After months of speculation, rumor and hype, Cisco finally announced its Unified Computing System (UCS) solution that most people know as the codename "California."Cisco's UCS moves the company into a new market, obviously known as "unified computing" (UC). UC is an emerging opportunity that effectively ties the network resources to server resources to allow for better management of both virtual machines (VM) and physical applications. This is being driven by the need to run and mobilize mission critical, real time applications like unified communications (the first UC!) on virtual machines and the need for a more dynamic, agile enterprise.
Based on much of what I read to date (much of it incorrect), I think it's important to clarify what Cisco has announced today, what it isn't, and what the industry implications are.
The Cisco UCS solution is made up of three broad components (each of which has smaller components) -- network infrastructure, blade servers and management software. All three are needed to fulfill on the vision of unified computing. However, much of the media focus has incorrectly but understandably been focused on Cisco's competition in the blade server market. The intention of this release is not to move Cisco into the same competitive space with the commodity blade server vendors. If a customer wanted to purchase a vanilla blade server to use as a file or web server, the UCS servers would not be the right choice.
What UCS does is create an environment where a VM can be moved from one server to another, in real time without anything breaking. In a high performance data center, communication between servers (including the movement of VMs) cannot tolerate the latency of IP so Layer 2 is used much more often, losing much of the dynamic nature of IP. If a server manager was to move a VM from one physical server to another, the network manager may need to make a number of changes to VLANs or other network policies to keep the application up and running. This, of course, can be cumbersome to impossible for some organizations, depending on the critical need of the application, and it has caused many companies to redesign their data center networks.
UCS allows the IT manager to move the VM through a management interface, and then all the changes required on the server and network are done automatically, allowing the application to continue running without interruption. This will allow companies to run more VMs for more applications. This concept is also important to the growth of cloud computing, where pieces of infrastructure will be used in various locations to deliver an application.
Just as powerfully, this concept can be applied to the infrastructure policies needed to deploy applications -- such as server identity, firmware, and Ethernet and storage networking. These policies are encapsulated into a construct called a service profile, which can be provisioned onto any blade in the environment. When the service profile is moved from one blade to another, all policy and identities follow. This both complements a virtual environment and brings mobility to a physical environment.
Now that the need has been established, it does raise the following questions: What does this do to Cisco's relationship with the other compute vendors, most notably IBM and HP; and what will the industry response to this be?
On the competitive front, this only further intensifies the competition with IBM and HP, who are significant channel partners. Although Cisco isn't competing directly for generic server business, they will compete for much of the high performance business, driving a wedge between these companies. In conversations with both IBM and HP, it's clear neither is happy Cisco is moving into their markets, and they also expect the competition to intensify and for Cisco to lessen its reliance on HP and IBM as channel partners and for IBM and HP to lessen their reliance on Cisco as a network vendor.
HP's response is obvious. They've rolled Procurve into TSG and have made claims that Procurve success now will be measured on Cisco share gain. HP also owns all of the assets to build their own UCS type of product and from what I understand, that product is well underway, to be announced later this year. Make no mistake, though, that the war between HP and Cisco had already begun prior to this UCS announcement. Procurve has been throwing daggers at Cisco for the last couple of years now, and that's been magnified since the integration of HP into TSG.
IBM's response is less obvious since IBM doesn't have network assets. However, what I've noticed in conversation in my channel checks is that IBM has been leveraging both Foundry and Juniper more for network infrastructure. Juniper's Stratus announcement earlier this month outlined how it plans to play in the emerging UC space and declared itself open and willing to work with partners, one of which will be IBM. This will be a much tougher path to go down, in that it requires coordination between multiple vendors and joint product development. Long term, it may have benefits, but by then I expect both Cisco and possibly HP to have gone down the same path but will have the early mover advantage.
The other vendor I expect to see something from--and I'm a little surprised I haven't--is the combined Brocade-Foundry. Here's a company that's arguably the technology leader in computing and network connectivity and has yet to stake their claim in the UC market.
For the rest of the network industry, Nortel, 3Com, Extreme, etc the outlook is bleak, but the outlook has been that way for a while. This will be hard to respond to, but their survival has been in doubt for a while. 3Com may be the only exception due to its H3C business which has allowed it to create a billion dollar business and round out its portfolio.
Now, before you pass judgment and think there's no way Cisco can pull this off, remember (quoting Caprica Six on Battlestar Galactica) that "all of this has happened before and will happen again." Cisco aggressively moved into voice business by catching the transition to IP. Many thought they would fail and they didn't. The transition to virtualizing the data center is equally big in the computing market as IP was to VoIP. What helped Cisco in the voice business is that the rest of the industry was caught sleeping, so we'll see how fast the rest of the data center industry responds.
The only other big question in my mind is how Cisco accounts for the integration and channel fulfillment that HP and IBM provide. If you've noticed in recent months, Cisco has announced expanded partnerships with Tata and Accenture, and at EMC's investor conference last week, EMC highlighted its growing relationship with Cisco. So I expect Cisco to methodically move more and more of its business to other areas to protect itself. However, realistically in the short term IBM and HP need Cisco and Cisco needs IBM and HP; but long term, expect these relationships to continue to change and become less important to both sides.
People will remember March 16th as the day Cisco entered the server market, but the gauntlet between these companies had already been thrown down. It will be an interesting next couple of years, and because I feel VM mobility is something people want and is a problem to solve, Cisco will be successful here. However, expect to see responses from the competition. Game on!
