Cisco's Unified Computing System announcement yesterday -- which was one of the worst kept secrets in the industry -- was a move Cisco probably had to make. Why? Several reasons. Let's take a look at some of them.
Networking market. Look, Cisco owns the networking market...and has for years. Their numbers speak for themselves. While they may be able to squeeze a few percentage points out of various switching and routing markets (WAN, storage, Wireless, etc), there's a point of diminishing returns. Cisco has an obligation to return value to shareholders....something their executive management is very good at. The networking market simply isn't a place where they felt Cisco had a significant revenue growth opportunity.
Competition. HP's position in the server market puts Cisco's data center Ethernet strategy at risk because HP sits atop of the x86 server market (ProLiant), has iSCSI ambitions (see LeftHand acquisition) and -- most importantly – have a viable and growing ProCurve business. This means HP can do without Cisco I/O modules in their blade chassis, but can also drive 10Gb lossless Ethernet market share. Not in Cisco's best interest.
But most importantly, Cisco has noticed -- correctly -- that the data center is a market in transition. Obviously, network, storage and computing are converging. Server virtualization has catalyzed this convergence. But Cisco believes that the next-gen data center is being held back by poor resource utilization and a lack of mobility/agility. Most of us agree that workload mobility is a key to a truly dynamic data center. But Cisco asserts that workload mobility needs a ubiquitous converged fabric, with rich network services, that can connect compute and storage (even across data centers -- think internal and external cloud). Cisco felt that rather than sitting back and have the competition define this new data center era, and have the market dictated to them, they would seize the opportunity to drive the dynamic data center and gain a greater share of the IT budget.
Thus, Cisco created UCS. Looking at UCS, they have an interesting solution. Their modular design is unique in the way that they tie their backplane with the I/O modules, but also how they unify chassis with the network. Cisco also stated (in the video announcement today) that they believe they can reduce capital and operational data center costs around 30%. This means they believe that UCS has a competitive advantage in terms of price (at least). But the question is, how long will this innovation last? Is USC innovative enough that Cisco felt like they needed to get into the server market? No lead in the x86 server market lasts very long. Competition in this space is extremely tough. Technological leads last 3-6 months at best.
On Cisco's data center blog, Doug Gourlay stated that Cisco is creating a sub-category of the server market- a specialized market we call Unified Computing. This is not the ‘Clash of the Titans’ or us ‘coming after HP or IBM or Dell’.
Pure hogwash. What Doug is saying is that Cisco has created it's own server market segment, and thus, really isn't in competition with the other blade server vendors. When you produce a product that is aimed at taking customers away from your competitors sweet spot (even if you believe your product is in a class by itself), you are competing. Make no mistake, Cisco has a target on their back.
There are some go-to market-issues too. Audience is a key issue. Cisco has the network administrator's mindshare...no doubt. But what about the server administrator? IBM and HP have significant mindshare with server administrators. How will Cisco overcome that barrier and gain the trust of server admins everywhere?
Furthermore, the market is full of customers with existing gear. In many data centers, it will be a tough proposition, especially in this market, to convince IT organizations to switch vendors. At best, IT organizations may slowly replace their existing infrastructure with Cisco's UCS during product lifecycle upgrades. But in many circumstances, IT organizations think more holistically when choosing a compute vendor. IT organizations look at the vendor's product portfolio breadth at then make a vendor choice. The lack of product breadth (x86 rack servers, mid-range servers) may be problematic going forward.
Some may ask, why didn't Cisco partner their way to a blade server solution? Well, perhaps they could have. Dell or Sun might have been willing partners (Dell is already a Cisco VAR channel and FCoE partner). But IBM and HP -- the leaders in blade space -- were not a realistic option. When Cisco gets into a market, they intend to dominate. Choosing the #3 or #4 blade server vendor as a partner probably isn’t what Cisco would deem as a winning strategy.
In the end, UCS was a move Cisco had to make to ward off competition AND increase shareholder value. Cisco has a strong brand, enterprise credibility, the technical chops and finances to pull it off. Is UCS a business risk? Sure. But the greater risk for Cisco is to do nothing.
[posted by: Drue Reeves]