A major refresh of the Cisco Unified Computing System (UCS) unveiled today relies on much higher levels of embedded IT automation to make it possible to manage denser unified compute platforms at scale.
Todd Brannon, director of product marketing for Cisco, says the M5 iteration of the Cisco UCS makes use of the latest Intel Xeon Scalable processor series that is optimized for systems based on the NVMe backplane technology Intel developed. Cisco says its tests show that the M5 Series platform is 86 percent faster than previous generations of UCS. Usage of NVMe is also making it possible for Cisco to add support for two graphical processor unit (GPUs) that can be deployed in the same system as Intel Xeon processors.
In addition to taking greater advantage of NVMe, the M5 series of blade and rack servers also makes use of 25G Ethernet networking to better integrate compute, storage and networking. To make it simple to manage various nodes within UCS, Cisco today also announced Workload Optimization Manager module with version 6.5 of Cisco UCS Director, the management platform Cisco embeds with UCS. That module analyzes the characteristics of any given workload to better match it to different classes of compute nodes that any given UCS system might have available at a specific time. Version 6.5 of UCS also makes it simpler to move virtual machines across nodes using software from VMware and automates a variety of Powershell functions.
“Our approach is to wrap our servers in software,” says Brannon. “That’s what sets us apart.”
Thank You
Nagesh Kumar
