I’m quite pleased with the ramp of the Intel Xeon processor 5500 series. As reported during the Nehalem-EX briefing (more info here), we expect the Xeon 5500 to reach more than half of our 2S server shipments by August 2009. All of our key OEM customers have embraced the new architecture with complete product offerings, which provides IT administrators a plethora of Xeon 5500-based systems to choose from.
For this blog, I am focused on data center advancements that Cisco is pioneering with their Unified Computing System (UCS). They intend to combine best practices for network infrastructure with data center virtualization. The Intel Xeon processor family, Intel Virtualization Technology and Datacenter Ethernet are foundational elements to Cisco’s strategy.
Already the industry has recognized the Cisco UCS blade platform with awards such as “Best of Interop 2009” for Data Center & Storage (link) and “Best Data Center Innovation” from BladeSystems insight (link). It is a bold move on Cisco’s part, and makes a lot of sense in light of the convergence of servers, storage and networking. See a video testimonial about Cisco UCS and Fiber Channel over Ethernet (FCoE) from our own Intel CIO, Diane Bryant, here.
Recently, Cisco announced further expansion of the UCS product portfolio with addition of three new rack-mount servers. All of them are based on the Intel Xeon processor 5500 Series and are expected to provide compelling performance, memory expandability and integrated virtualization capabilities. They are expected in the fourth quarter. You can read more about the new rack-mount servers here.
Have you had a chance to evaluate the new Cisco offering yet? Have you made any plans to deploy in 2009? What do you think of unified fabric and the concepts that Cisco has put forth? Let me know!
-steve
One of the key elements of the Cisco UCS story is that much of the economy of the solution is achieved by reducing the cost, complexity & power consumption of the network infrastructure elements that interconnect the servers. Coming from the industry leader in networking, this is indeed innovative thinking! Beyond the unified fabric which combines storage traffic over Ethernet using FCoE, Cisco has reduced the number of switches required to support the UCS system down to one switch (2 for redundancy) for the entire "pod" of servers (up to 320 physical server blades). These switches are expanded with proprietary low-cost "fabric extenders" which provide the needed connectivity to the server blades and hand off the data flows to be switched by the centralized pod switch. This architecture provides all of the needed network functionality at the lowest possible cost. This type of network innovation is not limited to a UCS solution, though. How will you reduce the overhead costs of your network infrastructure?