Cisco announced its new M5 generation of the Cisco Unified Computing System [Cisco UCS] last week concurrent with the announcement of the new Intel Xeon Scalable processors inside them. As it has previously, Cisco emphasized the modular unified computing element of its systems, with the tagline that the UCS is a system rather than a server. They highlighted the software innovation. And they stressed how all of this presents a strong opportunity for their channel partners.
“The Cisco UCS has been built around a couple key principles,” said James Leach, director of product management for UCS. “One is that it’s a system, not a server, focused on the end user goal of performing work, with the workload driving design. The second is that because it’s a system, it has modularized components, to future proof the system to add components without requiring a forklift change.”
The M5 roster has five new servers, three rack and two blades, a reduction from the last generation.
“Intel launched a single stack, which enabled us to have fewer servers in the portfolio with a much wider reach,” Leach said. The new Intel processors facilitate up to double the memory capacity of the M4. Cisco also said its lab testing indicate the M5 servers have up to 86 per cent higher performance than the M4.
The rack servers are:
The UCS C220 M5 Rack Server, a two-socket general-purpose enterprise infrastructure and applications server with 56 cores, 24 DIMM sockets and support for 10Gb Lights Out Management [Lom] systems and optional M.2 format internally mounted expansion cards.
The UCS C240 M5 Rack Server, another two-socket server, this one optimized for storage and I/O. It has 56 cores, up to six PCIe 3.0 slots, up to 3TB memory, and support for 10Gb LOM systems as well as the optional M.2 cards.
The four socket offering is the UCS C480 M5 4S Rack Server, for more demanding applications like in-memory databases. It has 112 cores, up to 6TB addressable memory and supports up to six GPUs, 12 PCIe 3.0 slots and 32 drives.
The two blade servers are the UCS B200 M5 Blade Server, a half-width blade form factor which supports up to two GPUs, with up to 20.5TB of storage, and the beefier UCS B480 M5 Blade server, with up to four GPUs, 56 cores, and up to 39TB of storage.
“We have added an increased amount of storage options, with up to 32 Small Form Factor drives to address customers who said they had more need for local storage in some workload bays,” Leach stated. “We also added some more density around GPUs. We have also done a lot of work around our integrated management controller with our own proprietary IP.”
On the software side, the big news is Cisco Workload Optimization Manager, powered by Turbonomic – formerly VMTurbo – which makes an autonomic hybrid cloud management platform that lets both on-premises and public cloud self-manage in real-time.
“Turbonomic is designed to get ahead of thresholds, and does this by converting the data centre into a marketplace,” said Joann Starke, senior manager for software automation data center and solutions group at Cisco. “Their analytics engine interprets monitoring data from workloads, and when it sees a change in supply or demand it changes it autonomically, without the need for administrator intervention.
“With Workload Optimization Manager,” Cisco brings a deep and rich integration between Turbonomic and the Cisco UCS chassis, blades, IO modules and interconnect,” Starke added. “It allows IT to do planning with a built-in toolset, so they don’t have to overprovision. It also works in the public cloud as well as on-prem, to keep cloud bills from overprovisioning down.”
In addition, Workload Optimization Manager is integrated into the latest release of Cisco UCS Director 6.5.
“Cisco UCS Director 6.5 also takes automation beyond basic infrastructure by running natively on PowerShell,” Starke indicated. It also includes automation improvements for FlexPod, Cisco HyperFlex, and added support for the UCS M5 series and UCS S-Series servers.
All of this is good news for partners, said Scott Mohr, director for the global partner organizations, solutions, architecture and engineering for the data centre at Cisco.
“We believe this gives them an opportunity to improve their penetration rate with customers,” Mohr said. “It gives them an opportunity to go back to their accounts and refresh them as well as increase penetration in HyperFlex. We believe this will lead to improved TCO and application performance. It also opens up cross-selling opportunities with the security portfolio teams around secure data analytics.”
Mohr also stressed new opportunities created on the graphics front, facilitated by Cisco’s partnership with NVIDIA.
“There’s a lot of graphics-intensive apps being consumed now, and partners are seeking out new ways to help,” he said. “The increased GPU densities here will accelerate customers’ graphics-intensive application initiatives, and will provide a better user experience for VDI.”
Mohr indicated that Cisco has already done a lot of work to get all the necessary support for the launch or partners.
“We are refreshing all content, updating bundles and adding M5 to all channel programs,” he said.