTIES, EXPRESS AND IMPLIED, INCLUDING
WITHOUT LIMITATION, ANY IMPLIED WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE OR NONINFRINGEMENT.

IN NO EVENT SHALL WE BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL,
PUNITIVE OR CONSEQUENTIAL DAMAGES, OR DAMAGES FOR LOSS OF PROFITS,
REVENUE, DATA OR DATA USE, INCURRED BY YOU OR ANY THIRD PARTY, WHETHER
IN AN ACTION IN CONTRACT OR TORT, EVEN IF WE HAVE BEEN ADVISED OF
THE POSSIBILITY OF SUCH DAMAGES. OUR ENTIRE LIABILITY FOR DAMAGES
HEREUNDER SHALL IN NO EVENT EXCEED ONE THOUSAND DOLLARS (U.S. $1,000).

Additional Trial Programs
We may include additional trial programs with your download of the
Programs licensed under this agreement. You will have 30 days from
the delivery date to evaluate these additional trial programs. Any
use of these trial programs after the 30-day trial period requires
you to obtain the applicable license. Any additional trial programs
are provided "as is" and we do not provide technical support or any
warranties for these programs.

No Technical Support
Our technical support organization will not provide technical support,
phone support, or updates to you for the Programs licensed under
this agreement.

End of Agreement
You may terminate this agreement by destroying all copies of the
Programs. We have the right to terminate your right to use the Programs
if you fail to comply with any of the terms of this agreement, in
which case you shall destroy all copies of the Programs.

Relationship Between the Parties
The relationship between you and us is that of
licensee/licensor. Neither party will represent that it has any
authority to assume or create any obligation, express or implied,
on behalf of the other party, nor to represent the other party as
agent, employee, franchisee, or in any other capacity. Nothing in
this agreement shall be construed to limit either party's right to
independently develop or distribute software that is functionally
similar to the other party's products, so long as proprietary
information of the other party is not included in such software.

Open Source Software
"Open Source" software - software available without charge for
use, modification and distribution - is often licensed under terms
that require the user to make the user's modifications to the Open
Source software or any software that the user 'combines' with the
Open Source software freely available in source code form. If you use
Open Source software in conjunction with the Programs (or if you plan
on licensing your own application under an Open Source license), you
must ensure that your use does not: (i) create, or purport to create,
obligations with respect to the Oracle Programs; or (ii) grant,
or purport to grant, to any third party any rights to or immunities
under our intellectual property or proprietary rights in the Oracle
Programs. For example, you may not develop a software program using
an Oracle program and an Open Source program where such use results
in a program file(s) that contains code from both the Oracle program
and the Open Source program (including without limitation libraries)
if the Open Source program is licensed under a license that requires
any "modifications" be made freely available. You also may not combine
the Oracle program with programs licensed under the GNU General Public
License ("GPL") in any manner that could cause, or could be interpreted
or asserted to cause, the Oracle program or any modifications thereto
to become subject to the terms of the GPL.

Entire Agreement
You agree that this agreement is the complete agreement for the
Programs and licenses, and this agreement supersedes all prior or
contemporaneous agreements or representations, including any and
all clickwrap, shrinkwrap or similar licenses. If any term of this
agreement is found to be invalid or unenforceable, the remaining
provisions will remain effective.

Last updated: 08/12/10

Should you have any questions concerning this License Agreement,
or if you desire to contact Oracle for any reason, please write:

Oracle America, Inc.
500 Oracle Parkway,
Redwood City, CA 94065

Oracle may contact you to ask if you had a satisfactory experience
installing and using this OTN software download.

DOWNLOAD PKGS FILES XFER (MB)
Completed 1275/1275 59454/59454 1156.5/1156.5

PHASE ACTIONS
Removal Phase 14947/25757
Warning - directory usr/share/icons/Neutral_Plus_Inv not empty - contents preserved in /tmp/tmp7tTc8r/var/pkg/lost+found/usr/share/icons/Neutral_Plus_Inv-20101119T184748Z
Removal Phase 25757/25757
Install Phase 36238/36238
Update Phase 55281/55281
PHASE ITEMS
Reading Existing Index 8/8
Indexing Packages 1275/1275
Indexing Packages 1275/1275
Optimizing Index...
PHASE ITEMS
Indexing Packages 1265/1265

A clone of opensolaris-b134b exists and has been updated and activated.
On the next boot the Boot Environment solaris11-151a will be mounted on '/'.
Reboot when ready to switch to this updated BE.

---------------------------------------------------------------------------
NOTE: Please review release notes posted at:

http://docs.sun.com/doc/821-1479
---------------------------------------------------------------------------

At this point we can reboot and bring up Solaris 11 Express!

$ shutdown -y -g0 -i 6
...
$ uname -a
SunOS katana 5.11 snv_151a i86pc i386 i86pc

$ cat /etc/release
Oracle Solaris 11 Express snv_151a X86
Copyright (c) 2010, Oracle and/or its affiliates. All rights reserved.
Assembled 04 November 2010

Now, lets upgrade our Containers using the zoneadm command:

$ pfexec zoneadm -z takeda attach -U
Log File: /var/tmp/takeda.attach_log.ggaiMK
Attaching...

preferred global publisher: solaris
Global zone version: entire@0.5.11,5.11-0.151.0.1:20101105T054056Z
Non-Global zone version: entire@0.5.11,5.11-0.134.0.2:20100528T233058Z

Cache: Using /var/pkg/download.
Updating non-global zone: Output follows
Packages to remove: 2
Packages to install: 84
Packages to update: 109
Create boot environment: No
Services to restart: 6
------------------------------------------------------------
Package: pkg://solaris/consolidation/osnet/osnet-incorporation@0.5.11,5.11-0.151.0.1:20101104T230646Z
License: usr/src/pkg/license_files/lic_OTN

Oracle Technology Network Developer License Agreement

Oracle Solaris, Oracle Solaris Cluster and Oracle Solaris Express
...
DOWNLOAD PKGS FILES XFER (MB)
Completed 195/195 10886/10886 114.6/114.6

PHASE ACTIONS
Removal Phase 3183/3183
Install Phase 11829/11829
Update Phase 8095/8095

PHASE ITEMS
Package State Update Phase 304/304
Package Cache Update Phase 111/111
Image State Update Phase 2/2
Updating non-global zone: Zone updated.
Result: Attach Succeeded.

$ pfexec zoneadm -z miroku attach -U
Log File: /var/tmp/miroku.attach_log.9faqwL
Attaching...

preferred global publisher: solaris
Global zone version: entire@0.5.11,5.11-0.151.0.1:20101105T054056Z
Non-Global zone version: entire@0.5.11,5.11-0.134.0.2:20100528T233058Z

Cache: Using /var/pkg/download.
Updating non-global zone: Output follows
Packages to remove: 2
Packages to install: 84
Packages to update: 109
Create boot environment: No
Services to restart: 6
------------------------------------------------------------
Package: pkg://solaris/consolidation/osnet/osnet-incorporation@0.5.11,5.11-0.151.0.1:20101104T230646Z
License: usr/src/pkg/license_files/lic_OTN

Oracle Technology Network Developer License Agreement

Oracle Solaris, Oracle Solaris Cluster and Oracle Solaris Express
...
DOWNLOAD PKGS FILES XFER (MB)
Completed 195/195 10886/10886 114.6/114.6

PHASE ACTIONS
Removal Phase 3183/3183
Install Phase 11829/11829
Update Phase 8095/8095

PHASE ITEMS
Package State Update Phase 304/304
Package Cache Update Phase 111/111
Image State Update Phase 2/2
Updating non-global zone: Zone updated.
Result: Attach Succeeded.

Now we can bring up our Containers and check them out.

$ zoneadm list -cv
ID NAME STATUS PATH BRAND IP
0 global running / ipkg shared
- takeda installed /zones/takeda ipkg shared
- miroku installed /zones/miroku ipkg shared

$ pfexec zoneadm -z takeda boot
$ pfexec zoneadm -z miroku boot
$ zoneadm list -cv
ID NAME STATUS PATH BRAND IP
0 global running / ipkg shared
1 takeda running /zones/takeda ipkg shared
2 miroku running /zones/miroku ipkg shared

root@takeda:~# uname -a
SunOS takeda 5.11 snv_151a i86pc i386 i86pc
root@takeda:~# cat /etc/release
Oracle Solaris 11 Express snv_151a X86
Copyright (c) 2010, Oracle and/or its affiliates. All rights reserved.
Assembled 04 November 2010

root@miroku:~# uname -a
SunOS miroku 5.11 snv_151a i86pc i386 i86pc
root@miroku:~# cat /etc/release
Oracle Solaris 11 Express snv_151a X86
Copyright (c) 2010, Oracle and/or its affiliates. All rights reserved.
Assembled 04 November 2010

So there we are! All of our Containers are updated nicely with our Global Zone. I've been able to do this on x86, SPARC, in LDoms, and even VirtualBox!

Enjoy!

Oracle Solaris 11 Express Released!

2010-11-15T17:18:00.000-06:00

Solaris 11 Express has finally been released! This has been a long time in coming and I'm very excited to finally see this day. Just so that folks are clear, this is a full express release with support for developers, system administrators, evaluators, enthusiasts, etc. on x86 and SPARC! It is totally free to use as long as it is not used on production. As you can see on the main link above, Oracle is selling a full suite of support for Solaris 11 Express, if you are looking for support or to use it in production. Oracle is dead serious about Solaris, so make no mistake about it!

Needless to say, I'll be busy downloading and upgrading my systems to this release. I'll make some additional blog posts once I have things in place to take it for a full spin on both x86 and SPARC. I'll leverage my Ultra 20, some VirtualBox instances, and some LDoms to make things interesting!

So what are you waiting for??? Go download Solaris 11 Express now!

Oracle Solaris Summit at LISA!

2010-11-09T10:35:00.000-06:00

At LISA this year, there is a Oracle Solaris Summit at the conference. I wish I could be there today to attend. However, thanks to technology, we can watch this event online! This is a excellent opportunity to get a look at what's happening with Solaris 11 Express!

Update:

The slides for this summit are located here.

Agenda:

Solaris Summit at LISA10

Tuesday, Nov 9, 2010 in Ballroom A4/A5, San Jose Convention Center.

Agenda:

9:00 - 9:30 am - Introduction to Oracle Solaris 11 Express, by Markus Flierl
9:30 - 11:00 am - Image Packaging System, by Bart Smaalders
11:00 am - 12:30 pm - Deploying Oracle Solaris 11 in the Enterprise, by Dave Miner
12:30 pm - 1:30 pm - LUNCH BREAK
1:30 pm - 2:30 pm - Advances in Solaris Networking with Crossbow and Beyond, by Nicolas Droux
2:30 pm - 3:00 pm - Oracle Solaris Containers in Oracle Solaris 11 Express, by Dan Price
3:00 pm - 3:15 pm - BREAK
3:15 pm - 4:15 pm - ZFS Features in Oracle Solaris Express, by Cindy Swearingen
4:15 pm - 4:45 pm - New Security Features in Oracle Solaris 11 Express, by Glenn Faden
4:45 pm - 5:30 pm - Deploying Applications Using SMF and Other Solaris 11 Features, by Liane Praza
5:30 pm - 6:30 pm - Beer and Snacks Reception for ALL

Interview With Andy Bechtolsheim and My Afterthoughts

2010-11-01T23:39:00.000-05:00

Dan Worth on V3 wrote an article about an interview with Andy Bechtolsheim who was one of the original co-founders of Sun Microsystems, who returned some years back to help design many of the servers that are deployed today from Sun Oracle. Andy has some interesting insights into Sun and the computing market in general. I do agree with him that Sun was too focused on bringing products to market for developers and technophiles instead of generating the kinds of profit it should have. And this definitely led to the need for change and why Oracle is managing things differently. Sun was too busy giving away software and technology for free and not pushing to get revenue. Oracle still gives away software, like Solaris, for free in non-production environments. But it expects payment now for production use. This has caused some flak from the user base and I have to side with Oracle on this one. Giving software away for free or open sourcing it is of course a "nice thing to do" and great for schools and start-ups. But all of that engineering work has to be paid for and of course stock holders expect a profit. Sun just wasn't making the money it should have from all of the intellectual property it had developed. I think Oracle has the right mindset. Give it away for free in non-production environments, but expect payment when used in production. To me this is perfectly reasonable.

Andy also makes a point that the way people use computers is changing. The revolution of mobile devices such as the iPhone and the iPad have truely changed the perspective on how computing should be. It should be easy and require very little effort on the part of the user or consumer. The heavy lifting should be on the back-end in a data center somewhere. And this is definitely the direction of the industry.

Recently in the OpenSolaris discussions list, the topic of UNIX workstations came up and how folks miss having a good SPARC workstation. I replied to this thread and gave my observations over the past 10 years. I think it's a good read and have included it below for you to enjoy!

I would argue that Sun lost focus on the workstation market around 2000. While it had some great workstations (Ultra 60, Ultra 80, Sun Blade 1000/2000), it was pretty clear that the focus for Sun was on building mid-range to high-end servers (Ultra Enterprise 4x00-6x000 and E10k). And in that space, pushing the Mhz up would have cost Sun on the cooling and power envolope. And with SMP on more and more sockets, Mhz wasn't a problem back then. Kinda similar to how things have been on the CMT processors until recently. Solaris was optimize to scale up and as a result, its performance on a single socket suffered. Not really a surprise if you think about it. Sun from that point on was making the lion's share from servers and not workstations anymore.

And to add to the situation, the release of the Sun Rays really helped to kill off the workstation line. Sun couldn't make a 1-2 CPU box that out performed an x86 box and Sun wanted everything in the data center anyways. Not to mention that the graphics cards, while very capable, were getting more and more expensive. When I left Sun almost 10 years ago, we were starting to deploy Sun Rays and taking everyones SPARC workstations away. In typical Sun fashion, we were eating our own dog food and making Sun Rays work right. This was the second time that Sun tried internally to deploy thin clients, remember the Java Station? But the good news is that Sun Rays worked great and the deployment worked. This led to major changes on the thinking inside Sun..

1. Pushed all the desktop workloads to the data center and enabled better remote access and work for field offices and telecommuters.
2. Reduced overhead in desktop support. I heard that it now takes less than 30 Sysadmins to manage the 30k+ Sun Rays at Sun.
3. It reinforced the vision of the data center handling everything and everyone using stateless devices (gee we're revisiting that again with mobile devices, lol)
4. It put into question the value of workstations beyond kernel and hardware engineers.

Sun Rays are still a great product and you can do a lot on them. Sun, the US Gov, lots of service desk orgs, and etc use them. I've even deployed them inside of two banks with great success.

Now on top of all this internal stuff at Sun.. you have to look at how the climate changed in the field for workstations at big organizations. There was a time when you'd walk in the Dilbert cube mazes of companies and find folks with a UNIX workstation and a PC desktop or maybe a Mac. Fast forward to about 5 years ago and most corporations decided to start out-sourcing or to downsize desktop support. As a result, they all pushed for standardized desktops and guess what? They are all windows desktops and most companies implemented a policy that if you want to use UNIX, you have to log into a box in the data center for that. So UNIX workstations, Macs, and even Linux were pushed out of the office space and back into the data center because of office policies.

I have yet to see this change, other than some folks using a Mac laptop here and there like me:) But ultimately, this is what really killed the UNIX workstation market. Corporations standardizing on Windows PC's and making it *illegal* to have anything else. And this affected not only Sun, but also IBM and HP. The UNIX workstation market just dried up. And this is also what has killed Linux as a desktop alternative.

Really at this point, the only alternative desktop that's making any progress is the Mac. The sales are up and everyone is buying iPhones and iPads. I think it's the only real chance to battle back against Windows and it's working better than even Apple thought possible.

So where does that leave folks like us who want a UNIX workstation? Up the creek, that's where! I'd love to have a quad-core CMT workstation with great 3D graphics. But it'll never happen at this point. I have an Ultra 20 and a Sun Blade 2000 that I use as servers and test gear at home. In the future, it'll probably be a Mac Pro and lots of VirtualBox instances. I don't know.

It is truly amazing to think back over the past 10 years and see how many of the ideas that were appearing inside Sun are now going mainstream. I'm very proud to have worked at Sun "back in the day" as it were and to have played my role in it. I think the days of large data centers, thin clients, and mobile devices in all shapes and sizes is indeed upon us.

Great Book on Solaris Virtualization

2010-11-01T23:17:00.000-05:00

While I was at the Oracle OpenWorld 2010 conference, I picked up a new book called "Oracle Solaris 10 System Virtualization Essentials" which is written by Jeff Victor, Jeff Savit, Gary Combs, Simon Hayler, and Bob Netherton. This is an excellent book for systems architects, engineers, and administrators to become familiar with all of the virtualization technologies for the Solaris platform. The book covers Dynamic Domains, Oracle VM Server for SPARC (LDoms), Solaris 10 x86 Virtualization (OVM for x86, xVM Server, Xen, VMware, and Hyper-V), Solaris Containers, and Virtualbox. The book includes an excellent chapters on choosing the right virtualization technology, applying virtualization, and management. This is definitely an excellent book for folks to become familiar and get started on learning these technologies. It is amazing to think there are soo many options for virtualizing Solaris these days. I've worked extensively with Dynamic Domains, LDoms, Solaris Containers, and VirtualBox. It's great to finally see a book that can introduce one to all of the options available and give you an insight in how to apply them.

I highly recommend this book! You can pick up a copy from Amazon or other book stores. Definitely want to thank the authors for taking the time to write this book!

Oracle VM Server for SPARC 2.0

2010-11-01T22:47:00.000-05:00

The Oracle VM Server 2.0 software was released during the week of Oracle OpenWorld 2010 in San Francisco. At that conference, I presented with John Falkenthal and Honglin Su on the 2.0 release. I've uploaded the presentations to the LDoms Community site:

Oracle VM Server for SPARC: Enabling a Flexible and Efficient IT Infrastructure, by John Falkenthal, Honglin Su, and Octave Orgeron, presented at Oracle OpenWorld 2010. (PPT)
Industry View: Oracle VM Server for SPARC 2.0, by Octave Orgeron, presented at Oracle OpenWorld 2010. (PDF)

The first presentation goes into detail over the 2.0 release while the second presentation gives an industry view. Unfortunately, the audio for the session has not been made available for download and contained valuable information. As such, I'll provide a brief overview of the new features in the 2.0 release:

SPARC-T3 Support - The SPARC T3 servers (1, 1b, 2, and 4) are supported with this release.
PCI-E Direct I/O - PCI-E slots can now be allocated directly to LDoms.
Dynamic Memory Reconfiguration - Memory can now be dynamically allocated to LDoms. This concludes DR capable resources that are available (CPU, Cryptographic Units, Virtual Disks, and Virtual NICs)
LDom Migration Enhancements - Increased performance and fewer restrictions around CPUs and Cryptographic Units
Enhanced Virtual Disk Multi-Pathing - Multi-Pathing at the Guest LDom level when multiple Service Domains are used as back-ends.
Unique Domain ID Tools & APIs - Unique Domain IDs for identification and asset tracking of LDoms.
Enhanced CPU Allocation & Affinity - Better allocation of CPU threads to prevent cache thrashing
SPARC-T3 Power Management - Power management features on the SPARC T3 servers where unused threads and cores can be powered off.
P2V Conversion Tool Enhancements - Enhanced support for UFS and VxVM boot volumes.

Over the next few blog posts, I'll demonstrate these new features where possible, so that you can see the benefits of using the 2.0 release.

The OVM for SPARC 2.0 download site is ready and the updated firmware is now available. It is important to update the firmware to support the 2.0 release features. I've also updated the LDoms Community site with the links to the 2.0 documentation.

SPARC-T3 Servers Announced!

2010-09-28T22:17:00.000-05:00

Last week while at Oracle OpenWorld, the SPARC-T3 servers were announced! These are the third generation of CMT servers to be released. Three rack mount servers and a blade module were announced. These are all based on what was going to be called the UltraSPARC-T3 processor, but what is now called the SPARC T3 processor. The SPARC T3 processor has the following specifications:

16 Cores x 8 Threads = 128 Threads Running at 1.65Ghz!
2 x Execution Units per Core with 4 x Threads Each!
1 x Floating Unit and 1 x Crypto Unit per Core!
6MB of Shared L2 Cache
2 x DDR3 Memory Controllers
6 x Coherency Links for Glue-Less SMP upto 4 Sockets!
2 x 10GbE NIU Controllers
2 x PCI-E 2.0 Controllers -> 8GB/s Bi-Directional I/O Each!

Each server bares this as the model name with a single number following it representing the number of sockets. Here are the high-level specifications of the SPARC T3 servers:

One of the big changes with this generation is the removal of the 1 rack unit servers as they were being outsold by the 2 rack unit and above servers. This makes sense as most clients want the extra PCI-E slots and internal storage capacity. The T3-1B is meant for the Sun Fire 6000 Blade chassis and follows the standard form factor there. The T3-1 server re-uses the chassis for the T5240 servers. The T3-2 and T3-4 servers re-use the chassis from the Sun Fire X-series x64 servers. The most impressive change is the T3-4 server which uses the PCI-E PEM modules which enable hot-swapping of the PCI-E cards. The T3-4 server provides 16 PCI-E PEM slots to work with, which is more I/O slots than the x4800 server the chassis is based on! These servers are definitely designed to help customers take advantage of virtualization and consolidation with Solaris Containers and Oracle VM Server for SPARC (LDoms)!

While walking around at Oracle OpenWorld, I had a chance to take pictures of these new servers:

Some of the interesting features of the SPARC T3 servers that most people at the conference probably missed are:

Integrated Graphics Console (15-pin VGA)
RFID Tags on Face Plate
Enhanced Power Management (Even at the CPU and Memory Level)
More Hot-Swappable Components!

Already Oracle has released benchmarks using its middle-ware and database products to compare the SPARC T3 servers against Intel Nehalems and IBM POWER7 servers! This is really exciting as for the first time there are some real world benchmarks out at the server announcement release. This will go a long way to showing that SPARC can still pack a punch and compete even against IBM POWER7 servers!

In the next post, I'll talk about Oracle VM Server for SPARC 2.0!

Presenting at Oracle OpenWorld!

2010-09-19T11:44:00.000-05:00

I'll be presenting as a guest speaker at Oracle OpenWorld! Here are the details:

Oracle VM Server for SPARC: Enabling a Flexible, Efficient IT Infrastructure
Wednesday, September 22, 13:00 | Moscone South, Rm 301

Oracle VM Server for SPARC (previously called Sun Logical Domains) provides highly efficient, enterprise-class virtualization capabilities for Oracle's Sun SPARC Enterprise servers with chip multithreading (CMT) technology. Oracle VM Server for SPARC leverages the built-in SPARC hypervisor to subdivide supported platforms' resources by creating partitions called logical or virtual domains, provides the flexibility to deploy multiple Oracle Solaris operating systems simultaneously on a single platform, and enables you to create as many as 128 virtual servers on one system to take advantage of the massive thread scale offered by the CMT architecture. Learn more in this session.

This will be a great chance to meet enthusiasts, engineers, architects, system admins, etc. that are interested in LDoms. This is also a great chance to learn about the upcoming features!

See you there!

Oracle Open World is Next Week! See you there!

2010-09-13T22:41:00.000-05:00

Oracle Open World 2010 is next week, Sept. 19-23!

This is the conference to attend this year for developers, system admins, engineers, architects, etc. for Sun and Solaris technologies. I will be at this conference next week and looking forward to the exciting sessions on cloud computing, virtualization, Solaris, hardware, etc. I'll be posting pictures and impressions from the conference as the week unfolds. It's also a chance for me to visit my old city San Francisco!

I hope to run into some of you there at the conference and have a few drinks! See you there!

Cross Roads for OpenSolaris

2010-09-13T22:23:00.000-05:00

As many OpenSolaris enthusiasts and users have learned, Oracle has changed the game for OpenSolaris. The tap as it were has been turned off and we are no longer getting code drops and updates! Oracle has decided to drop the OpenSolaris binary distribution in favor of a new Solaris 11 Express program that will be announced later in the year. Oracle will only do a code drop onto the OpenSolaris.org repository after an official Solaris update release. This means that we won't see any code drops until Solaris 11 goes GA and when there are follow-on updates. This definitely cripples things for users and developers of the OpenSolaris distribution until Solaris 11 Express comes out. We can only hope that Solaris 11 will be a cut of OpenSolaris with IPS and more of the non-open sourced components put back in.

In the mean time, there have been several developments. First off, several of the key Solaris developers have left Oracle to work for the likes of Nexenta, Joyent, etc. There is a new OpenSolaris based community call Illumos that has been created and will continue on the good work of replacing the closed source bits of OpenSolaris with new code that will be open sourced. This is being led by Garret D'Amore who is now working at Nexenta. The Illumos project will allow OpenSolaris based distributions to use a common source base for the ON core. I'm very excited about this project and look forward to joining in. This may provide the base of my own distribution if I ever get around to doing something like that:)

Another development is that there will be another distribution starting called OpenIndiana which will be announced tomorrow! Needless to say, the number of OpenSolaris based distributions is increasing! These are definitely exciting times and I expect this trend to continue.

Sun/Oracle ZFS and NetApp Patent Dispute Over!

2010-09-13T22:08:00.000-05:00

Finally Oracle has put this issue to rest by settling with NetApp over the ZFS vs WAFL patent dispute. From what I remember this all started with STK going after NetApp over other patents, and then NetApp going after Sun over ZFS. Sun went after NetApp for Sun IP being used in NetApp products without any license. And needless to say this all became crazy and stretched on for a while. The last I heard was that Sun was able to show that there is indeed "prior" art on the whole "copy on write" and "write anywhere file layout" concept and that if anything NetApp could loose its own patents. Enter Oracle, and things go silent. It may be some time before the public finds out the details. But if I were to make a prediction, it is that Oracle got its legal army ready to take NetApp down and NetApp backed down and decided to settle.

So the good news is that all the FUD around ZFS is over! With ZFS being open sourced through OpenSolaris, it should be safe to assume that users of the source code are indemnified and protected. As a result, Coraid, Compellent, GreenBytes, Nexenta, etc. can all go happily back into the market place and sell their products without worry. This is great news! As anyone who has really delved into using ZFS can attest, it really is a killer product. Seeing how Oracle has dropped its HDS agreement and is using the 7000 series OpenSolaris ZFS based arrays in its ExaByte 2 product, that Oracle has big plans for ZFS. The need for expensive EMC, HDS, NetApp, etc products may be coming to a close. All of those vendors have been placing greater and greater focus on modular commodity storage. With out the custom ASICs and firmware, there isn't much to prevent anyone from turning a server with off the shelf storage and OpenSolaris ZFS into a SAN/NAS/iSCSI solution.

Oracle Solaris 10 09/10 (Update 9) Released!!

2010-09-13T20:21:00.000-05:00

Oracle has released the latest update for Solaris 10 09/10! It is great to see that the Solaris 10 updates and patches are alive and kicking! I know there was some skepticism from many as to Oracle's intentions for Solaris and all I have to say is WAKE UP! Solaris is alive and well!

According to the What's New documentation, the following new features are available:

Auto-Registration (For interactive installs and can be turned off)
Install-Time Update (ITU) Tools Support for SPARC
Solaris Zones Cluster Node Update (Via DVD or Jumpstart)
Solaris 10 P2V Tool for Containers
Solaris Containers HostID Emulation
Solaris Containers zoneadm attach -U option for updating packages
Oracle VM Server for SPARC (LDoms) 2.0 Dynamic Memory Reconfiguration Support
Oracle VM Server for SPARC (LDoms) 2.0 Virtual Disk Multi-Pathing between Service Domains
Oracle VM Server for SPARC (LDoms) 2.0 Static Direct I/O Support for PCI-E End-Point Devices for Guest Domains
Oracel VM Server for SPARC (LDoms) 2.0 Virtual Domain Information (virtinfo) Command and API
ZFS Device Replacement Auto-Expand Feature
ZFS "zpool list" out-put enhancement "USED,AVAIL" replaced with "ALLOC, FREE"
ZFS "keep" tag for snapshots with the "zfs hold" command to prevent ZFS receive snapshots from being over-written by source changes
ZFS Triple Parity (raidz3) Support
ZFS intent-log enhancements for performance and device replacement and removal
ZFS pool recovery enhancements
ZFS I/O process threads visible for performance metrics per ZFS pool in the global zone
ZFS poll split support
Fast Crash Dump - 2-10x performance increase in crash dumps with compression
Intel Xeon 5500 Energy Performance MSR support
Modern Disk Sector Size Support (512, 1024, 2048, and 4096 byte support) for newer hard drives
iSCSI Initiator tunables for network response times
Sparse Files Support for cpio command
x86 64-bit libc string functions with SIMD SSE instruction acceleration
Sendmail auto-rebuild feature for sendmail.cf and submit.mc
Automatic Boot Archive Recover support
net_access privilege now part of the "basic" privilege set
x86 Intel AES-NI cryptographic support on Xeon 5600 processors
New Language Support (South Africa, English Singapore, Chinese Singapore, Malaysia, Indonesia, Bengali India, English India, Gujarati India, Kannada India, Marathi India, Telugu India, Tamil India)
Solaris iSCSI Boot Support (Already present in SPARC OBP and some x86 iLOMs)
iSCSI Acceleration over RDMA (iSER)
New Hot-Plug Support for PCI-E and PCI SHPC components with the cfgadm command
Adaptec RAID AAC power management support
x86 HP Smart Array Raid Controller HBA support
x86 Broadcom NetXtreme II 10GbE GLDv3 Support in bnxe driver
x86 Broadcom HT1000 SATA Controller Support in bcm_sata driver
x86 AMD SB700/750 and nVidia nForce 780a Chipset Support
GNU Tar 1.23
FireFox 3.5
Thunderbird 3
Less 436
Bind 9.6.1
GLDv3 Driver API Framework
IP Over Infiniband (IPoIB) Performance Enhancements
Infiniband OFED RDMA CM Kernel KPI Support
Infiniband Performance Enhancements
X11 Xorg setxkbmap dynamic keyboard switching support
Intel Shared Code ixgbe support
Broadcom NIC support for 5718, 5764, 5785 chipsets in bge driver
x86 Fully Buffered DIMM Idle Power Management on Intel 5000/7000 Memory Controller Hubs (MCH) for Sun Fire x4450, x4150, x6250 servers
Solaris FMA support for AMD Instanbul Based Servers
Solaris FMA support for Netra CP3250 Blade DDR2 DIMMS
Solaris VTS 7.0ps9 Updates and Enhancements
Solaris mdb debugging enhancements for kmem and libumem

The Solaris documentation is available in the normal place.

Some of the interesting things you may have picked up on the list above of new features is the Oracle VM Server for SPARC v2.0 items. As you've probably guessed, LDoms is now called Oracle VM Server for SPARC. I'll always call it LDoms. If you click on the links in the "What's New" list, you'll see that the links do not work yet. However, this clearly shows that LDoms 2.0 is literally around the corner! The LDoms support features that are listed are things that I've seen pop up in OpenSolaris, not mention a ton of the other features that have been back-ported into Solaris 10. Needless to say, I'm very excited about LDoms 2.0 coming out and can't wait to get my hands on it. There is a session next week at the Oracle Open World conference that will talk about the next generation T-Series servers and another one on LDoms being presented by Honglin Su and John Falkenthal. Can't wait to see what will be presented there!

One of the things I did after downloading Solaris 10 09/10 was to look at the sun4v support to see if the UltraSPARC-T3 servers have turned up. Sadly, I don't see them in /usr/platform. I do see the following models:

SUNW,Netra-CP3060
SUNW,Netra-CP3260
SUNW,Netra-T2000=SUNW,Sun-Fire-T200
SUNW,Netra-T5220
SUNW,Netra-T5440
SUNW,SPARC-Enterprise-T1000=SUNW,Sun-Fire-T200
SUNW,SPARC-Enterprise-T2000=SUNW,Sun-Fire-T200
SUNW,SPARC-Enterprise-T3120=SUNW,SPARC-Enterprise-T5120
SUNW,SPARC-Enterprise-T5120
SUNW,SPARC-Enterprise-T5220=SUNW,SPARC-Enterprise-T5120
SUNW,Sun-Blade-T6300
SUNW,Sun-Blade-T6320
SUNW,Sun-Blade-T6340=SUNW,T5140
SUNW,Sun-Fire-T1000=SUNW,Sun-Fire-T200
SUNW,Sun-Fire-T200
SUNW,T5140
SUNW,T5240=SUNW,T5140
SUNW,T5440=SUNW,T5140
SUNW,USBRDT-5240
SUNW,USBRDT-5440

The mysterious T3120 is in the list, but I haven't heard any news on if it'll ever make it to market. I totally expect that the UltraSPARC-T3 servers will be announced next week and start shipping before the end of the year. I can't wait to see one and have a chance to take it for a spin!

One of the most heated debates that has emerged since the release of Solaris 10 09/10 is the change in the license. There is a lot of contention over it the following:

Except for any included software package or file that is licensed to you by Oracle under different license terms, we grant you a perpetual (unless terminated as provided in this agreement), nonexclusive, nontransferable, limited License to use the Programs only for the purpose of developing, testing, prototyping and demonstrating your applications, and not for any other purpose.
All rights not expressly granted above are hereby reserved. If you want to use the Programs for any purpose other than as permitted under this agreement, including but not limited to distribution of the Programs or any use of the Programs for your internal business purposes (other than developing, testing, prototyping and demonstrating your applications) or for any commercial production purposes, you must obtain a valid license permitting such use. We may audit your use of the Programs. Program documentation, if available, may be accessed online at http://otn.oracle.com/docs.

The gist here is that Oracle gives you a perpetual free license to use Solaris 10 if you are developing, test, prototyping, or demonstrating applications on it. For most developers, system admins, and engineers this is pretty simple. However, for normal end-users this can present an issue and possible gray area. I'm not sure what the answer here is other than to suggest that the enthusiast community of Solaris, OpenSolaris, etc. users contact Oracle to clarify this issue. Reading the rest of the license pretty much states that if you are using Solaris for production purposes, you have to buy a license which makes sense. From a business perspective, Oracle is giving you a free license ride for your development, testing, and QA servers and pushing the costs to your production servers. Typically, there are more non-production servers for most companies. The support costs for Solaris have also changed. Instead of the ~$240 a socket that it use to be, you now pay 8% of the net (discounted where possible) cost of the server per year. While I understand this may throw a wrench into the mix, Solaris support use to be cheap and Sun wasn't making enough money on it. So this is just Oracle being Oracle and turning something that is otherwise a cost center into a revenue stream. IDEAS has done an interesting write up on the Solaris support changes, an interesting read.

One last note on the license is that it's for Oracle Solaris, Oracle Solaris Express, and Oracle Solaris Cluster!! So this is the license that Solaris 11 Express will be using.

Internal Oracle Memo Leaked on Solaris and OpenSolaris

2010-08-15T16:02:00.001-05:00

On August 13th, an internal memo to the Solaris Engineers and Managers at Oracle was leaked out onto the OpenSolaris mailing lists:

Solaris Engineering,

Today we are announcing a set of decisions regarding the path to Solaris 11, and answering key pending questions on open source, open development, software and binary licenses, and how developers and early adopters will be able to use Solaris 11 technology before its release in 2011.

As you all know, the term “OpenSolaris” has been used colloquially to refer to any or all of a collection of source code, a development model, a web site, a logo, a binary release, a source license, a community, and many other related things. So it’s taken a while to go over each issue from an organizational and business perspective, and align on the correct next step. Therefore, please take the time to read all of the detail here carefully. We’ll discuss our strategy first, and then the decisions and changes to our policies and processes that implement that strategy.

Solaris Strategy
———————-

Solaris is the #1 Enterprise Operating System. We have the leading share of business applications on Solaris today, including both SPARC and x64. We have more than twice the application base of AIX and HP-UX combined. We have a brand that stands for innovation, quality, security, and trust, built on our 20-year investment in Solaris operating system engineering.

From a business perspective, the purpose of our investment in Solaris engineering is to drive our overall server business, including both SPARC and x64, and to drive business advantages resulting from integration of multiple components in the Oracle portfolio. This includes combining our servers with our storage, our servers with our switches, Oracle applications with Solaris, and the effectiveness of the service experience resulting from these combinations. All together, Solaris drives aggregate business measured in many billions of dollars, with significant growth potential.

We are increasing investment in Solaris, including hiring operating system expertise from throughout the industry, as a sign of our commitment to these goals. Solaris is not something we outsource to others, it is not the assembly of someone else’s technology, and it is not a sustaining-only product. We expect the top operating systems engineers in the industry, i.e. all of you, to be creating and delivering innovations that continue to make Solaris unique, differentiated, and valuable to our customers, and a unique asset of our business.

Solaris must stand alone as a best-of-breed technology for Oracle’s enterprise customers. We want all of them to think “If this has to work, then it runs on Solaris.” That’s the Solaris brand. That is where our scalability to more than a few sockets of CPU and gigabytes of DRAM matters. That is why we reliably deliver millions of IOPS of storage, networking, and Infiniband. That is why we have unique properties around file and data management, security and namespace isolation, fault management, and observability. And we also want our customers to know that Solaris is and continues to be a source of new ideas and new technologies– ones that simplify their business and optimize their applications. That’s what made Solaris 10 the most innovative operating system release ever. And that is the same focus that will drive a new set of innovations in Solaris 11.

For Solaris to stand alone as the best-of-breed operating system in Oracle’s complete and open portfolio, it must run well on other server hardware and execute everyone’s applications, while delivering unique optimizations for our hardware and our applications. That is the central value proposition of Oracle’s complete, open, and integrated strategy. And these are complementary and not contradictory goals that we will achieve through proper design and engineering.

The growth opportunity for Solaris has never been greater. As one example, Solaris is used by about 40% of Oracle’s enterprise customers, which means we have a 60% growth opportunity in our top customers alone. In absolute numbers, there are 130,000 Oracle customers in North America alone who don’t use our servers and storage yet, and a global customer base of 350,000 (the prior Sun base was ~35,000). That’s a huge opportunity we can go attack as a combined company that will increase Solaris adoption and the overall Hardware server revenue. Our success will also increase the amount of effort ISVs exert optimizing their applications for Solaris.

We will continue to grow a vibrant developer and system administrator community for Solaris. Delivery of binary releases, delivery of APIs in source or binary form, delivery of open source code, delivery of technical documentation, and engineering of upstream contributions to common industry technologies (such as Apache, Perl, OFED, and many, many others) will be part of that activity. But we will also make specific decisions about why and when we do those things, following two core principles: (1) We can’t do everything. The limiting factor is our engineering bandwidth measured in people and time. So we have to ensure our top priority is driving delivery of the #1 Enterprise Operating System, Solaris 11, to grow our systems business; and (2) We want the adoption of our technology and intellectual property to accelerate our overall goals, yet not permit competitors to derive business advantage (or FUD) from our innovations before we do.

We are using our investment in core Solaris innovation and engineering to drive multiple businesses, through multiple product lines. This already includes our Solaris operating system for Enterprise, and our ZFS Storage product line, and will soon include other Oracle products. This strategy is all about creating more value from a set of common software investments: it makes everything you do more valuable and used by more people worldwide. It also means you as an individual engineer or manager have an even greater responsibility to understand the broader business and technical contexts in which your engineering is deployed.

Solaris Decisions
————————

We will continue to use the CDDL license statement in nearly all Solaris source code files. We will not remove the CDDL from any files in Solaris to which it already applies, and new source code files that are created will follow the current policy regarding applying the CDDL (simply, that usr/src files will have the CDDL, and the very small minority of files in usr/closed might not have it). Use of other open licenses in non-ON consolidations (e.g. GPL in the Desktop area) will also continue. As before, requests to change the license associated with source code are case-by-case decisions.

We will distribute updates to approved CDDL or other open source- licensed code following full releases of our enterprise Solaris operating system. In this manner, new technology innovations will show up in our releases before anywhere else. We will no longer distribute source code for the entirety of the Solaris operating system in real-time while it is developed, on a nightly basis.

Anyone who is consuming Solaris code using the CDDL, whether in pieces or as a part of the OpenSolaris source distribution or a derivative thereof, would therefore be able to consume any updates we release at that time, under the terms of the CDDL, LGPL, or whatever license applies.

We will have a technology partner program to permit our industry partners full access to the in-development Solaris source code through the Oracle Technology Network (OTN). This will include both early access to code and binaries, as well as contributions to us where that is appropriate. All such partnerships will be evaluated on a case-by-case basis, but certainly our core, existing technology partnerships, such as the one with Intel, are examples of valued participation.

We will encourage and listen to any and all license requests for Solaris technology, either in part or in whole. All such requests will be evaluated on a case-by-case basis, but we believe there are many complementary areas where new partnership opportunities exist to expand use of our IP.

We will continue active open development, including upstream contributions, in specific areas that accelerate our overall Solaris goals. Examples include our activities around Gnome and X11, IPS packaging, and our work to optimize ecosystems like Apache, OpenSSL, and Perl on Solaris.

We will deliver technical design information, in the form of documentation, design documents, and source code descriptions, through our OTN presence for Solaris. We will no longer post advance technical descriptions of every single ARC case by default, indicating what technical innovations might be present in future Solaris releases. We can at any time make a specific decision to post advance technical information for any project, when it serves a particular useful need to do so.

We will have a Solaris 11 binary distribution, called Solaris 11 Express, that will have a free developer RTU license, and an optional support plan. Solaris 11 Express will debut by the end of this calendar year, and we will issue updates to it, leading to the full release of Solaris 11 in 2011.

All of Oracle’s efforts on binary distributions of Solaris technology will be focused on Solaris 11. We will not release any other binary distributions, such as nightly or bi-weekly builds of Solaris binaries, or an OpenSolaris 2010.05 or later distribution. We will determine a simple, cost-effective means of getting enterprise users of prior OpenSolaris binary releases to migrate to S11 Express.

We will have a Solaris 11 Platinum Customer Program, including direct engineering involvement and feedback, for customers using our Solaris 11 technology. We will be asking all of you to participate in this endeavor, bringing with us the benefit of previous Sun Platinum programs, while utilizing the much larger megaphone that is available to us now as a combined company.

We look forward to everyone’s continued work on Solaris 11. Our goal is simply to make it the best and most important release of Solaris ever.

-Mike Shapiro, Bill Nesheim, Chris Armes

While there are some positives about Oracle investing more heavily in Solaris and getting Solaris 11 ready for release next year, there are some down sides to this memo. It clears states the following:

OpenSolaris Distribution Ends!
Restricted ARC cases and discussions on OpenSolaris.org
Oracle will re-instate the old Solaris Express program for binary only pre-views of Solaris 11.
OpenSolaris code drops will only happen after official Solaris releases, making Solaris Express the only place for non-Oracle employees to pre-view new Solaris features.
The CDDL License will continue, but closed components in /usr/closed will continue as well

This leak has caused havoc in the OpenSolaris community, especially after the months of silence from Oracle. Last month, I attended a OpenSolaris Governance Board meeting on the phone to discuss this very situation and worries over the future of the community. At the time the OGB decided to give an ultimatum to Oracle to answer our questions and concerns or the OGB would dissolve itself and turn control back over to Oracle. In light of what has happened, this is inevitable.

However, there are some truths that we as a community must accept. First off, while there has been extraordinary support from developers, users, system administrators, and even OpenSolaris based distributions, there was limited code development outside of Sun or Oracle. No matter what, we are completely dependent upon Oracle for releasing code, doing fixes, and handling all the bug reports. For those in other open source projects such as Linux and the *BSD variants, this should be no surprise. Even in those communities, the number of contributions outside of their core kernel developers is very small. The big difference here is that those core developers are employed by Oracle to work on Solaris. This is not a bad thing, as that core has been the driving strength for innovation and quality in the Solaris product. In the OpenSolaris community, there were higher hopes that developers would clamor around to work on bugs and code enhancements. Unfortunately, this didn't happen. I strongly believe this had a lot to do with the difficulty in understanding where to begin and how to jump through all the hoops to get code committed. If the community would have donated a lot of code back and worked on fixing tons of bugs, I would think that Oracle would feel differently on this. But seeing how all Oracle was getting was a lot of beta testers for new features and having the code sit out there for the competition to reverse-engineer, the value to paying customers and share holders is very low. It was great for all of us who are vested in OpenSolaris and in Sun technology as a whole.

Ultimately, this leaves the OpenSolaris community in a bind and without strong leadership to turn this around somehow. So instead, some brave folks in the community have pulled together to continue the work we all wanted to see with the OpenSolaris community by creating a fork of the code and to replace all the closed bits that were left. I'd like give a big thanks to Garrett D'Amore for starting the Illumos Project for doing what the community soo desperately needed to do from the start. I strongly encourage the OpenSolaris community to hold things together, but to support the Illumos project.

Being a community leader in the SysAdmin and LDoms communities, this has affected me deeply as I'm unsure about the future directions of these communities. But pledge to continue forth with leading these communities and to supporting the OpenSolaris movement.

Everyone hang in there! We can keep our OpenSolaris movement and community together. We live in interesting times and it's up to us to shape the future. Let us all be thankful that the source code for Solaris was open sourced and now we can take it to the next level.

Oracle Speaks

2010-06-09T23:23:00.000-05:00

While many Sun customers are up in arms over the silence over at Oracle, there have been a few blurbs here and there as to the future of the products we depend upon. Here are two good resources:

These atleast give some idea that Oracle is investing in Solaris, OpenSolaris, and a huge amount of Sun's software portfolio. This should ease some tensions, but Oracle definitely has to spend more effort to reassure customers. I am very happy to see commercials on TV again for Sun products:

Virtualization Product Landscape

2010-06-09T23:14:00.000-05:00

The virtualization space has been growing by leaps and bounds. It has attracted vendors from all over the map to the party. Not to mention the evolution of virtualization into cloud computing. There has been an explosion of software products for creating, brokering, monitoring, securing, and managing virtual infrastructures. Keeping track of all of these developments is challenging to say the very least. However, I have found a good reference table that I wanted to share with everyone. Enjoy!

Brocade and Juniper Gang Up on Cisco

2010-06-09T21:46:00.000-05:00

Over the past year, there has been a lot of buzz in the industry over Cisco with its UCS and Nexus 7000 products providing a converged fabric for virtualization, consolidation, and cloud computing needs. However, the problem with the Cisco party is that you're stuck with their framework and products. And while UCS is impressive, it doesn't address the need for 4-8 way socket servers, high performance I/O, or traditional I/O needs beyond networking and storage. And you may say, what kind of server would need that? How about a Netbackup media server, payment processing server with crypto cards, or a data warehouse? UCS is really targeted for general purpose usage. Many companies are finding that while blade servers have their place, rackmount servers have greater flexibility and less vendor lock-in. To help aid companies in this effort and still get the benefits of converged networks, Brocade and Juniper have jumped into the ring with their own converged networking products. Both products have different architectures, protocols, and features that will change the way networks are designed and used. Brocade seems to have a really innovative solution with its VCS architecture, TRILL protocol, QoS capabilities, and VM profiling.

These types of products will be the foundation of cloud computing by enabling low latency, scalable, and agile infrastructure for virtual machines, storage, and networking services. Would definitely like to test out these products with LDoms, Containers, and of course VMware.

Oracle and Fujitsu Signing up for APL2?

2010-06-09T21:11:00.000-05:00

It would appear that Fujitsu and Oracle are going to renew the APL agreement for the next generation servers which should be based on the eight-core SPARC64 VIIIfx processor. This is the processor that is being used by the Japanese government to build a supercomputer. I wrote about this about a year ago and have been waiting to see if these processors would see the light of day in a commercial setting. It is interesting how the Sun and Fujitsu agreement back in 2004 lead to the M-Series servers, which have filled the gap left in the mid-range and high-end servers from Sun after the venerable UltraSPARC-IV+ servers were EOL'd. With an eight-core base, the next generation M-Series servers should be pretty impressive. I am curious if there will be over-lap between the T3 servers and these future M-series? Perhaps the T3 and T4 series will cover entry level up to something like an M5000 and the SPARC64 VIIIfx will cover the high-end. Only the future will tell.

More Layoffs? Really?

2010-06-09T21:03:00.001-05:00

Well needless to say there are more lay-offs happening for freshly converted ex-Sun employees. I'm not surprised considering the size of the acquisition. But I do hope that these employees will either find their way back to Oracle or another company soon. These have been difficult economic times and hope the best for them.

UltraSPARC-T3 vs. Power 7

2010-06-09T20:39:00.000-05:00

While we await the release of the UltraSPARC-T3 servers from Sun Oracle, the big question is how does it stack up against IBM's Power 7 processor?

Ars Technica took an interesting look at the feature set of the processors themselves in this article. One of the best lines in the article describing the differences between these processors is, "If Niagara 3 is an army of guys with shovels, POWER7 is a giant bulldozer". The T3 processor definitely takes prize for the largest amount of cores, threads, and I/O options integrated(Crypto, PCI-E, and 10GbE). However, the Power 7 processor makes up for this with heavy cores, higher clock speeds, an on-chip 32MB L3 cache, and large I/O and memory pipes. In many ways you can see how IBM learned a lot from Sun on building multi-core designs. One could even argue that IBM took the idea of a 32MB cache right from the UltraSPARC-IV+ playbook for addressing latency issues for heavy cores. And to think a few years ago IBM was laughing at Sun for the T1 processor saying that CMT wasn't relevant. Well always nice to see someone eat their own words and invest millions!

Hopefully we'll see the UltraSPARC-T3 servers appear soon before long to compete with IBM and show that Sun Oracle can still pack a bunch. Having a 1-4 socket server with UltraSPARC-T3 processors could lead to a new wave of ultra-dense SPARC servers for web, database, and throughput processing. When combined with LDoms and many of the networking and storage virtualization advancements in OpenSolaris, one has a killer product for new projects and consolidations.

The State of Things

2010-06-09T19:56:00.000-05:00

Well the OpenSolaris OGB elections went well, but I did not make it onto the board this year. I'll try again next time. However, seeing how difficult things have been in the OpenSolaris community it can't be an easy time for the OGB. With the acquisition by Oracle, users, customers, and enthusiasts alike have been struggling to understand what the direction is at Oracle across the board for Sun products and even the communities. There was supposed to be another distribution release of OpenSolaris 2010.x, but it has not materialized which of course had caused "great confusion and delay". Many community members such as myself and Ben Rockwood have been vocal and cautious at the same time.

However, as Ben as pointed out, OpenSolaris was never fully out in the open and has always relied on the employed engineers at Sun, now Oracle. There are many reasons behind this that originate in politics, internal Sun processes, committees, etc. As a community we are truly left on the outside to take either the source code or the OpenSolaris distribution to use and tinker with. It does take discussions on the mailing lists, through bugs and RFE's, and community coordination with the perspective Oracle developers to cause change. Every time I find something, I'm not happy about, I file a bug or RFE on the OpenSolaris site and keep my eyes on things. Sadly this is as good as it gets right now. But what should it really be like?

One might think that having the code and processes completely opened would solve all of the issues. Unfortunately, one only has to look at Linux and any of the BSD's to see that any changes to the Kernel or core utilities is controlled by a tight-knit group of developers. Some are friendly and others are down right ruthlessly rude to out-siders. In much the same way that the OpenSolaris community struggles to get it's hands deep into the development cycle and sift through the bureaucracy of Sun/Oracle, developers in the Linux and BSD spaces have the same issues. The only real difference is that OpenSolaris development is owned and controlled by a company. The rest of the open source tools, libraries, and applications that all Unices leverage have their own ups and downs for developers and distribution builders alike.

So what is the real issue here in the OpenSolaris community? I would summarize the issues as the following:

Lack of communication on the roadmap and direction for OpenSolaris from Oracle
Lack of community, customer, and end-user involvement in design decisions and standards
Community dependence on an Oracle release of the official OpenSolaris distribution
Difficult processes for developers to follow to contribute
No accountability between the OGB and Oracle

So how should the community address these issues? Well the OGB should be focusing on addressing these issues in the open. That would mean they would have to pressure Oracle to be more involved, open, and to communicate. While the community is indeed very passionate about Solaris and OpenSolaris, I do believe the voice of reason bring things to a level playing field where all the stakeholders can mutually agree on how to address these issues. The OGB should organize a townhall meeting to tackle the issues at hand. Oracle has everything to gain by working with the community. It'll help build credibility, mind-share, and most importantly trust.

For example, if the community and the developers at Oracle were to work on building better processes and even standards for OpenSolaris, it would go a long way to mending the situation at hand. But the first step is for the OGB to act now and push Oracle into the right direction. Oracle has to listen to our community, because many of us are also their customers whether they know it or not.

Nominee for OpenSolaris Governance Board

2010-03-10T19:33:00.000-06:00

Recently I nominated myself for election on the OpenSolaris Governance Board. I'd like to introduce myself to those in the OpenSolaris community that may not be familiar with me and explain why I am running for office.

First off, I started out with AT&T UNIX back in 1994 after years of using MS-DOS, Windows, and MacOS. Needless to say, there was just something magical about UNIX that caught my attention because it's a platform that you can tinker with and learn something new or old about every day. I was unable to have my own UNIX system and settled on Slackware Linux for a while until I was able to buy a used Sun SPARCstation IPX. Eventually as the next few years went by, I had a collection of UNIX systems ranging from my IPX to a DEC Alpha 500au workstation and a PC running Slackware.

I studied and got my Solaris SCSA certification and moved to Silicon Valley to work as a Taos contractor at Sun Microsystems. I was soo happy to become a Sun employee and work with some of the best people on the bleeding edge of technology. I remember building Sun Clusters on E6500s and A5000s. This was before most people even knew what clustering was let alone thought about using it for infrastructure services. It was around this time that I started to become involved with the Solaris 9 beta program, which was exciting as it was the first Solaris version that included tools that I would normally install from Sunfreeware or compile myself. I learned a great deal during my time at Sun. Not long after that I left Sun to move across the country to be with my future wife and eventually start a family.

Since then I have worked for small businesses and large enterprises helping them design and implement Solaris based solutions. Working in financial services and E-Commerce environments has given me incredible insight and experience into how Sun technology can be leveraged to solve complex problems. I always kept my ties back to Sun tight and involved myself in many beta programs covering all sorts of products such as Solaris updates, Sun Rays, Sun Studio, Solaris 10, JDS, etc. During my time at JP Morgan I was able to test the latest and greatest from Sun including T and M series servers and LDoms. When OpenSolaris started, I helped start the SysAdmin community and eventually the LDoms community which I have made significant contributions to.

Today I am an architect helping Fortune 500 companies design virtualization and cloud computing solutions. While this keeps me very busy, I still dedicate time and resources to reaching out and promoting OpenSolaris and Sun technologies. I've seen our community evolve over the years through the many challenges and accomplishments. With the recent acquisition by Oracle, it is important that the community continues to push forward and keep the momentum of OpenSolaris going. I'd like offer my experience and industry perspective to the OGB and help guide OpenSolaris towards success with both the open source community and with the corporate world. My platform is simple:

Push for openness and transparency between the OpenSolaris Community and Oracle Solaris developers.
Close the gap between Solaris and OpenSolaris by reducing the number of closed source binaries.
Push for OpenSolaris to be easier to build from source and for the community to contribute to the code base.
Promote OpenSolaris development interests from schools, universities, and the open source community in large.
Push for enterprise readiness of OpenSolaris and parity between x86 and SPARC builds.
Increase visibility of OpenSolaris to both the enterprise and start-ups.

Please let your voice be heard and vote for a new OGB and for me!

Thank you everyone for the opportunity, I look forward to serving the community!

Sun Oracle

2010-01-29T00:00:00.000-06:00

Finally I had a chance to watch all the webcasts and read through the presentations, which are located here. I can not even begin to describe how excited I was about the event and wanted soo badly to watch it live and ignore work. But alas, that was not meant to be and I had to wait until I got home to check it all out. Needless to say, I'm very pleased with the outcome and was really excited to through all the material. Things that I'm most pleased with are:

New Sun Oracle Logo!
Increased investment in SPARC, Solaris, Servers, Storage, Virtualization, etc.
UltraSPARC-T3 will be delivered this year! Plus there are 3 other SPARC processors in development!!
LDoms takes center stage as the hypervisor on SPARC for Oracle VM
Oracle Enterprise Manager will merge with OpsCenter
Sun Oracle is hiring across the board! Great news for Sun employees who were laid off and for people looking to be part of something new and exciting.

Also, the old www.sun.com site has changed! It is sad to see the old Sun web site go away and merged with the Oracle web site. I will miss the old blues and silvers. Atleast the servers will keep the silver colors and the Sun logo will continue with just a font change and minus the "microsystems" part. It'll also be unusual to say Oracle Solaris now, so that will take some getting use to.

I also noticed that OpenSolaris was not mentioned, so hopefully Oracle will clarify that at some point soon. Personally, if it were up to me, Oracle should dump Linux and invest heavily in completing Solaris 11 and base it on OpenSolaris. But that's just me!

EU Approves Sun Buy-Out by Oracle!

2010-01-22T10:51:00.000-06:00

Finally, the EU has given approval for the acquisition to move forward and complete! Fortunately, Oracle has been given the green light with no additional conditions or terms. This is great news as Sun will not be broken up. In addition to this news is that Larry Elison will make an announcement on Janurary 27 through a webcast on the acquisition and plans for Sun. For the event, the following topics have been highlighted on the Oracle website:

Offer a broad range of products including servers, storage, networking, and software
Integrate all the components–hardware, operating system, database, middleware, and applications–for unmatched performance, reliability, and security
Simplify IT management and reduce system deployment and integration costs
Continue to drive innovation in SPARC, Solaris, the Java platform, and many other technologies

These are all good signs and look forward to more details! Even the market analyst are responding positively to this.

LDoms 1.3 Released!

2010-01-21T08:00:00.000-06:00

What better way to start the new year than with a new release of Logical Domains! The new version 1.3 has many new features:

Support for link-based IPMP. See Using Link-Based IPMP in Logical Domains Virtual Networking in Logical Domains 1.3 Administration Guide.
Support for migration of domains that have cryptographic units. See Migrating Cryptographic Units in an Active Domain in Logical Domains 1.3 Administration Guide.
Support for CPU dynamic resource management. See Using Dynamic Resource Management Policies in Logical Domains 1.3 Administration Guide.
Support for automated migrations. See Performing Automated Migrations in Logical Domains 1.3 Administration Guide.
Support for the hostid and mac-addr properties by the ldm set-domain command. See the ldm(1M) man page.
Support for memory compression that improves migration performance. See Migrating an Active Domain in Logical Domains 1.3 Administration Guide.
Support for autosave XML interfaces.
Support for booting from a disk that is larger than 1 terabyte.

Support for features such as link-based IPMP and dynamic reconfiguration of cryptographic units have been anticipated or a while now and I'm glad they have arrived. However, I'm very excited about the CPU dynamic resource management feature where you can specify when CPU resources can be automatically scaled up or down. This is potentially a very useful feature as it can allow the logical domain manager (LDM) to automatically add or remove CPU resources to meet performance requirements and utilization. The other feature that I'm excited about is the automated migration capability. This brings LDoms closer to some of the commodity virtualization products on the market that can automate these types of tasks already. However, I do hope that Sun will in the future enhance the feature to support an encrypted password store instead of plain text files.

Overall, I'm very excited about this release and look forward to upgrading this weekend! I'll also have to see if the 1.3 release is in IPS for OpenSolaris and available for compiling. If so, I'll post examples on how to install it from IPS and how to compile LDM. Plus any tidbits along the way when I upgrade.

The OpenSolaris LDoms Community site has already been updated with the new documentation links and the download site, so enjoy and happy upgrading!

Happy Thanksgiving!

2009-11-26T16:26:00.001-06:00

Happy Thanksgiving Everyone! I wish the best for everyone and there will be some new blog posts in the coming weeks, I have tons to blog about.

Sun and Oracle Release Fastest OLTP and Data Warehouse Appliance

2009-09-17T13:49:00.000-05:00

On Tuesday, Sun and Oracle announced the Oracle Exadata V2 OLTP and Data Warehouse appliance. This appliance replaces the previous generation Exadata appliance built on HP equipment and uses all Sun hardware. This includes Sun Galaxy x64 servers (X4170 & X4275), Sun FlashFire Storage, and Sun's own fully non-blocking Infiniband switches integrated into Sun cabinets for easy expansion. The end result is an appliance able to blow away the competition completely and utterly on OLTP and Data Warehouse workloads. In the webcast link above, Ellison and Fowler make great jabs at IBM and demonstrate significant savings over over 5 times. Each fully configured cabinet can perform over a million IOPS, ridiculously insane as Fowler said. Without a doubt, this is a strong endorsement of Sun technology and demonstrates how Sun and Oracle will work together going forward. Also think the new Sun/Oracle logo looks great! To think we'll see more in next month when Oracle demonstrates how Sun T5440's and Oracle 11g can beat the pants off of IBM and Power6!

SPARC Roadmap Leaked

2009-09-14T12:41:00.000-05:00

Well the Register has an interesting article that leaks a slide of the SPARC processor roadmap. While most of the info there is not new to those who are customers or pay attention to conferences, it's still a great read and shows how Sun/Oracle has realigned the roadmap without the UltraSPARC-RK "Rock" processor in the mix. Some interesting points from the slide are:

Transition from 40->28nm process thanks to TSMC.
Only minor speed-ups for SPARC64 Jupiter CPU's until 2012.
APL2 will arrive in 2012. Would assume this would be centered around the SPARC64-VIII "Venus" processors.
Rainbow Falls @ 1.67Ghz, 16 cores with 8 threads each, 1-4 socket servers.
Yosemite Falls @ 2.5Ghz, 8 cores with 8 thread each, 1-4 socket servers.
Yellowstone Falls @ 3Ghzs, 4 cores with 8 threads, 4-192 socket servers!
Cascade Falls @ 3Ghzs, 16 cores with 8 threads, 1-8 socket servers.

As you can see, the SPARC64 roadmap is much slower, even though it is using the Venus processor in the Project Keisoku super-computer project for the Japanese government. With APL2 not arriving until 2012, it may not be the primary future focus for Sun/Oracle.

Which brings us to the follow on processors for Rainbow Falls. Looks like we'll see fewer cores with Yosemite Falls and Yellowstone Falls. However, look at Yellowstone Falls, it can scale up to 192 sockets!! This may become the replacement for what ROCK should have been. I doubt that the lessons learned from ROCK will be lost and if anything enhance the CMT line-up. Looks like Sun/Oracle will be able to fill up the volume, mid-range, and perhaps even high-end servers with CMT processors. I would have to assume that single-threaded performance will finally be addressed. And with clock speeds going into the 3Ghz range, we should see some significant performance increases.

Considering the performance benchmarks for Oracle products on CMT, it makes sense that Oracle would place the recently promised increase in SPARC development on CMT. This would give the biggest return on investment as the vast majority of Oracle products today are already multi-threaded and scalable.

Oracles Reveals Intentions for Solaris and SPARC, Watch out IBM!

2009-09-10T01:04:00.000-05:00

Finally! We are starting to see more coming out of Oracle about their intentions for Solaris and SPARC. This re-enforces what Larry Ellison had stated earlier in the year that Oracle would increase the investment in Solaris and SPARC. This is great news to Sun customers and of course bad news for IBM and HP who were hoping to make further gains on spreading FUD. Now we just have to wait for the outcome of the EU investigation into the merger, which is centered around MySQL. Once that is sorted out, I fully expect to see a major shake-down at Sun and a huge increase in staffing and resourcing on key products, which is badly needed. This should help bring more hope to customers and users alike. Solaris and SPARC will continue on!

Flash Storage?

2009-09-01T20:53:00.001-05:00

It would appear that Sun will be announcing a single rack unit flash storage array with SAS interfaces, battery backup, and 80 flash modules. Without a doubt this is a huge leap forward, even for a JBOD. The performance and capacity could be used for all sorts of applications from accelerating databases, messaging, HPC applications, image or audio processing, to mundane things like VTL caching. This will be Sun's first step into selling products that could very well change the storage landscape. These same flash modules could be leveraged on larger arrays or even internal storage for servers. I have no doubt that Oracle will leverage this kind of technology with integrated high-performance database appliances.

UltraSPARC-RF Details

2009-08-31T21:43:00.000-05:00

Recently at the HotChips Conference, details about the next generation UltraSPARC Coolthreads processor, UltraSPARC-RF or Rainbow Falls as it is known, has been released. It is interesting that the designation is now RF and not KT. The presentations are focused on the challenges with creating a 16-core processor and the cryptographic technology that is embedded with the processor. Definitely a great read and insight into how Rainbow Falls works and what is in store for the future.

The most interesting parts deal with the cache coherency, flow-control, and the inter-node SMP coherency. Having servers with these processors will enable the CMT line to reach beyond the T5440 to the mid-range and possibly high-end server segments. With each processor having 16 cores and 8 threads, a 4-way server would have 64 cores and 512 threads. I would imagine that with the split coherency plane and cross-bar switch, one may be able to glue at least 4 UltraSPARC-RF processors with zero cost in specialized ASICs. This could significantly reduce the costs and complexity of such platforms up to 4 sockets. Beyond that, a specialized high-speed and low-latency interconnect would be required, as is the case on the T5440 today. It does beg the question if it's possible to create larger servers that could scale out across some standardized interconnect, perhaps Infiniband, to create modular SPARC servers capable of scaling out.

Of course a lot of details are left out in the presentations on the features of the Rainbow Falls processor. Some of the questions that come to mind are:

The NCU is depicted in one slide. Is that the integrated 10Gb Ethernet controller?
Is PCI-E v3.0 with PCI-IOV integrated? (This is still in the works from the PCI-SIG)
Given the distributed and even heat profile of the die, will the processor fit into a typical air-cooled envelope?
What process is the processor implemented in? (32nm or smaller?)
What is the power envelope?
Will DDR3 or FB-DIMMs be used?
How will the cache coherency affect cache trashing in LDoms?
How will the processor perform against T2 or SPARC64?
How will it perform against single-threaded applications?
What kind of clock speed would be deliverable with this architecture?
Will this be manufactured by TI or TSMC?

Unfortunately, we'll have to wait to see how things will play out with the server line-up for Rainbow Falls. But it would stand to reason that this will be the workhorse for future CMT servers.

DoJ Approves Sun Acquisition

2009-08-31T20:54:00.000-05:00

This is a little old and I apologize. But the US Dept. of Justice has approved the acquisition of Sun Microsystems by Oracle. This brings the acquisition a step closer to completing since the share holders have already given approval. Now it is up to the EU to give approval and things can proceed to completion. Of course there has been a lot of speculation in the news as to Oracles commitment to Sun products, in particular SPARC and Solaris. Beyond the words spoken by Larry to Reuters very little has been stated publicly. Of course it's the right thing to do until everything is approved and completed. However, there are some glimpses into how Oracle will position Sun and Oracle products together.

Probably the best that is just starting to get attention is a pre-announcement of comparison of Sun Solaris SPARC servers outperforming the highest-end configured IBM Power configuration on the TCP-C transaction benchmark. It's this kind of marketing that Sun has desperately needed over the years. Unfortunately, we'll have to wait until October 14 to find out more about the findings and details. Of course to those who have deployed large Oracle databases, this is no surprised as Sun servers are cheaper and perform just as well, if not better than the IBM Power servers in real-world usage. But the important things to pick up on this bit of marketing information are:

Oracle is serious about positioning Solaris and SPARC with its software.
Oracle wants to go "head-to-head" with IBM.
Oracle will spend marketing money and effort on proving the value of Sun and Oracle products.
Oracle is very serious about selling integrated products.

Also, if you look closely at the picture, you'll be able to tell that the servers are T5440's! It'll be very interesting to learn more when the announcement is made in October.

How UltraSPARC-T2/T2+ Beats Power6+ and Itanium

2009-07-23T08:57:00.000-05:00

I came across this great blog entry from Sun on how the new 1.6Ghz UltraSPARC-T2/T2+ processor can outperform the Power6+ and Itanium processors. It's an excellent read and uses results from SPEC to demonstrate how CMT throughput matters more than clock speed or cache size.

I have used the SPEC CPU2006 benchmarks in the past to do comparisons of different servers. While it may not reflect well on "real world" workloads, it's a good base indicator. When one factors in the results from Oracle, SAP, SPECjbb, etc. a more realistic picture can be obtained. Of course the best way to understand how a server will behave with your application is actually run your application on it and simulate the workload. There are many commercial products out there, such as HP's Mercury tools that can automate such tests. In environments that I have worked in previously that have sensitive performance requirements, having such tools are invaluable for sizing and comparisons.

1.6Ghz T2/T2+ Systems Released

2009-07-22T09:12:00.000-05:00

Today Sun has announced the availability of 1.6Ghz UltraSPARC-T2 and UltraSPARC-T2+ processors for the T-Series servers. As a result, the higher-end configurations that had the 1.4Ghz 8-core processors will now have the 1.6Ghz processors. The 1.2Ghz 8-core processor configurations will now have 1.4Ghz 8-core processors. And the 1.2Ghz 4-core processor configurations will remain on the low-end. Pricing wise, it looks like the 1.4Ghz 8-core processors are priced the way the 1.2Ghz 8-cores were priced. The 1.6Ghz 8-core processors seem to be priced at the old 1.4Ghz 8-core price point. So it's a nice speed increase for the product line. Here are some more enhancements:

Solaris 10 5/09 OS Preloaded
LDoms Manager and MIB 1.2 Pre-install
ILOM 3.0
CMT Tools 1.0 Pre-install
GCC 4 for SPARC Systems 4.0.4 Pre-install
Sun Studio 12 Pre-install
SYS, FW, Download UTIL Pre-install
MAI, 10 GBE ETCSYS CFG
Live Upgrade, ABE Pre-install
SATA DVD-RW Drives instead of PATA
EOL of 1GB FB-DIMMs
4GB 800Mhz FB-DIMM Option for T5440 1.6Ghz Configurations
New Disk Backplanes on certain models
32GB Solid State Disk Option
New Perforated Chassis Covers

The great thing is that the T-Series servers are pre-loaded with the latest and greatest software and firmware, making them ready for LDoms. I'll assume that at some point the UltraSPARC-T1 systems will be EOL'd, to make room for the UltraSPARC-KT servers when they come out.

As customers leverage these servers for consolidation and virtualization, the need for good management tools will become extremely important. Luckily, while xVM OpsCenter is taking a long time to support LDoms, there is Wand from Ecoviv that can fill the gap with a centralized web management interface.

It's rather unfortunate that IBM's FUD machine continues to spread lies about CMT, but as they say the proof is in the pudding. With the 1.6Ghz announcement, there is also a new benchmark with the T5440 running the SAP SD benchmark where it beat competing Power6 and Itanium configurations on both price and performance. There are some other interesting customer wins that are mentioned with the benchmark above, located here. The T5440 also beat an HP DL580 G5 in the SPEC jAppServer2004 benchmark by 74%. Altogether, this demonstrates that CMT can compete in web, application, and database tiers.

Realistically, all the major ISV vendors have already moved to the multi-threaded model and developers can leverage SunStudio and JavaStudio to optimize their code. This will continue to become prevalent as even commodity processors for desktop, laptops, and even mobile devices continues to shift toward multi-core and multi-threaded processors. The clock speeds of processors can no longer be pushed without some significant manufacturing changes or everyone switching to liquid-cooling. And even then, the issue of large latencies for memory and I/O would become more obvious. With such constraints and the increasing emphasis on better power performance, CMT is still ahead of the pack.

Interview with SunNews on Blog Talk Radio

2009-07-17T10:14:00.000-05:00

Yesterday I had the pleasure of being interviewed by Chhandomay Mandal for SunNews on Blog Talk Radio. The interview was on OpenSolaris and my involvement in the community. A transcript can be found here. Below is the podcast from the interview, hope you enjoy! It was a lot of fun to participate in!

Sun Stockholders Approve Oracle Bid

2009-07-16T17:16:00.000-05:00

Today the stockholders voted with ~62% approval for the Oracle acquisition bid. From my reading of the SEC filing documents, it would appear that Oracle will retain the Sun branding and that Sun will be a subsidiary of Oracle. Now we just have to wait for US and EU approval for the acquisition. Unfortunately, it's during these "hurry-up-n-wait" moments that things tend to go into a limbo state. Hopefully, Oracle and Sun management will shift things into gear quickly to avoid such lingering limbo states.

I know there has been some apprehension and negativity from crack-pot journalists and anti-Sun folks out there about this deal. However, I remain rather up-beat about this as Oracle has been very positive about creating a on-stop shop for fully integrated stacks of hardware and software that include Solaris, SPARC/x64 servers, MySQL, JAVA, storage, etc. Once things are on-track, IBM and HP will have a rude awakening that Oracle is after them with a full product line-up, powerful marketing, and a sales force that can push products.

I also realize that not all things are bright and wonderful. Things change at companies and Sun is not the same company I use to work for years ago. But it has not affected my use or enthusiasm for its products or services. I have no doubts that there will be layoffs of redundant corporate components, contractors, engineers, and of course everyday employees. Not to mention that some low-hanging or non-profitable products or R&D work will get canceled. Unfortunately, that's just part of business and house-cleaning. I do think that cuts and RIFs have been made recently to get things in order as it where. To any employees that may face such a grim situation, I wish you well and ask that you don't give up. Once the dust settles, there will be new opportunities at Sun/Oracle. Hopefully Oracle management will be able to take Sun to new heights and back into a leadership position in the market.

LDoms 1.2 Released!

2009-07-09T19:02:00.001-05:00

Well it's that time of the year again when a new version of LDoms has been released! Version 1.2 has some great new features and enhancements:

Physical-to-Virtual Migration Tool
Configuration Assistant
Power Management for the CPU
Jumbo TCP/IP Frames for Virtual Networking
Restriction of Delayed Reaction for the Control Domain
Domain Dependency Configurations
Auto-Recovery of Configurations
Exporting Virtual Storage Back-ends Multiple Times
API to Support LDMD Auto-Discovery

The included P2V tool allows one to take a sun4u server running Solaris 8-10 and convert it into a Solaris 10 sun4v image! This is very similar to the Solaris 8/9 Migration Assistant for Containers. The Configuration Assistant gives you an ncurses and Java GUI for doing the initial configuration of your server. Definitely a handy tool for getting people started. The other enhancements will help with power management, networking, configuration management, sharing storage, and enabling management products such as Wand.

I've already updated the main LDoms Community page with all the links to the software, firmware, and documentation. I'll have some more posts over the weekend as I upgrade my environment.

Also for your viewing pleasure, here are some links to demos done on rapid provisioning and mobility with LDoms. Also a sneak peak at xVM OpsCenter managing LDoms is here:)

I'd like to thank all the Sun Engineers and Managers involved with the 1.2 release! Continue doing an excellent job and keep those features coming:)

PS, if you can't find the 1.2 release on the main download site, here is the direct link.

He's Dead Jim!

2009-06-16T09:02:00.000-05:00

Well the rumor on the street this week is that Sun has killed off the UltraSPARC-RK or ROCK processor. This was suppose to be a 16 core processor for the upcoming SuperNova servers. This would be an unfortunate blow to the SPARC family and to Sun customers as there was a lot of hope that ROCK would take the high-end of CMT against the M-Series servers and allow customers to consolidate large amounts of applications. Hopefully, this is all just rumor and speculation. I could only imagine this happening if one of the following were true:

The 8-core SPARC64 VIIIfx "Venus" processor out-performs ROCK by a significant amount, thusly making it impractical to move forward with ROCK.
The 16-core UltraSPARC-KT "Niagara 3" processor out-performs ROCK by a significant amoutn, thusly making it the new high-end CMT processor.

Unfortunately, without real sources to confirm any of this, it's difficult to know for certain if ROCK has been canceled or just delayed again. There have been numerous delays over the past few years and some have been rumored to be the result of layoffs and politics. Either way, it's not looking good for the ROCK processor or SuperNova systems at this point. If it has been canceled, it'll be a big blow to the LDoms Community and users as those servers would have been LDOM enabled and given a higher level of RAS features than are currently available on volume CMT servers. Time will tell..

SPARC64 Strikes Back on Super-Computing

2009-05-14T20:58:00.000-05:00

It would appear that Fujitsu has learned a lot from Sun's CMT processor designs and is testing out its new SPARC64 VIIIfx or Venus processor which is an 8 core SPARC64 chip that can perform 128 Giga-flops! Here are some interesting points about this processor:

8 core SPARCv9 with extended SIMD extensions for supercomputing, known as HPC-ACE.
Dedicated L1 data and instruction cache for each core.
Shared 6MB L2 cache
45nm CMOS process on 2 cm squared die
2x the number of transistors as the SPARC64 VII, yet only one third the power consumption.
Integrated DDR3 Memory Controller
Hardware optimization for non-parallelized traditional workloads
128 Giga-flops performance per chip, more than 2.5 times the fastest Intel processor.
Liquid cooled!

Fujitsu is first targetting this processor for super-computer applications into a 3D torus arrangement that will allow for over 100,000 nodes to be connected together and deliver around 12.8 Peta-flops! This will out perform the FX1 SPARC64 VII super-computers which run at 40 Giga-flops per chip. The FX1 is only sold in Japan, you can read more about it at this link.

Here is a picture of a Venus node:

As you can see, the front of the module is on the right-hand side where the latches are. You can also see what look like the liquid cooling connector on that side. This would suggest that the node plugs into a larger chassis and that the I/O components are on a separate module. I would not be surprised to see that the frame uses a custom switch-less fabric on the back-end, with mainframe level RAS, and SPARC Enterprise M-Series features.

It remains to be seen when this processor will make its way onto the M-series or a future SPARC product line for Fujitsu and Oracle/Sun. However, with the investment from Japan in deploying this in super-computers and Fujitsu's focus on making traditional workloads benefit from this new processor, it is very likely we'll see Venus based SPARC servers in the near future. The M-series was designed to handle several generational upgrades. It has already gone from Dual-Core to Quad-Core. As such, an 8-core upgrade is only logical. How this will play into Sun's ROCK processor is in question as well. It would appear that Fujitsu has indeed learned a lot from Sun around building multi-core and multi-threaded processors. ROCK and next-gen Niagara processors may be more focused on general purpose computing, while Venus is focused on large number crunching super-computer tasks.

Mergers and Legal Dribble

2009-05-14T20:19:00.000-05:00

The SEC filings on the merger are available for the public to review. The document has lots of interesting details about the merger process, conditions, etc. Reading through the document provides an interesting lesson in how mergers are designed and how things get sorted out. Items such as severance pay for the executives, beneficial stock holders, waiving of rights, etc. are all very fascinating to read through. It provides an interesting glimpse into the legal process and how companies organize such activities. I was not surprised to see my old employer BGI holding a significant amount of securities in the company. I think those who hold on to their stock and enjoy the gains from the conversion to Oracle common stock will be happy over time.

Of course some of the other interesting bits of information surround the history behind the merger itself. Some of the other parties who were interested, have been masked over. Reading through the document, it's obvious the other suiters were HP, Fujitsu, and of course IBM. I never thought that Sun would be a good fit for HP, other than to kill off the competition, considering HP's rather bloody and unfruitful past mergers. Fujitsu has done worse in the server market space outside of the Asia region, and has such a limited market in the US, it didn't make sense unless they were willing to commit to the US market. While the R&D mentality of Fujitsu would have brought something to the table with Sun, they seem to suffer the same sales and marketing issues Sun has been unable to address over the years. Ultimately, I still think the Oracle merger makes the most sense seeing how Oracle has strong ambitions, sales, and marketing power. The vision of supplying the whole solution stack makes increasingly more sense as one looks at the IT landscape. Oracle stands to change the game and go headon with IBM and HP.

Speaking of a changing landscape, Rackable is merging with SGI and taking on its name. SGI has been through some rough times over the years. I remember when I first lived in Silicon Valley seeing the huge campus for SGI and thinking they must be doing well. It's amazing how fast things can change.

And speaking of Oracle, it is buying out Virtual Iron which will give Oracle access to some interesting Xen management tools. It might also be interesting to see how Oracle will cherry-pick components from Oracle VM, Virtual Iron, and Sun's xVM. Oracle is definitely tooling up on components for its stack. Considering that Oracle will own Dynamic Systems Domains, Logical Domains, Solaris Containers, xVM Server, VirtualBox, Sun Rays, etc. through the Sun merger, they will have the widest virtualization portfolio in the industry. A solution for each situation and platform. Something to think about:)

More Rumblings and Solar Flares

2009-05-11T17:36:00.001-05:00

There has been some talk lately over Sun's SEC filings which list some interesting turn of events:

Sun shareholders acting to block the Oracle take-over.
Violations of bribery laws

The first one, I am not surprised at since Sun's management has indeed failed to meet their responsibilities to the shareholders for manning the ship and keeping the profits up. I'm not sure how this will help, since a failure to complete the acquisition would probably place Sun into further uncertainty and force it to go private or bust. Hopefully this will get resolved peacefully. But I do understand the frustration and anger shareholders must have for being let down constantly over the past 8 years.

The second item though, is indeed troubling and could hurt Sun's position in the GSA program. Idealistically, business should not resort to such tactics. But that is not the way of the world. Recently news of the EU filing anti-competitive charges against Intel for paying manufacturers to not sell AMD processors demonstrates this clearly. Not to mention that Microsoft is having its own issues abroad as well. I've seen enough business deals change direction due to vendors using such tactics or discounting hardware to bargain-basement prices just to make the sale, but charge you an arm and a leg on support and licensing. The sad thing is that customers forget that they have choices and need to leverage that to not only get a good price, but to have a good business partnership based on trust. Hopefully those responsible for the bribery issues will be dealt with. Atleast Sun has been honest about this issue and is helping the authorities.

Ecoviv Beta Testing Wand for LDoms

2009-05-08T12:34:00.000-05:00

The company Ecoviv is beginning to beta test their management product for LDoms, called Wand. This product was previously known as Virtvue, but has some significant enhancements:

Web based remote management of LDoms
A robust three tier architecture that leverages Solaris’ Cacao Agent
Support for LDoms 1.1
LDoms Life-Cycle Management
Live Migration
Provisioning of the Logical Domains Manager and of Guest Domains
LDom Search and Discovery
SNMP Support

This is great news and I'll hopefully get into the beta program. This would be a fun product to write an article or white paper on.

Oracle Confirms Commitment to SPARC

2009-05-07T22:41:00.000-05:00

Reuters did an interview with Larry Ellison over the acquisition of Sun that has some great news. In the interview the subject of software vs hardware came up. He confirmed that Oracle will retain Sun's hardware business and increase development of the SPARC processors as important to Oracle's strategy for delivering a complete solution stack. This is great news for SPARC customers to hear and understand. Larry Ellison even specifically stated that Oracle wants to compete directly with IBM in the data center for dominance. Oracle is aware of the value of the SPARC64 and the UltraSPARC CMT lines as they are power efficient, scalable, and heading in the right direction. I highly recommend this article for customers, users, engineers, etc. to get the message out that Oracle is committed to Solaris and SPARC. I think with Oracle's focus on competing with IBM Power systems, we'll see some significant focus on ROCK and the next generation CMT systems. Hopefully, we'll see SPARC kick Power7 in price and performance.

Solaris 10 05/09 Released

2009-05-04T19:11:00.000-05:00

The latest update to Solaris 10 has been released. There are new features and fixes for Solaris Containers, SMF, IPSec, 10GbE, iSCSI, Infiniband, SunSSH, and even SunVTS. No major updates on platform support or virtualization. Seems this should be a fairly stable update.

In other news, it's encouraging to finally see builds of the Indiana or OpenSolaris.com flavour for SPARC. Unfortunately, this release can only be installed by the new AI installer over the network and can not be booted from CD/DVD or through Jumpstart. Another major limitation is that Xsun is not included, which limits framebuffer support significantly. Until recently, we only had Solaris Express or one of the other distributions such as MilaX or MarTux. You can check these different distros out on Genunix. Overall, it is encouraging to see the number of distros for OpenSolaris grow and mature. Hopefully, the Indiana codebase and support will mature quickly on SPARC. (We can only hope that Oracle will see the value in making cheap UltraSPARC-T2 workstations).

Switching from SPARC to and Opteron

2009-05-02T22:50:00.000-05:00

A couple of weeks back, I bought an Ultra20 off of Ebay to replace my aging Sun Blade 2000. I've been pretty happy with my Sun Blade 2000 over the past few years. However, it was beginning to worry me that due to the closed source drivers for the Expert3D card, I would eventually be left in the dark with OpenSolaris and probably Solaris 11 when it comes out. And due to the fact that Sun no longer makes a SPARC workstation, it seemed like it was time to move on. I have for years run Solaris Express and recently OpenSolaris on VMware and VirtualBox on my work laptops. I do miss having OpenBoot and the Sun Logo lit up on my workstation, but I'm very happy with the transition. Everything obviously runs faster and getting use to IPS has been pretty easy. I did do a few upgrades to my Ultra20:

Opteron 180 2.4Ghz Dual-Core
NVidia GeForce 8400 GS PCI-E
Sony DVD+-RW Drive
IO Gear USB->Serial Adapter (Still need serial access to gear I test)

While it's not the fastest x64 system out there, it was amazingly cheap on EBay and works great! Here is a picture of the system before the upgrades:

Here is my old system, which I'm selling on EBay:

Here is my new desktop on the Ultra20 using the Avant Window Navigator which gives a kinda MacOS X feel and is very snappy:

I've taken the time to recompile many of the apps and games I had on my Sun Blade 2000. I'll have to see if I can package them up into IPS and distribute them for other folks. Overall, the transition has been fairly painless and a lot of fun. At somepoint, I'll have to upgrade the motherboard, CPU, and RAM to a newer Opteron generation or a Phenom II x4 so I can do xVM Server testing. Now all I need is my own T1000 or T5120 for doing LDoms on.

Fujitsu Out-Sourcing to TSMC

2009-05-02T22:27:00.000-05:00

Well according to reports, Fujitsu is divesting itself of semi-conductor fabrication and out-sourcing it to TSMC. This is not as surprising if you consider the amount of consolidation in the technology sector. This is something that has already happened in hard drive manufacturing where IBM sold out to Hitachi and in July Toshiba will own Fujitsu's hard drive business.

What I find really interesting is that Sun had already selected TSMC to fabricate the ROCK or UltraSPARC-RK processors and the next generation Niagara processors. Fujitsu was already producing the SPARC64 processors for the SPARC Enterprise M-Series servers. It'll be interesting to see if TSMC will be manufacturing the SPARC64's as well in the near future. It's unfortunate that Texas Instruments will not build a next generation fab for the Niagara chips, but it is an expensive investment. Even AMD has spun off its chip manufacturing business due to the expenses involved. This leaves Intel, Samsung, TSMC, IBM, and the Foundary Co. as the major chip manufactures for processors.

While there was speculation that Fujitsu would buy Sun, considering how Fujitsu has been divesting itself of manufacturing hardware, it didn't seem to make a lot of sense. The SPARC Enterprise servers are not manufactured by Sun or Fujitsu, only certain parts are. The rest is manufactured and assembled by Flextronics. Even HP sold off its share of the Itanium and Alpha business to Intel. To make things even more interesting, with the rise of SSD storage in servers and storage arrays, the future of the storage market will have little to do with spinning disks anymore and a whole lot more to do with chip manufacturing. This is something that Intel and Samsung are investing heavily into and even Sun has worked closely with both to further this trend.

The Niagara processors have been ahead of the general purpose CPUs in the multi-core and multi-threaded game. It is interesting that since the introduction of the T1000/T2000 servers, Intel and AMD have stepped up on the creation of dual, quad, etc core processors. Even IBM has realized that Sun was right and is now having to eat its own words against CMT as it develops the Power 7 processor featuring multi-cores and threads. Another trend in this area is the inclusion of features such as memory controllers and large caches into processors. Of course the UltraSPARC-T2 and UltraSPARC-T2+ processors have additional features such as cryptographic units per core, on-board PCI-E controller, on the UltraSPARC-T2 a 10GbE controller, and on the UltraSPARC-T2+ SMP interconnects. The concept of a "system on a chip" will continue to expand and become prevalent.

A New Oracle/Sun World

2009-04-24T08:55:00.001-05:00

Now that a lot of the dust has settled around this wild and crazy saga, it's time to take a serious look at what the impact will be on the industry and those who work in it.

I think it's important to take a quick look at their history together:

Oracle has been around since 1979 and Sun since 1982. So over a quarter of a century in the industry, which is a long time.
Oracle started out with its RDBMS product and Sun with its first workstation.
Both shot ahead of the competition and become key players in their fields.
Both expanded their "Enterprise" reach in the late 90's and profited well from the internet bubble (Oracle OPS, Sun E10K, etc.)
Both have partnered together over the years to bring to market complementary products (Oracle DB, Solaris, SPARC, Sun Cluster, etc.).
Both had a vision for network computing (NCI, Javastations, SunRays). With SunRay sales growing still to this day.
Both have invested heavily into Java.
Both jumped into the middleware market through acquisitions (BEA, iPlanet, etc.) and have significant market penetration as a result.
Both own some interesting open source IP; BerkeleyDB, InnoDB, MySQL, etc. Sun has open sourced the majority of its assets (OpenSolaris, Solaris Cluster, OpenDS, Glassfish, Fishworks, xVM, etc.).

As you can see there are a lot of parallels between the companies.

There has been a lot of speculation around what Oracle will "prune" from Sun once things are all said and done. Unfortunately, most of the speculation is still stuck in a pre-Oracle/Sun world. What many people have not caught onto is that Oracle knows it bought a systems company that has both software and hardware. This was not by mistake or on a whim. Oracle understands that their customers typically buy Sun products to provide the infrastructure (SPARC, Solaris, and storage) when buying from Oracle. While Oracle products tend to cost more than the hardware it goes on, it's still lost opportunity. What Oracle has done is to take control of the very ecosystem their bread-winning products rely upon. And by doing so, it now has a full portfolio of solutions to go-to market with. Let's sum it up:

Servers (x64 Galaxy Servers, 6000/8000 Blades, T-Series CMT, M-Series SPARC64, and the future Niagara/ROCK platforms).
Storage (Open Storage, Unified Storage, SSD products, Fishworks, etc.)
Operating System (Solaris/OpenSolaris, Oracle Unbreakable Linux)
Middleware (Java Enterprise System, Glassfish, OpenDS, Weblogic, Oracle ERP/CRM, Peoplesoft, JD Edwards, etc.)
Database (Oracle DB, Siebel, Berkeley DB, InnoDB, MySQL, JavaDB)
HA (Sun Cluster, Oracle CRS)
Virtualization (Solaris Containers, LDoms, Dynamic System Domains, xVM Server, SunRays/VDI, Oracle VM, etc.)
Grid/HPC (N1GE/SGE, Sun HPC Tools, Infiniband switches, pNFS development)
Management Software (xVM OpsCenter, SunMC, Oracle Enterprise Mgr, etc.)
Development Software (Java, Sun Studio, JavaStudio, Netbeans, JDeveloper, etc.)

While there are some redundancies between the two, they are really only in the middleware, database, and development tool space. Now many think that Oracle will rid itself of such redundancies. However, people who think that haven't really paid attention to Oracle's product line-up. A quick look will demonstrate that Oracle typically will rebrand its software acquisitions and keep them going, great examples being Weblogic, PeopleSoft, JD Edwards, Siebel CRM. What this shows is that Oracle understands how to fit acquisitions into its portfolio and realizes that its difficult and inefficient to consume point solutions into a single product. It also understand that each product has its place in the market and there is money to be made through that. So these redundancies will probably go down one of these paths:

Continue with a Oracle rebranding
Replace an Oracle product
Merged with an Oracle product
EOL'd

There is a good chance that many of the redundancies will either complement Oracle's portfolio or enable some major enhancements. Of course some things will probably be axed that don't have a profitable future, but I think those will be fewer than what people may think.

Now on the servers, storage, operating system, HA, virtualization, and grid components these are new areas for Oracle. As posted in my last blog entry, Oracle realizes the value in these components. These are the components that will enable Oracle to compete directly with IBM and HP by selling everything from the server and storage, all the way up the stack as an integrated solution. This is exactly what Oracle has published and said their intentions are. What most people may not realize is that while Sun may have had issues with running its business operating expenses, it is in the top 5 server maker ranks. What it lacked was good sales and marketing to go up against IBM and HP consistently. Sun has always been an innovated company and currently has an amazing product line-up. The challenges have been:

Poor economy has shifted businesses from buying servers like crazy to cutting costs and leveraging things like virtualization.
Technology has improved to the point where you can do more with less. An few v880s can be replaced with a T5220.
Competitive pricing has driven the returns down on hardware. Consider that even an UltraSPARC-T2 server is relatively cheap these days.
Poor marketing makes it harder to sell things.
Poor strategies and mixed messages create confusion.

So what will Oracle do? I think it's clear that Oracle will make the tough choices that Sun management would not dare to. This means making cuts to bring the expenses in-line with the profit margins. It will also find ways to quickly build integrated solutions that combine the servers, storage, operating system, virtualization, and the software stack together. But the key thing is that Oracle has a powerful sales and marketing force that can turn things around for the Sun products. If anything, Oracle will sell more and bring in greater profit margins. The angle here will be rather difficult for IBM and HP to compete with, or for customers to resist. Oracle is now in the position to sell you the whole enchilada. Say you want to build an e-commerce company or a new trading system that needs a large data warehouse.. BAM! one-stop shopping with Oracle. Now it's just one throat to squeeze when things don't work. No more calls involving everyone(Sun, EMC, Symantec, Oracle, etc.) and their grandmother. Now it's just one company to call for the solution and the support. Oracle has completely changed the playing field.

Think that's crazy? Well wake up.. it's here and now. IBM, HP, EMC, Cisco, etc all envision the same future where you'll buy a whole stack or an on-demand service (think Salesforce.com or Amazon) from one vendor. The whole point of utility/cloud computing and getting everyone and everything connected is to make technology ubiquitous and turn it into a pay-for service. Just think of how buying music has changed. You use to goto a music store and buy a CD, then play it on your stereo in your house or car. Now you download it on iTunes and play it on your iPod where ever you happen to be. There will be more consolidation in the technology industry and this is only the beginning.

So what does this mean for customers, sys-admins, engineers, etc? It means that there is a brighter future in Sun products and that Oracle will not be doing to Sun what IBM undoubtedly would have done. For professionals, this means there will be a growing demand for your talents and a need for enterprise level thinking around integrating Oracle's solutions.

Oracle + Sun = Win!

2009-04-20T09:48:00.000-05:00

Well today is a big day! Sun has announced that is being purchased by Oracle! At first, I wasn't sure what to think. But considering that Oracle already owns the majority of the software stack that Sun systems are purchased to run (Oracle DB, Weblogic, etc.). This completes the stack (Sun servers, storage, Solaris, Java, MySQL, Java Enterprise System, virtualization, SunRay thin clients, grid computing, etc.).

Probably the best thing to take a look at is this great document from Oracle which outlines the benefits and Oracle's committment to Sun. Here are some key points from the document:

Page 4:
Sun’s key product assets are expected to provide substantial long-term strategic customer advantages
• Ensure continued innovation and investment in Java technology
• Optimize Solaris and Oracle for better performance, reliability, and manageability
• Protects massive customer investment in SPARC
• Open Storage built with industry standard servers and components

Page 5:
Protects and extends customers’ investment in Sun technologies
• Accelerate growth of Java as an open industry standard development platform
• Sustain Solaris as an industry standard OS for Oracle software
• Continue Open Storage and Systems focus and innovation

This is great news! It demonstrates Oracle's understanding of Sun's value and products. It does not surprise me as Oracle has a huge software portfolio and many of their customers make heavy use of Sun equipment and software. The really amazing thing about this is that it repositions Sun against IBM and HP. Oracle has almost 3 times the number of employees, a huge sales force, numerous channel partners, and excellent marketing. Oracle has an aggressive sales force, which is exactly what Sun needs to get things back on track.

Now we just have to see how things turn out in the coming weeks as this acquisition goes through. I'm sure there will be some restructuring as things proceed. Overall, I'm happy that Sun is not being bought out by IBM or HP.

More Rumors

2009-04-16T13:37:00.001-05:00

Well it would appear that not only has IBM pulled away from the table, but regulators have advised such action, according to Reuters. I am not surprised at all by this as it would create a monopoly in the commercial UNIX market and reduce consumer choice. Especially given that Sun products are heavily used in government and military applications. In the end, this is good for Sun customers and shareholders. While the stock price has dipped, it should force the board of directors to focus on strategies to get Sun back on track as oppose to an exit strategy. Realistically, Sun will have to make some deeper cuts and focus on the products that bring in revenue. The other aspect would be to cut the R&D budget and align the resources there to support continued revenue generation.

With the reorganization that took place last year, the Systems Group stands as the major bread winner for Sun. This consists of the server, storage, Solaris, virtualization, support, and management software products. These are definitely the areas where Sun should focus it's resources and look at making cuts in the other divisions.

There are some interesting lessons to be learned from the situation Sun is in:

Sun is a hardware company. This has been a debate issue inside of Sun for a long time. But it's painfully clear that giving away software for free does not help the bottom line. While it may help sell the hardware, it does not ensure continued revenue growth when it is given away for free. As such Sun is dependent on continued hardware sales for real growth.
If enterprise software gets the job done with little fuss and is given away for free, there is little incentive for buy a support contract. This is especially true in a down economy where it may be cheaper to pay for support per incident than to pay an annual support contract. While giving it away for free can lead to wide adoption, it does not lead to massive amounts of customers paying for support contracts.
Value Added Resellers and Systems Integrators are key to sales. It is not widely known that years ago Sun depended heavily on these sales channels for revenue and that it was very successful. Unfortunately, Sun decided to take on this process itself. This led to VAR and SI companies going bust or merging. Ultimately, Sun was not able to replace this valuable channel process and it was not until late 2008 that the tide turned with Sun shifting the focus back to VARs and SIs. Of course, this was "too little, too late" to affect sales as it takes time for VARs and SIs to ramp up.
You can't make money if you don't market your company properly. This too has been a rather sticky item. Without good commercials, ads, and marketing campaigns it is very hard to get the message out. You might be surprised to hear what every-day people say when someone mentions Sun to them. Comments such as "Oh, they're still around?", should pretty much raise up the red flags all over the place. However, that is how bad the lack of marketing has been for Sun over the past couple of years. They have great products and are making sales. But it a whole lot harder to get things out in volume when non-technical investors, business owners, and executives who sign-off on a PO are not aware of a company beyond the Java and "dot-bubble era" history.
In a troubled economy where poor sales are unavoidable, it's important to sell customers what they need and at the right price. With the plethera of old Sun gear in corporate datacenters, and customers looking to cut costs, this is an exceptional opportunity to help customers virtualize and consolidate. This is the area that I specialize in and do consulting around. Sun should do more in this area to drive up hardware sales and professional services. When you see vendors like VMware giving away services to make the sale, it's easy to see how competitive this market can be. But this is the big focus for everyone today. Sun has a lot of virtualization solutions, but really needs to get xVM Server out of beta.

Interestingly, according to CNBC the regulatory concerns were an issue, but also the executive pay-outs. If the rumors in this story are true, we may see Scott McNealy take the helm again at Sun. We can only hope that there will be more changes to get things back on track at Sun. Right now the continued path of silence and jingoism is not helping customers or investors.

IBM Deal Breaks Down! Yay!

2009-04-05T20:37:00.000-05:00

It would appear that according to the NY Times, the deal between IBM and Sun broke down this weekend over pricing! While I'm still very upset that Sun would even consider selling out, I'm very happy that the deal is off. This will definitely make the folks on Wall St. react negatively tomorrow. However, it was an enormous mistake for Sun's executive management to put soo much stock in the perception that Wall St. has towards Sun. The reality is that while Sun is struggling, so are other companies. However, how many server makers in the world can say they are in the top 5? Sun is still doing business and making money! It may not make the kind of money IBM or HP make, but if you were to compare the UNIX divisions of those companies and look at the units sold, you'd see that Sun is doing well.

What really has to happen is a change of management at Sun. I've seen an epidemic develop over the years in America where the bean counters are running companies and not visionaries who actually understand the products and customers. This has developed into a situation where companies are soo fixated on short-term market returns, they have no clue where things are going for the long-term. If there is anything to learn from this difficult economy, is that such shortsightedness leads no where and devistates companies, customers, employees, and most of all families.

I hope that the shareholders of Sun call for a change of guard and push for someone with real vision and insight to lead Sun. Sun has the right products without a doubt, but needs to position them better, market them properly, and most importantly.. get sales folks and channel partners to step it up. It's time for Sun to make it on its very own and fight the good fight!

Another week passes..

2009-04-03T15:47:00.000-05:00

Still no announcements from either side. But the current rumor is that the buy-out will happen on Monday. Hopefully, this is all just a bunch of hogwash or at worst Sun is just selling off the weak parts of its business. I would not mind a smaller and more focused Sun. One focused on OpenSolaris, CMT/ROCK, x64, Virtualization, and storage appliances. I think that angle could be very profitable without the burden of the rest of the business. However, I get the feeling that things are not soo rosy or idealistic.

To that point, I'm rather disappointed that the number of signatures on the petition I created has been soo low, it's not funny. I was really hoping that thousands of people would sign the petition. Instead, I got comments that it's hopeless and pointless. However, I still think that if enough people were to show their lack of support for selling Sun by signing the petition, it could open the eyes of Sun management and even the shareholders that would have to approve such an aquisition.

Well here's the best to keeping our fingers cross that Sun still be here for many years to come.

Keep Sun Microsystems Inc. Independent

2009-03-30T18:20:00.000-05:00

I was recently asked through my blog how one can make their voices heard by the management of Sun. I thought long and hard about this and decided the best thing was to start a petition. I urge all readers of my blog to sign this petition and to tell their friends and colleagues to do so also. I'd like to get as many signatures at possible. Here is the link:

http://www.petitiononline.com/smi09/petition.html

I will attempt to make the management at Sun aware of this petition so that they understand how their customers, shareholders, employees, former employees, etc. feel about the recent turn of events.

I also would like to extend my best wishes and hopes to all the employees of Sun that have been RIF'd today.

Hopefully all of us can come together and make our voices heard!

how many people are at the table?

2009-03-27T11:34:00.000-05:00

As the saga continues.. it appears Sun has been shopping around Silicon valley trying to sell itself. There are some reports that it talked with Oracle, HP, Cisco, Lenovo, Fujitsu, and even EMC. I must say this is just insane and truly sad. To think that Sun and its management has carried on doing this behind the scenes. I'm curious if this is happening as a result of Southeastern Asset Mgmt company pushing for a sale. They upped their stake in Sun last year to something like 20% of the stock share. I'm very surprised that the US SEC has not come down on this already as the stock price has been affected by what could be in-sider trading information. Not to mention the fact that removing Sun from the server industry would cross the lines with anti-trust and anti-competitive regulations. With the massive investment the government has in Sun technology and services, I'm surprised they are not stomping all over this. If you remember back when AT&T had a large share of Sun and was thinking of buying it, the government stepped in and stopped it.

But if it's to be.. I'd rather see Sun get bought out by Cisco, Apple, Google, Fujitsu, or Hitachi. Going to IBM or HP is just a death warrant for Sun.

Sun Saga Continues

2009-03-25T22:08:00.000-05:00

Well it's been over a week now and still no news on if the IBM will buy Sun or not. The rumor mill now is that IBM lawyers are looking over Sun's contracts and projects. I'm really hoping that this does not go through and that the US SEC or the EU will step in and stop this based on anti-competitive reasons. Sun's stock has not been doing well for years, no dispute there. However, if you pay attention closely.. you'll see that the volume of traffic has always been high and if anything Sun stock is actually a good market indicator. Most tech stocks have been vastly over valued for years, look at Googles. There's no way that if you liquidated Google today that it would be worth that much, let alone generate that much cash. I think the main thing the public and customers fail to remember is that the stock market is based on perception. If it wasn't the current mortgage and credit hole we're in would not be here.

So does Sun really need to be bought out to be successful? It is the 4th largest server manufacturer, ahead of Fujitsu and behind Dell. Is that really bad? Common sense would just say that if Sun was not making money it would not be in business. To put this into perspective, I've worked for many large companies and do consulting work where I get help companies do consolidation and virtualization. So is it true that everyone and their grandma is dumping Solaris for Linux on x86? I would say no. What's really happening is that applications that use to sit on mainframes, AIX, HP-UX, Tru64, etc. are moving to either Solaris on SPARC for bigger workloads and production stuff and Linux on x86 for general purpose apps and things Windows use to do. Many companies are moving in the direction of Solaris on SPARC replaces the big-iron expensive stuff, Solaris x86 or Linux for the low-hanging fruit and general purpose stuff, and windows for all the desktop apps and infrastructure. So why are the sales not the way they were years ago?? Well it's simple, you don't need as many Sun servers to do the work these days and the pricing is competitive. So basicly, Sun has more or less engineered itself into lower profits.

For example, look at the previous post about the Internet Archive. You have 800 white-box Linux servers being replaced by 63 x4500 servers running Solaris 10. Would Sun have made more money selling 800 or a 1000 servers to them? Well sure, but that's not what they needed considering the performance and capacity of current technology. See, technology has gotten faster, smaller, and cheaper. This is also the case in the SPARC world. For example, I've seen many customers replace expensive v880's and v890's being used as big Netbackup media servers because of the I/O capacity being replaced by T2000s and T5220s. Yes, that's right the CMT servers that IBM and HP make fun of for not being so hot on single threaded apps are killer for multithreaded apps and streaming media (something you'd expect on a backup media server). Again though, something (v880/V890) that cost a lot of money can be replaced with something faster, smaller, and cheaper.

Many Sun customers are also faced with another complicated situation.. hundreds or thousands of aging servers that are 5, 7, 10, or more years old that just keep running. But of course as data center space, power, and cooling become bigger issues.. these servers become targets for upgrades. And now you end up in a different field all together. For example, lets say you have a bunch of E450s or E4000s. They are not only old, but long sense EOL'd. Such servers could easily be replaced with a T5220 or T2000. But do you need to buy a box to replace another or do you look to virtualization? And this is a decision many companies are faced with today. You have old servers that just work, are paid off, but cost more to operate. On the other side you have new, faster, and cheaper technology that can replace it.. but it's all overkill. I've seen people consolidate a couple of v880s into a single T2000. So many companies are looking into or are already consolidating and virtualizing their legacy Sun footprint into a new and smaller Sun footprint.

My point here is that Sun is obviously delivering on good technology. The problem is that there isn't as much money to squeeze out of it. And this is where having good support and professional services could add some much needed revenue. Unfortunately, this is an area where Sun has struggled to compete in. I've seen the QA work alone in the Solaris patches go down in quality over the past few years as people have been RIF'd and customer support go through some tough times. I do question the cuts Sun has made over the past few years. Even with the recent RIFs I'm not sure how things have shaped up. I know when I worked at Sun, the management layer was starting to become rather hefty and the other layers being spread out more. When you need to shrink things down, it's best to get rid of things that aren't going to bring you a long-term ROI and bring your assets together, meaning fewer management layers. There are times where you need more grunts making good products and fewer managers.

Unfortunately, I've seen Sun slowly hemerage good people and somehow not remove the un-needed fat. And with how Wall St. looks closely at the numbers, the solution to most things is more RIFs at Sun. This isn't fixing things. So does selling out really help? Well if IBM were to run Sun as a subsidurary and fix the management and marketing issues, Sun would probably do well. But I find it hard to believe that is the plan. So I hope Jonathan is reading this and will hear my plea to not sell of Sun, but instead to make some drastic changes.

Sun has already been reorged. Why not spin off the software and cloud components as private companies? That would leave the systems group as the primary company with the majority of the revenues. And think of it this way, much of the complaints on Wall St. would probably go away.. fewer employees, operating costs, etc.. and a strong revenue stream vs capex/opex. Or take the whole company off the market and go private? Restructure, get investments that would goto marketing and sales development. Then when things are better.. go back on the market strong.

Well anyways, these are just some of my ramblings on the subject. We just have to wait and see how this unfolds.

Wayback Machine Powered by Sun

2009-03-25T22:02:00.000-05:00

Well it was time to upgrade the Internet Archive with a Sun modular data center shipping container with 63 x4500 servers running Solaris 10. Obviously they are making use of ZFS and lots of 1TB disks. This shows how great technology can help businesses consolidate their footprint. The previous setup consisted of 800 Linux servers with only 4 drives each. Not bad going from 800 down to 63 servers.

Sun Selling Out??

2009-03-18T12:12:00.000-05:00

Well the rumor mills are spinning big time today. It looks like Sun is doing the rounds and seeing who might buy them out. The rumor today is that IBM is interested. I think this would be the worst thing for Sun to do, hands down! Sun has a great product line-up. The issue is not the products or the engineers behind it. It's the poor marketing and management of those assets.

Sun would be better off selling to Apple or Cisco who understand marketing. Apple would be my ideal choice because they need help on getting into the server market. A good idea would be to see MacOS X drop the Darwin stuff and adopt OpenSolaris. Imagine MacOS X GUI and apps running ontop of OpenSolaris. That would be a killer combo for the consumer and enterprise markets.

But to sell out to IBM would be a huge betrail for all Sun customers and supporters across the board. When it comes to the commercial UNIX landscape, it's just Solaris and AIX these days. And realistically, most shops have reduced or eliminated their AIX and HP-UX footprints in favor of Solaris and Linux. Sun needs to get its butt in gear and help continue this trend.

The only way this would be good is if IBM were to dump AIX and push forward on Solaris. There has already been a port of OpenSolaris to the POWER platform that IBM could work on completing and marketing. But honestly, I can't see IBM making that choice. I've been through this once before years ago when I worked on and used Digital Unix (Tru64) and it was eaten by Compaq and trashed by HP. I'd hate to see Sun make the mistakes of Digital and Compaq all over again.

What's really needed is a change of management at Sun. If they want to make money, they have the right products and solutions. It's just a matter of selling it well and getting out infront of customers again. I mean when was the last time we saw a Sun commercial on TV? I've asked this countless times to Sun execs and management. IBM, HP, Apple, Cisco, etc. all have TV commercials and are in consumer and enterprise customer minds as brandnames. This is not rocket science!

I urge everyone to contact their Sun sales rep and complain. Also to write letters to Jonathan Schwartz and send him a clear message that selling out is not the solution.

Supercomputer Win for Sun

2009-03-17T11:48:00.000-05:00

I've noticed over the past few years that Sun has been winning contracts to build Supercomputers which is great. The Ranger supercomputer at the University of Texas at Austin is pretty impressive. Today, Sun won a contract in South Africa, read more here. One of my favorite websites that has an interesting collection of Supercomputers is HPCVL. Many of the Supercomputers being built today are using non x64 technology which proves that not all problems can be solved with general purpose equipment. It requires a combination of technologies. Just as any company has a wide range of hardware and OSs for their needs. The same is true in the supercomputing world.

Virtualization and Businesses

2009-03-17T10:44:00.001-05:00

Virtualization is a big buzzword these days. I know because it's what I specialize in for work and blog about:) There are definitely some major misconceptions about it that I'd like to point out for people to consider:

Virtualization? Oh you mean VMware right?
You can virtualize anything and it's easy
It'll reduce the need for SA's because there will be less to manage
Virtualization product XYZ will solve all our problems out of the box

Well I hate to tell you, all of those assumptions are dead wrong. Like anything else in IT, "it depends". For example..

Many people have heard the "virtualization" buzzword, but automatically think the only solution is VMware. Obviously, it's been around longer than VMware has, look at VM on the mainframe. And virtualization is not just at the OS or hardware layer. You'll find it at the storage, networking, application, etc. layers. Each platform has it's own set of solutions, look here.

Many applications can be virtualized, this is true. However, there are apps that are not supported in a virtualized environment or will not work well. Not all commercial databases are supported in a given virtualized environment. Another example, it would not make a lot of sense to place a Netbackup media server into a virtual machine because it needs direct access to things such as tape libraries and any additional I/O overhead will cost you. You'd have to go for an option that allows for direct I/O access, such as Dynamic System Domains (DSD) or an I/O Domain in an LDom environment. Not something you could stick into a VMware or VirtualBox instance. So the key thing here is that the application support and I/O requirements decide the real options.

While the number of physical hosts can be significantly reduced in a virtualized environment, it does not remove the need for SA's. If anything, it places more emphasis on thier training and skills. You'll still have hundreds or thousands of OS and application instances to manage. Plus, the utilization, performance, and responsiveness of your virtualized environment will become paramount. Most standalone servers in traditional environments are only 10-30% utilized. So there is head room for usage spikes and for issues to grow in. In a virtualized environment, there are constraints that can cause contention for resources. As a result, an SA has to pay more attention to the metrics around their servers and adjust things as required to meat SLAs. It also requires SAs to know more. You're virtualizing not just the hardware resources, you're also virtualizing the networking and storage resources. So skill requirements in those areas increase.

And of course my favorate bad assumption is that some product XYZ will solve all your issues out of the box. Well I've seen many companies spend millions on software that promises to do all your work and make you dinner. Of course, it's just good marketing and never delivers what was promised. In the virtualization world, there is an investment in the hardware, the hypervisor (if it's not free like Sun's solutions.. zones, LDoms, and DSDs are free!), the management tools, provisioning tools, monitoring tools, HA components, etc. I have yet to see any virtualization solution that comes with everything out of the box, let alone not require some learning and configuration. A good example here is how people think that VMwares VMotion somehow will save the day. While it is good that a virtual machine can be moved around due to hardware failure or low resources, it doesn't do much beyond that. So lets say there is an application failure in your virtual machine, VMware has no idea about such things and even if you VMotion it elsewhere, the issue is still there, just on a different physical server. So the need for application monitoring or HA software does not disappear. If anything, the need for knowing what's going on in your environment becomes more important. So the reality here is it takes a combination of different solutions to build a truely robust virtualized environment.

Well I hope that gives everyone some things to think about and consider. Like I said, in IT.. it depends;)

UltraSPARC-T2 Reference Design Kit

2009-03-17T10:24:00.000-05:00

On the note of the lack of SPARC workstations, Sun released a Reference Design Kit for folks to build their own UltraSPARC-T2 solutions. Here is the link. The kit is basically a T5220 system board in an Extended ATX form factor, 16GB's of RAM, and reference materials to help you build your own boards. I took the time to inquire as to the cost of this design kit, because it would be kewl to build my own UltraSPARC-T2 workstation for home. Unfortunately, the starting price is $10K ! Yikes! So a big fat no to that idea!

Of course if Sun were thinking about building a nice workstation, I'd probably start with this board design and do the following:

Strink it down to ATX form factor, perhaps remove 4 DIMM slots, or place the iLOM onto a card. Reuse the Ultra20/24 case?
Go with 4 core UltraSPARC-T2 to bring the costs down.
Switch the e1000g chipsets to just two single-port ones. Don't need quad ethernet on a workstation, but dual is good;)
Remove the XAUI ports and add two more PCI-E slots.
Have it come standard with 2 x SATA hdds, sound card, 3D video card, and DVD+-RW/CD+-RW drive.
Enjoy!

I'd think something like that could sell for $3k easy. Oh well, enough with day dreaming about having an UltraSPARC-T2 workstation I could run LDoms on..

The Death of SPARC Workstations

2009-03-17T10:15:00.000-05:00

Well recently on the OpenSolaris desktop forums, the discussion of video cards for the upcoming OpenSolaris 2009* release has become a hot topic. Alan Coopersmith gave a list of cards that will be supported in Xorg:

XVR-50 (onboard graphics on some servers)
XVR-100 (PCI board)
XVR-300 (PCI-E board)
XVR-2500 (PCI-E board)
AST2100 (Service Processor/Remote KVM)

Of course people are not happy as this will prevent them from running OpenSolaris 2009* on their SPARC workstations. My response to all this is as follows:

Well there are definitely more folks out there with PCI and UPA video cards for their SPARC workstations. Realistically, many people are getting nice bargains on Ebay and putting together nice SPARC workstations. I know because that's how I've gotten by for the past 10 years and most of the people I've worked with during that time have done the same. If Sun would realize that there is a large market for such things and that people are paying less there. It should be a signal that Sun would stand to make a profit if they built a good solution at a reasonable price.

First off, it's a damn shame that Sun was never able to sell a descent SPARC workstation at a reasonable price to compete with the Ultra20/40. Second, the fact that Sun couldn't keep the SPARC workstation line alive, notice the only workstation now is the Intel Ultra24. Hmm.. I guess changing the stock symbol from SUNW to JAVA makes sense now. I think a lot of this stems from Sun's disconnect with the customer base which does want affordable performant SPARC workstations. When Sun was selling things like the Ultra25 or Ultra45, the prices were just as much as a low-end SPARC server, but with processors that were years behind. Not a good business model if you ask me.

What we're seeing here with OpenSolaris 200* is a real disregard for what makes money at Sun. It's not the x64 servers, it's the T-Series and M-Series servers which are obviously SPARC based. Notice there isn't a UltraSPARC-T* or SPARC64 workstation anywhere for developers, engineers, SA's, medical imagery, etc. I guess if you want to develop on current SPARC technology, you better work for a company that can spare a Zone or LDom in the datacenter for you. Oh well, most companies have taken away ppls *NIX workstations and forced them to use Windoze anyways it seems.

I've been very pleased with the progress of OpenSolaris 200*, but dissappointed that the SPARC side has languished until now. Of course, most of the benefits that have been delivered thus far have been around the desktop, installation, and package management. From this rather short and unimpressive list of cards that'll be supported on SPARC, Xorg will be pointless for most SPARC customers and users. People will end up sticking with SXCE or Solaris 11 when it comes out. Of course, the big question I have is, can Sun afford to keep all of these different streams of Solaris development alive and kicking? I get the impression that the majority of development and integration is on the OpenSolaris 200* side and that Solaris proper is just getting patched up with features from OpenSolaris until OpenSolaris can replace it. So if that's the long-term goal, what hope is there for customers and users on SPARC workstations?

Seeing how there won't be any new SPARC workstations, anything that's done to support them will be minimal at best. Other than waiting for the folks who have Ultra25's and Ultra45's to sell their gear on Ebay, not much of an upgrade path either. I have a SB2000 with dual 900Mhz procs, 4GBs of RAM, and a Expert3D card. It's a great workstation and I've enjoyed it a lot over the past few years. But seeing the "writing on the wall", I've just bought a used Ultra20 that'll be arriving today. I'm planning on upgrading the hell out of it and putting a AMD quad-core Phenom mobo and proc in there to get it kicking and screaming. Yes, I just wanted the case for the Ultra20! So I've given up on SPARC workstations, which is sad. But it seems it's time to just drink the x64 cool-aide and get on with things.

Well later on today when my Ultra20 arrives, I'll post some pics of it. It's sad that the company that use to be known for creating workstations only has 1 workstation to sell, the Ultra24. Not that it's a bad box or anything.. it's just not SPARC. But I've decided to move on and dive into the x64 world for my workstation/desktop needs.

BTW, I have lots to blog about and apologize for not posting for such a long time. Time to get 2009 rolling:)

LDoms 1.1 Details

2008-12-24T12:22:00.000-06:00

Our good friend Alexandre Chartre has a great blog entry with some details! It's an excellent read!

I think the ability to do Warm Migrations, enhanced dynamic reconfiguration, and enhancements to the networking stack are great reasons to upgrade to the 1.1 release. I'll have more details shortly!

LDoms 1.1 Released!

2008-12-24T11:45:00.001-06:00

I'm very pleased to announce that LDoms 1.1 has been released today! This is major update to the product. Here are the new features:

Warm and Cold Migration
Network NIU Hybrid I/O
VLAN Tagging Support for Virtual Network Interfaces (VNETs) and Virtual Switches (VSWs)
Link Aggregation Support for Virtual Switches (VSWs)
Virtual I/O Dynamic Reconfiguration
Virtual Disk (VDISK) Multi-Pathing and Failover Support
iostat (1M) Support within Guest Domains
Public XML Interface and XMPP Connection with Logical Domain Manager

Here is a list of other improvements:

Improved Interrupt Distribution (CR 6671853)
Performance improvements for virtual IO (Several CR's: 6689871, 6640564)
Solaris Installation to Single Slice Disk (CR 6685162)
Numerous Improvements and Extension to our Domain Services Infrastructure (CR 6560890)
Improved Console Behavior when not using Virtual Console (CR 6581309)
VDisk EFI Label Support for Disk Image (CR 6558966)
VDisk Support for Disk Managed by Multipathing Software (Veritas DMP, EMC Powerpath) (CR 6694540, 6637560)
LDoms Manager Improvements to IO FMA (CR 6463270)
ldm list -l now displays MAC assigned to guest (CR 6586046)
Improved Error Messages (CR's 6741733, 6590124, 6715063)
ldm list -o provides fine-grained control of configuration display options (CR 6562222)
More accurate utilization percentage reporting (CR's 6637955, 6709020)
Can now explicitly set a domain's hostid (CR 6670605)
Improved persistence of VIO service and VDS volume names (CR's 6729544, 6771264)
More predictable behavior when deciding which cpus to DR out of a domain (CR 6567372)
More accurate annotations in ldm ls-spconfig output (CR 6744046)
Better support for large, fragmented memory configs for a domain (CR 6749507)
Supports setting persistent WANboot keys from OBP (CR 6510365)
Lots of Bug Fixes (over 100)

Here are the links:

I'll have some blog posts about LDoms 1.1 soon!

Enjoy! Happy Holidays!

SPARC Enterprise T5440 Announced and LDoms Cookbook

2008-10-15T19:50:00.000-05:00

Finally, the SPARC Enterprise T5440 has arrived! It's nice to see this beast released into the market. It's amazing to think that in 4 rack units, there are 256 threads, 32 cores, 4 CMT processors, 64 FB-DIMM slots, and 8 PCI-E slots. Definitely a lot of power in a relatively small package. This would be a great fit for highly multi-threaded applications or for consolidating applications. With the use of Solaris Containers or LDoms, one could use racks of T5440's to migrate and consolidate many applications. Considering that most servers today are less than 30% utilized, this is virtualization power house.

In other news, Sun has started a Cookbook wiki for LDoms. Definitely a great start on a large topic that is becoming more complicated and challenging.

Sun Netra T5440 Released!

2008-09-30T22:52:00.000-05:00

Sun has announced a Netra version of the UltraSPARC-T2 Plus architecture, the Netra T5440. The first thing I noticed about this server is that it's not the four-way UltraSPARC-T2 Plus server that has been talked about, the T5440. This Netra server is only a dual socket system. Sure it has a lot of I/O capacity for telco requirements, but it's missing two processors! It should have been called something else. I have no doubt that when the four-way non-Netra T5440 comes to market, this will cause confusion for customers. Still it is an interesting server and is heavily influenced by Telco requirements.

Sun BluePrints on LDoms and I/O

2008-09-30T22:43:00.000-05:00

A good friend of mine at Sun, Peter Wilson has released two excellent Sun BluePrints on I/O best practices for LDoms:

LDOMS I/O Best Practices: Data Reliability with Logical Domains
LDOMS I/O Best Practices: Network Availab ility with Logical Domains

Both of these documents can help you understand the basic best practices around storage and networking for LDoms. Without a doubt these are some of the more complex areas, as there are many configuration options.

Check out the updated links on the OpenSolaris LDoms Community site!

Review: SPARC Enterprise T5140/T5240 and Virtualization

2008-09-30T22:40:00.000-05:00

I've completed my review of the T5240 with a focus on the hardware features and the ability to leverage virtualization. I think it will give people a good technical understanding of the hardware and help them brainstorm the possibilities.

Here is the link to check it out!

Sun Cluster and LDoms

2008-07-14T13:17:00.000-05:00

I have been searching for information on Sun Cluster and LDoms to see if I can integrate some info on my T5240 review. I came across this great wiki link on the subject. Here are some key requirements and things to keep in mind. I'll have to take a look at this and see if I can get it working this week!

George Carlin and Napalm

2008-06-23T23:32:00.000-05:00

It's definitely a sad day. We just saw George Carlin here in Houston this year. It was the second time that I've had the honor of seeing him perform live and it was amazing! He really opened the minds and ears of a society that was closed minded and unimaginative. He reminded America how to question authority and even our own social oddities. He rambled on about the mundane and even the most serious issues we still face today. Most importantly, he reminded America how to laugh at itself in the face of it's own shortcomings and failures, to place things into perspective.

"Frisbeetarianism is the belief that when you die, your soul goes up on the roof and gets stuck."
- George Carlin

Scalent and CA Support LDoms

2008-06-23T23:20:00.000-05:00

Scalent and CA recently announced the ability of their products to manage and integrate with LDoms:

It's good to see 3rd party products emerge to fill the void that currently exist with the LDoms product line. Hopefully, we'll see xVM OpsCenter take stage at some point in the near future to provide a native management tool. In the meantime, it looks like we'll see virt-manager first.

New Server to Review!

2008-06-03T22:08:00.000-05:00

Yesterday, a wonderful package arrived on my doorstep..

It's a T5240! I'll be writing a review of the server, with a focus on virtualization and practical uses. I'm also thinking of doing some sort of benchmarking to show the difference between my Sun Blade 2000 and the T5240. Keep your eyes out for more posts as I get started on this server!

Simulating the UltraSPARC-T2

2008-06-03T21:46:00.000-05:00

Two months back, I started playing around with the UltraSPARC-T2 development tools that are available. Included with the tools are the following simulators:

SAM (SPARC Architecture Modeler)
Legion (Fast Instruction Simulator)

These tools are intended for developers to simulate the architecture on other SPARC systems. I took the time to compile the tools on my Sun Blade 2000. It does take some considerable overhead to simulate a single UltraSPARC-T2 processor core. However, the learning experience is worth the effort. Both tools enable you to simulate the complete stack (processor, POST, hypervisor, and OBP), and enable you to run a guest OS. Definitely an interesting way to learn about the architecture under the hood!

Here is a screenshot of SAM:

Here's a screenshot of Legion:

Now that the hypervisor API has been released, I'll be playing with this more. Hopefully, I'll be able to get my arms around the programming and do something useful:)

Meeting the LDom Engineers

2008-06-03T21:18:00.000-05:00

So last week, I had a chance to meet some of the engineers behind LDoms in person! The meeting was amazing! Communicating with engineers at Sun is much easier thanks to the revolution of the OpenSolaris community. But having a chance to actually meet some of the engineers is always a pleasure. I talked with them about the future of LDoms and especially about our LDoms community. The good news is that the LDM software will be open sourced! So far, we have the processors, hypervisor, and the Solaris hooks available. This means the last piece of the puzzle will be available for developers and enthusiasts to participate through. I'm pleased with the discussions around opening up the development of LDoms. I also talked with them about ideas for the future and how to build out our community. I'll have more information as things develop.

Sad news though, Ashley Saulsbury, one of the engineers behind LDoms, has left Sun:( Definitely a sad situation. Needless to say, I think the development of LDoms is in good hands. Ashley was kind enough to publish the API for the hypervisor, before parting. I've setup a link on the community page. The document goes over the details of how the hypervisor, OS, and LDM software interact. So far, it's a very interesting read and definitely fills in the gaps with the other documentation out there today. Big thanks to Ashley for all of his dedication and hard work in developing this technology! I wish him luck with his future ventures!

LDoms 1.0.3 Download Link Fixed!

2008-05-20T09:02:00.000-05:00

The download link for LDoms 1.0.3 has been fixed on Sun's web site. Now everyone can start downloading!

LDoms 1.0.3 is Out!

2008-05-15T12:21:00.000-05:00

LDoms version 1.0.3 has been released. It will be available from the normal download site before the end of today, or through this link right now. I have updated the documentation links on the OpenSolaris LDoms Community page to have the 1.0.3 versions available. This release has a number of network and storage enhancements. These enhancements will require Solaris 10 U5, Solaris 10 U4 with the 127127-11 patch, or Solaris Express. I'll have more details shortly!

OpenSolaris.com

2008-05-05T17:22:00.000-05:00

Sun has released the results of Project Indiana as OpenSolaris. While this is has been a heavily debated decision amongst OpenSolaris users and developers alike, the release has been made. There are some very interesting components in this release, such as IPS, ZFS boot, new installer, etc. However, I must point out that this release does not cover the SPARC platform and does not take into account installation methods such as Jumpstart. This is definitely a different cut of the Nevada code and a divergence from the Solaris and Solaris Express methodologies. It'll be interesting to see how things play out as this new distribution evolves. I'll have to take this for a spin on my laptop or inside VMware.

Sun BluePrint on Using LDoms

2008-05-01T21:43:00.000-05:00

Here's an interesting article on leveraging LDoms for application servers and databases. The article has some interesting ideas and tools for doing the benchmarks. I might have to leverage those for some tests in the future!

One of the interesting things I saw in the article is the use of multiple VSWs for separating back-end and front-end traffic. This demonstrates one of the strengths of LDoms where it is possible to create virtual switches for different purposes. One thing of note is that multiple VSWs were used to isolate a single guest domain and front-end client for more throughput. However, it would appear that the utilization on the virtual switches easily allowed for increasing number of guest domains per virtual switch as was demonstrated when the number of guest domains was doubled.

I think this brings up some common misconceptions that I see regularly in the industry. Just as CPU utilization is observed to be less than 10% in many environments, the same holds true to networking. Many shops deploy 1GbE nowadays, but are under utilizing the capacity. The bad part being that many servers are provisioned with multiple 1GbE connections for different VLANs, such as production, backups, management, monitoring, etc. This leads to a huge waste in resources as ports and cabling increase substantially per server rack.

The use of LDoms can reduce some of this waste by sharing each NIC through VSWs. This allows multiple guest domains to use the same NIC and drive the utilization up. It'll be nice when Link Aggregation and VLAN Tagging become available for LDoms in Solaris 10. This will reduce the requirement of probe-based IPMP in guest domains and further consolidation of networks onto fewer physical links. With the ability to use to 10GbE on the T-series servers, this opens many possibilities. Just think how nice it would be to only have two network connections per server with all of your VLANs and guest domains running together on them.

Another interesting item in the article is the need to do performance and capacity planning for proper consolidation ratios for virtualization. I can't stress how important it is to look at the utilization (CPU, Memory, Network, Storage, I/O, etc.) of your servers now. Then test those workloads on a T5220 or T5240. If the application works well with a few threads or cores, you can size your guest domains. I do like the comments about looking for the resource factor that will limit the size of your guest domains. Some applications may need more memory than CPU for example. So the amount of memory can determine the number of guest domains for a given distributed application.

One of the common issues I'm seeing is that people have a hard time believing that a V880 can be replaced by a T5220. While single-threaded applications are a challenge for the CMT processors, fewer and fewer commercial applications are single-threaded. With the advent of highly scalable databases and Java applications, multi-threading is becoming the mantra for programmers globally. The other thing to consider is that older generation servers from Sun used PCI, while the current servers use PCI-E, which is considerably faster. For example, consider that an 8 lane PCI-E slot can handle 250MB/s per lane. A dual 1GbE card only has four lanes, but the Ethernet is moving at Gigabits, not GigaBytes. And given that most network ports are under utilized, you can see that current technology has a lot of head room for bandwidth. It's really amazing to think of soo much bandwidth being available on todays servers. Of course, there are always applications that can push that bandwidth utilization up. Think of the the need for pushing high volumes of data over networks or SANs for things like VOIP, video streaming, data mining, backups, etc. Definitely interesting applications that servers like the T5120-T5240 are designed to handle easily.

Fourth LDom Article Available to Community!

2008-05-01T21:36:00.000-05:00

My fourth article on LDoms is now available to the community here. For those who have not had a chance to read it or do not have a USENIX membership, the article wraps up my series of articles on LDoms. Focus is given to more advanced topics such as hardware, split pci, and running different operating systems. I highly recommend signing up for USENIX membership so you can receive the ;login: magazine and be part of the organization. Enjoy!

UltraSPARC-T2+ Servers Announced!

2008-04-09T19:02:00.000-05:00

Today was a big day! Sun announced the T5140 and T5240 servers which are the first dual socket UltraSPARC-T2+ CMT servers. This means that you can now scale out to 128 threads in a 1U and 2U form factor! The UltraSPARC-T2+ processor is a little different from the UltraSPARC-T2 processor. Here are some of the differences:

SMP enabled
Coherency links to support SMP
Only 2 FB-DIMM Memory Controllers instead of 4
Relaxed DMA ordering
Removal of the 10GbE Controller from the processor

As you can see, the ability to support SMP required some extra room on the processor. The coherency links provide the glue to enable SMP between UltraSPARC-T2+ processors. The number of FB-DIMM memory controllers is cut in half, but the ability to address a larger amount of RAM is enabled. The 10GbE controller is removed, and supplemented by a Neptune ASIC on the mainboard to manage the XAUI ports and the four onboard 1GbE ports. Altogether, not a bad compromise to support SMP. We still have the same number of cores and threads per chip, with the added benefit of having two chips!

The T5240 has the added benefit of double the number of physical RAM slots thanks to a mezzanine board that plugs in over the mainboard. We also see double the amount of disks on the T5140/T5240 from the T5120/T5220. This definitely allows a greater number of disks to be used for LDoms! Also, the other big change is in I/O. We have two PCI-E controllers, one in each CMT chip! We also have PCI-E x8 lanes across the board. This is an improvement over the T5120/T5220 which only had two x8 slots and four x4 slots. This increases the bandwidth available for pushing I/O. Here is a great blog entry on the subject.

Speaking of LDoms, we now have 128 threads to work with! Obviously, you'll want to use LDoms 1.0.2 to support these systems. Considering that once the control domain is configured, we still have 15 cores (120 threads) to work with for guest domains. Even if one were to break each core up into four guest domains, that's 60 guest domains! Performance wise that would probably be the smallest configuration you would want to go since each core has two execution units (each handles 4 threads). Any smaller and one would risk cache thrashing. Added with the ability to have 64GBs or 128GBs of RAM, one can definitely take on larger apps. I'm curious if a split PCI-E configuration can be achieved with these servers and if there is any benefit? Still this is very impressive for a 1U/2U form factor with such a small physical footprint.

Wish I had one of these to play with already! It would be fun to write a review of a T5140/T5240 server!

The next system of interest will be the T5440, which will be a four way UltraSPARC-T2+ system. That has to be a beast of a box!

Fourth LDom Article Out!

2008-04-06T19:15:00.000-05:00

The fourth and last article in my series "An Introduction to Logical Domains" has been published. Those who are USENIX members will be able to read it in the April issue of ;login: or through the website. After the run of the article, I'll make it available on the OpenSolaris LDoms Community site.

The fourth article covers topics such as hardware, cpu affinity, split PCI-E, DR, configuration management, HA, running other OS's, and even a comparison with other available technologies. Definitely a lot of topics to cover! I could probably write a book on the subject. The challenging part is keeping up with a technology that is evolving constantly.

Speaking of which, I spotted a T1000 on Ebay that was going for less than $200, but that quickly increased outside of my price range! I think it'll be a bit longer before I have a permanent LDom box at home.

LDoms 1.0.2 Released

2008-03-18T21:44:00.000-05:00

So the latest version of the Logical Domains Manager (LDM) 1.0.2 has been released. Also with that are patches for the ALOM/iLOM firmware. Check it out here. In the documentation, you'll see mentions of the T5140/T5240 which are the Victoria Falls servers. I'll have some more posts about this version soon, have to test it out first:)

More Info on the Victoria Falls Gear

2008-03-07T00:32:00.000-06:00

Looks like once again that The Register has info on the "next thing" from Sun. They have some basic info on the T5140 and the T5240, which will be the 1U and 2U dual socket UltraSPARC-T2+ servers due next month. While the core and thread count does not increase per CMT chip, the fact that we'll have a dual socket 128 thread SMP server in 1U and 2U form-factor is another big step forward for Sun. The T5440 will be the four socket version, so 256 threads! No doubt these servers will enable customers to migrate their larger databases and app servers to the CMT line. It also means that customers will be able to consolidate more apps into LDoms and Solaris Containers. The T5440, could support a boat load of LDoms. Even if one only divides each core into two guest domains, minus one core for the primary domain, that means a total of 62 LDoms! That is of course being conservative, since you could create 256 LDoms with one thread each. That is an amazing amount of virtualization in 4U. Hopefully, I'll have a chance to take a look at the UltraSPARC-T2+ servers at some point. Hopefully, this will drive the price of the T1000 down:)

T5120 Returned:(

2008-03-06T00:41:00.000-06:00

Last week, I returned the T5120. Very sad day indeed. Without a doubt, it was one of the most enjoyable systems I've gotten to play with other than a E25K and a M5000. Although, I could probably do without the extra noise, I'm glad I had ear plugs! It's definitely quiet with just the SB2000 at home.

I did notice that the price on the T1000 has dropped to just under $2800. While that is somewhat more reasonable, it's still outside of my price range. I don't even want to think about how much the T5120 I had would cost. I'm thinking about possibly doing a presentation on LDoms for a conference this year. Just have to figure out which one I want to go to and start doing the prep-work. The tough part may be in getting a box to bring along. I think the T1000 would be interesting since I could hook up some cheap SATA drives and the memory is also cheap. I'll just have to keep my eye on Ebay for a good deal on a T1000 since that's the low-end.

Third LDom Article Now Avialable to Community!

2008-03-03T11:51:00.000-06:00

My third article on LDoms is now available to the community here. For those who have not had a chance to read it or do not have a USENIX membership, the article covers advanced topics in networking and storage. I highly recommend signing up for USENIX membership so you can receive the ;login: magazine and be part of the organization. Enjoy!

Virtual Disks and LDoms

2008-02-10T03:49:00.000-06:00

Alexandre Chartre has a great blog post on the creation and use of virtual disks with LDoms. I definitely recommend it as a primer on the subject.

More ROCK Info!

2008-02-10T03:42:00.000-06:00

I was searching around to see if I could any new information on the ROCK processor and came across this great article that even has a block diagram of a ROCK core cluster! It looks like it'll be a beast of a processor and capable of some very interesting features. Check it out!

Review of the T5120 Available

2008-02-10T03:39:00.000-06:00

I would have had this available sooner. However, I went a little overboard and ended up with a PDF that was over 20MB's in size. Needless to say, I had to scale some items back and even take out some pictures. But here it is! Enjoy!

Almost done with T5120 Review

2008-02-08T07:55:00.000-06:00

Hopefully, I'll get the review of the T5120 completed today. I ran into some issues this week where our iMac died:( I do most of my multimedia stuff on that machine. Luckily, I can connect up my camera to the SB2000 without any issues and get to my pictures:) I'll probably post the review on the LDoms Community site. I'll have to put some pictures together our dead iMac. We'll just have to get a new one shortly!

Third LDoms Article Out

2008-02-08T07:49:00.000-06:00

My third article on LDoms is out! I'll post a copy on the LDoms Community site after the publication run. In the meantime, you can read it in print or at the ;login: site. The article has some very useful information on networking and storage. I used the T5120, Solaris Express build 77, and LDoms 1.0.1. Enjoy!

Why water and keyboards don't mix!

2008-01-27T10:38:00.000-06:00

Well the other night I was busy playing with libvirt on the T5120 and I took a sip of water. And the condensation from the class fell into my keyboard! Arrghh! I tried to dry it up, but soon noticed that I was getting extra letters when I pressed a single key!

So I decided to try and save my keyboard from doom as it's my only keyboard on hand. First, have to take the keys off:

Looking closely, I found the water had leaked further inside! Have to go deeper..

Arggh.. it's under the rubber membrane...

I found the water pooled between two of the film layers you see here. The film layers are pretty interesting, they look like a star chart:) I dried that up, cleaned things up inside, put the keyboard back together. And now it's working again! I probably could have waited for it to dry up on it's own, but I live in Texas and humidity usually doesn't leave the room quietly:) Still an interesting process.. I had never seen the inside of a Sun keyboard before!

Articles and other stuff..

2008-01-27T09:26:00.000-06:00

Last week, I finished my 4th article on LDoms for ;login:. I'm very excited about completing the series of articles. They have been challenging and fun to write. I intend to write more articles in the future, but I may focus on other technologies. I do intend to update and merge the LDoms articles into a single white paper later in the year. Right now, I have to finish my review of the T5120. Afterwards, it'll have to go back to Sun:( That'll be a sad day, I've had lots of fun with the T5120. I wish I could keep it!

That brings me to a topic which I think many SA's, engineers, and developers are probably all thinking about right now. "Gee, I wish I could have a box like that too!". I think Sun should consider making a UltraSPARC-T2 workstation that is affordable. With the UltraSPARC IIIi workstations stagnating and probably getting close to EOL time, Sun needs a new workstation to fill the gap. Having a UltraSPARC-T2 workstation that is under $1500 would be ideal. I don't see any reason why it wouldn't bring in a fair amount of volume from the user base. Many people in my field have Sun workstations at home to work or learn from. I use my SB2000 all the time, it's my primary machine! The biggest problem for people such as myself is the cost of getting a SPARC workstation. I got my SB2000 on ebay! So having something affordable and not completely server oriented would be nice. So here's my wish list for an UltraSPARC-T2 workstation:

Kewl enclosure. Perhaps something between an old IPX and a Mac Mini? Sun Logo that lights up (best feature of the SB2000 besides it's power)! Metallic shell, like the front of the T5x20's.
UltraSPARC-T2 with 4 cores
4 x FBDIMM slots (have to stick with the memory controller tech in the chip)
2 x GigE NIC ports
2 x SATA 2.5" laptop hard drives (gotta keep the space down)
1 x DVD-RW/CD-RW slot drive
4 x USB 2.0(3.0?) ports
1 x Nvidia video chipset with 2D/3D support, DVI port, no skimping here!
1 x Integrated sound card with all the normal in/out ports.
2 x PCI-E low-profile slots (combo XAUI slots?.. if it can be done within the costs)
Type 7 keyboard/mouse USB kit.
Silent fans!
External power brick to save space and reduce heat.

I think something like that would sell well and perhaps even kickstart Sun's workstation business. It could be called the SPARC Enterprise T5 Workstation. I would buy one! To have something that small and powerful would help developers out a lot as they could have something that's accessible for developing on.

Well enough with my day dreaming:)

A Peek Under the Hood of the T5120

2007-12-20T11:22:00.000-06:00

It's time to take a peek under the hood of the T5120. Lets start with some of the items on the front:

The drive sleds or spuds are pretty nifty! I like the new look and the latch mechanism. The grills on the front of the spuds, allows better air-flow. In the above photo, you can see a SAS drive in the spud on the left and a filler spud on the right. Lets take a look at the top:

On the top there is a panel that allows you to access the fan modules without having to take apart the server. Each fan module consist of two small fans with a carrier that has a connector on the bottom. It's very well organized and each module plugs easily into a slot. The chassis for the T5120/T5220 is the same that found on the current Sun x64 Galaxy boxes. So you can see there are other slots for extra modules, probably to keep those x64 procs nice and cool. As you can see, the UltraSPARC-T2 only needs 4 fan modules. Lets take a look at one:

How cute? If you look closely, you'll see that each of the modules that are plugged in have a green light. Meanwhile, the socket our module was plugged into has an amber light. I noticed that the chassis fault light went on and even a message was displayed on the ILOM console. One thing I found kinda odd was that the access panel can not stay open on its own and there is even a warning label that it might injure your hand if you are not careful. I used a pen to keep the panel open so I could take the pictures. When it's running, it probably makes sense that the door can't stay open as it would disrupt the air-flow.

Before we open the top, lets look at the instructions:

It's nice to see instructions and information like this on a server. Definitely makes it easier to service and prevents you from having to carry a printout into a data center! Hmm.. even the bottom of the lid has more instructions:

Well, lets get to it!

Finally, the inside of the T5120! Just infront of the fan modules, youc an see a clear plastic air duct that directs the air-flow over the DIMMs and the CPU. The T2 processor has a copper heat sink and the DIMMs seem to have heat sinks too. Beyond that, you can see the ASICS, PCI-E raiser slots, and the power supplies to the right.

I'll post some more pictures shortly:)

T5120 First Impressions

2007-12-20T10:59:00.000-06:00

It has been a couple of weeks since I've posted. I know:( The holidays are a busy time of the year. But enough with the excuses..

So the T5120 is setup in my home office. I had a hard time figuring out where to place it, but found that my storage bins (full of computer junk) works great! The machine is absolutely amazing!

It came pre-loaded with Solaris 10 Update 4, JES, and the LDM 1.0.1 software. Here is the hardware configuration:

T5120 ( 1 x UltraSPARC-T2 8-Core @ 1.4Ghz)
32GB RAM
2 x 146GB SAS HDDs
1 x DVD-RW

It's amazing to think that in a 1RU box, there are 64 threads on a single socket CPU!

The sound of the machine powering on is like a jet getting ready to launch off an air-craft carrier. It definitely moves plenty of air through the front! Obviously, it's not designed for home or personal use! Although it would be interesting to see a workstation version in the future. I'm sure many SA's, engineers, and developers would love to get their hands on a workstation with this feature set.

I'll try to make a video of the system booting so you can see and hear it:)

JET Module for LDM Software

2007-12-08T09:29:00.000-06:00

I took a few minutes to put together a JET module for deploying the LDM software. I've packaged it up and you can download it from here. When a client is being built, you'll see the following:

LDM: Installing ldm....
LDM: Installing ldm....
LDM: Register postinstall script 'add_packages' for boot 1

During the last phase, the software will install:

LDM: Running 002.ldm.001.add_packages
LDM: Installing SUNWldm.v from: /var/opt/sun/jet/js_media/pkg/ldm/1.0.1/sparc
LDM: SUNWldm.v installation complete
LDM: Installing SUNWldmib.v from: /var/opt/sun/jet/js_media/pkg/ldm/1.0.1/sparc
LDM: SUNWldmib.v installation complete

In the future, I'll have to see if I can squeeze the firmware and configuration pieces into the module. The firmware would be the more complicated piece. For now, it just installs the software.

New Arrival...

2007-11-16T09:03:00.000-06:00

So a package arrived from Sun...
Hmm.. what could it be??

Why it's a Sun SPARC Enterprise T5120!! For the next few weeks, I'll have this wonderful machine to test, blog about, and use for writing another article;) So expect to see lots of posts over the next few weeks!