Bladewatch.com

HP.com

Type:   Firmware – Lights-Out Management
Version:  1.70 (5 Dec 2008)
Operating System(s):     Microsoft Windows Server 2003, Microsoft Windows Server 2003 for 64-bit Extended Systems, Microsoft Windows Server 2008 W32, Microsoft Windows Server 2008 x64

I was talking with Chris about upgrading the ILO on DL380 G4′s the other day and he mentioned a newer versio for the ILO2 card on the newer HP servers do check it out, firmware upgrades are often mandated by first line support and can include useful updates/fixes. This url has the FAQs.

Share and Enjoy

Businesswire

MANCHESTER, England & BEAVERTON, Ore.–(BUSINESS WIRE)–VirtenSysâ„¢, Ltd., announced it has delivered pre-production units of its I/O Virtualization (IOV) Switch to leading server and storage customers. VirtenSys IOV systems dramatically reduce data center operational expense and complexity by virtualizing all the I/O resources used to connect servers and storage platforms to the networks. This results in reducing rack and blade server management costs by more than 60 percent, power and cooling costs by up to 30 percent and equipment costs by as much as 50 percent. VirtenSys is the first-to-market with IOV products that use the native PCI Express® bus available in every server. The VirtenSys solutions enable the best price/performance and non-disruptive deployment of server I/O connectivity in the data center.

I/O Virtualization Switches drive the evolution of server and storage platforms into more efficient IT infrastructures where all the resources are virtualized, dynamically allocated on demand, independently scaled and optimally utilized. Data center configuration and management becomes a remote, automated process that eliminates physical re-configuration or human intervention. I/O Virtualization also enables the rapid adoption of new usage models and IT services such as green IT, cloud computing and software-as-a-service.

We need to continue the innovation of the IT platform, reduce the barriers to success and ‘add value’ if you like to the end user community, the business so to speak. Using virtualization technologies we can empower a more flexible infrastructure, a more on demand IT. It will be a a two way thing, we need to virtualize the application and the infrastructure. In the infrastructure we should be looking at:

• Server virtualization – commoditize the server, consolidate do more with less
• Storage virtualization – be able to allocate storage more on demand using whatever storage we have and present it where we need it
• Network virtualization – to be able to plug the server in and put it on whatever network we need without having to change the physical patch (as much as possible)
• I/O virtualization – provide the network and the storage down the one pipe, one pipe, one team doing the patching and port allocations – anything we can do to end those delays, sorry we’re waiting on storage/networks allocating the ports…

On the application side we need to be doing everything we can to abstract our application from the infrastructure, the IP, the server using effective coding and appropriate technologies like:

• Citrix/application virtualization
• HPC/Grid technologies
• Web technologies

Removing or reducing the client role in the application, the user simply clicks a url, or double clicks an icon, a thing which runs the application within an exe. Reduce the amount of client pc configuration/support.

Share and Enjoy

Bit-Tech.net

Intel’s six-core Dunnington Xeon CPUs have already been doing the rounds since September 2008, but AMD’s rival Opteron processors have now finally caught up. The company recently demonstrated its own six-core ‘Istanbul’ CPUs, and the demo included showing off a 24-core, quad-socket server based on the new CPUs.

The Tech Report attended the demo, and revealed some interesting findings. The first factor to note is that Istanbul CPUs will be fully compatible with existing Socket F Opteron systems. To perform a drop-in upgrade to Istanbul processors, the server’s motherboard will just need a BIOS update and support for dual power planes.

Secondly, the performance of the new CPUs appears to scale well, more so than you would expect from the simple addition of more cores. AMD compared a 16-core system based on four Shanghai-based Opterons with a 24-core system based on four Istanbul-based chips, on which you could see all 24 cores in Task Manager on Windows Server 2008. To show the difference between the two machines, AMD ran the Stream benchmark on both machines, in which the Shanghai system managed a throughput of around 25,000MB/sec, while the Istanbul system managed a highly impressive 42,000MB/sec.

I was talking with a colleague about processor improvements and what might be coming soon and he mentioned this article talking about AMD’s new processors, it’s an interesting read, do check it out.

Share and Enjoy

Computerworld

February 9, 2009 (Computerworld) The U.S. government has awarded IBM a contract to build a supercomputer capable of performing at 20 petaflops, which is more powerful than all of the systems on today’s Top500 supercomputer list combined.

Terms of the deal were not disclosed.

Nuclear researchers at the U.S. Department of Energy will use the planned Sequoia system, to be built in IBM’s plant in Rochester, Minn.

The fastest systems today perform at a little over 1 petaflop, or 1 thousand trillion floating-point operations per second. The single-petaflop barrier was passed last May by IBM’s Roadrunner supercomputer.

It’s always great to see how we can continue to improve performance, what new possibilities and opportunities we can create through hpc and grid technologies.

At the same time, it’s aways cool to see what range of technologies are being used to achieve this, 10GB Ethernet? PowerPC or Cell/B.E etc. Granted the application needs to be coded to gain the most benefits from the platform, but then that’s always going to be the case, do check out the article. I wonder if this is based on any paticular technology like their BladeCenter products?

Share and Enjoy