At a recent press event, Cisco CTO Padmasree Warrior said the networking giant wasn’t looking to undercut the existing server market leaders–Hewlett-Packard and IBM–with its forthcoming virtualized blade servers and Unified Computing System strategy. Rather, Warrior said UCS was about building better, next-generation data centers.
Obviously a “next-generation data center” will mean displacing the existing data center that’s dominated by HP and IBM servers, but Cisco has been careful in most cases to avoid calling out its long-time server allies.
HP, however, isn’t being so cordial. The owner of one-third of the global server market has published a scathing assessment of Cisco’s UCS strategy, titled “The Real Story About Cisco’s ‘One Giant Switch’ View of the Datacenter.” The self-published article goes into great detail in a blunt critique that boils down to UCS being too complex, inefficient, insecure, costly and, most damning, designed to just sell more Cisco gear.
The article states:
“In a network centric view of the datacenter the network admin rules it all. Cisco’s so called unified computing requires a change of control where all data center traffic is now piped through the center of the network adding yet another coordination touch point. This scheme requires investing in a very large, complex, and expensive networking system, similar to ‘one giant switch’ in the middle of the data center.”
HP goes to great lengths in its assessment to pick apart the UCS strategy on points ranging from difficulties in purchasing UCS equipment to technology complexity to cost and insecurity to lack of scalability to customers being locked into a single vendor platform.
While Cisco has said that UCS is designed to leverage the efficiencies of virtualization to create data centers with higher resource utilization and small physical and energy footprints, it hasn’t shied away from the complexity issue. UCS, which is still mostly concept than a product, will only be available initially direct from Cisco and through a handful of large integrators. Cisco says specific capabilities are required to design and install UCS.
HP not only picks up on that theme, but says the actual infrastructure design of UCS is inefficient because it will require data on one virtual server to exit the physical machine, transverse a network switch and return to the same physical machine to reach another virtual server.
Unifying the whole of data center technology as Cisco describes sounds good in theory, but HP says the security risk may be too great. While HP doesn’t directly poke at UCS’ security, it does quote from reports that say the unified data center would essentially create a giant single point of failure that, if cracked, would give a hacker complete access to the overall infrastructure.
HP even speculates on the ability of businesses to absorb UCS because, beyond technology, it requires changes in management processes and operational organization. As HP describes in its article, network managers run the network; server managers run the server–a new breed of admin or cross-over management will be required.
The HP article soundly criticizes UCS for being overly dependent on proprietary technology, such as the Nexus switch, and closed-standard protocols to work. The result, it says, will lock customers into a single platform that will be expensive and difficult to scale.
“Cisco is attempting to make themselves, through the Cisco network, the control point for everything in the datacenter using an unproven proprietary approach requiring massive investments and difficult migrations by customers,” the article states.
HP does, as expected, position itself as the alternative to Cisco’s UCS strategy, calling its ProCurve servers and distributed data center products affordable, scalable and proven.
Analysts have speculated on the impact UCS will have on the long-standing alliances between Cisco and HP, IBM and Dell. This HP article clearly shows that HP, for one, is not taking the threat to its server business lightly.