Cisco is making a big play in the cloud, or for the cloud, with CloudVerse, an architecture that in some ways has the same goals as the one that Alcatel-Lucent announced last month (CloudBand). Everyone is trying to get a place in the cloud, hoping in part no doubt that the inherent fuzziness of the cloud will admit them. Cisco, though, has more credentials in the space than most and so you have to take their approach seriously.
In functional ingredient terms, everyone has to build the same cloud. Clouds are a system for creating a sharable (multi-tenant, in public cloud instances) pool of resources, of assigning those resources to applications/users, and of managing the process to insure efficiency and cost-effectiveness. In practical terms, what separates cloud players is what they can offer that’s special within this mix and how broadly they can cover the missions. The former gives them differentiation and may assign a service target (one that values the special something) and the latter gives comfort to buyers who doubt their own skills. These days, that’s pretty much everyone.
Cisco’s secret sauce in the cloud is completeness of solution, in my view. Cisco has every piece of technology you need for cloud-building, including the servers. Where Alcatel-Lucent focuses on data center interconnect because it can’t populate a data center, Cisco can articulate a simple story. Clouds start with big resource pools that we can build, and grow out through big networks that we can also build. Yes, Cisco needs to articulate the cloud story better to align its capabilities with the real business trends, but they have an easy story to create and sell. If they get good at it, it will put a LOT more pressure on competitors to do something smart on their own, and precious few network vendors can realistically match Cisco’s scope. With HP in management ruins and IBM out of networking at least in terms of having their own products, Cisco may be the only one-stop shop in the cloud mall.
Arch-rival Juniper is taking the other approach, the “cloud secret sauce” with its new drive to promote data center fabrics as the logical heart of the cloud. This story also has inherent strength since clearly a cloud has to start with a resource pool, but like Alcatel-Lucent’s CloudBand the Juniper tale needs more collateral to be as effective as a potential Cisco story. The challenge for Juniper is articulating a cloud position that their fabric can be the centerpiece for, as Alcatel-Lucent’s is to articulate the overall scheme of cloud services that justifies and empowers their interconnect vision. What this says is that cloud positions have to be holistic; if they aren’t naturally complete in a product sense then they have to be complete in a vision sense. That fits with the fact that cloud prospects, particularly network operators or other public cloud-builders, want a strategy that they can drop in with a minimum of extra integration effort and delay.
I think everyone in the cloud space is minimizing the most important truth about the cloud, and that’s the fact that the cloud isn’t about an alternative to enterprise data centers at all. The real notion of the cloud has been transformed by usage and expectation now, transformed into a vision of a flexible and elastic and universally available pool of knowledge and compute power. Yes, in theory, that new pool could be used to implement the old stuff, but more than that it could transform how we use computing and information, coupling it more tightly to our behavior. Yes, I know this is my usual “mobility/behavior” theme, but the fact is that where work and workers can be concentrated into buildings the notion of ubiquity of availability of IT resources becomes rather lame; they already are by virtue of having limited the scope of the use of the resources. What makes the cloud different is that it’s not requiring massing up humans to justify it; instead it’s presuming they are distributed and that’s clearly the case.
The jury is still out on just what “distribution vehicle” will drive all of this, or whether a single driver is even necessary. You can say this is about smartphones or tablets or game consoles or anything else and be wrong, I think. This is really about the notion of a general set of information appliances that become more integrated with our lives and work practices, and that are agents for a new model of information empowerment. This shift will generate a lot of winners, and losers too. Which camp will you be in, Dear Vendor?