Cisco says the rollout represents its third-generation fabric computing platform and addresses data center scale and promptness in responding to changing business needs.
As expected, the new servers in Cisco’s Unified Computing System (UCS) fabric computing platform support Intel’s new Xeon processor E5-2600 line – also known as “Romley” and “Sandy Bridge” -- and includes multiple form factors, up to eight times the memory capacity and four times the I/O of previous UCS servers.
Also, the UCS Manager now supports Cisco’s UCS rack mount servers, enabling those form factors to reach management parity with the UCS blade servers. Cisco also added fabric extenders, interconnects and I/O modules to support the new servers.
Cisco says it now has 11,000 UCS customers since the platform’s introduction in 2009. And at a $1.3 billion annual run rate, UCS is the fastest growing product in Cisco’s history, company officials said.
The new servers include one blade and two rack-mount units, all based on the Intel Xeon E5-2600 processor. The Cisco UCS B200 M3 Blade Server comes in a half-blade form factor supporting 24 DIMM slots and up to 80 gigabits of I/O bandwidth.
The UCS C220 M3 Rack Server is a one-rack-unit (1RU) unit, targeted at business workloads like Web services to distributed databases. The UCS C240 M3 Rack Server is a 2RU server designed for storage-intensive workloads, from big data to collaboration.
The new rack mount servers and Cisco’s existing rack mount units can now be managed by Cisco UCS Manager. UCS blade and rack mount servers can be managed uniformly within a single domain by UCS Manager, Cisco says.
In the second half of 2012, Cisco says it will deliver the capability for UCS Manager to control multiple UCS domains, which would improve scale. The centralized manager will be able to govern and orchestrate thousands of servers in or between global data centers, Cisco says.
Naturally, the network has to support these increasingly virtualized and distributed workloads. So Cisco rolled out the chassis I/O module 2204XP. It offers 80Gbps and 160Gbps down to each chassis to handle workload bursts. The module also offers load balancing across all ports.
Supporting it is the VIC 1240 interface card, which connects the server to the chassis I/O module and delivers the 80G to the server.
Another enhancement is the UCS 6296UP Fabric Interconnect, which is used to interconnect Nexus fabric extenders to scale UCS deployments. The 6296UP doubles the switching capacity of the current UCS fabric from 960Gbps to 1.92Tbps. Cisco says it also reduces end-to-end latency by 40% (under 2 microseconds), watts per port by 36% and supports Unified Ports – Ethernet, FibreChannel or FCoE – on all ports.
Read more about data center in Network World's Data Center section.