HP and Microsoft are partnering to create a new vision of communications services that's hosted on switch blades but runs enterprise software. In the carrier world, there's a growing call for a new and different approach to creating network services and standards, something that could change public networking. A lot of ink will be spilled talking over each of these two developments, and the tension between the "enterprise" and "carrier" models of future services. It's not the tension between these two activities that is important, it's where and why they converge.The notion that UC/UCC might be heading for a network-hosted implementation approach has already been introduced with developments like the Siemens prototype implementation of UC on Amazon's EC2 cloud. The Microsoft/HP alliance is at one level offering a different approach, the hosting of UC/UCC on switches (ProCurve One blades), but both approaches have the common theme of pulling UC/UCC applications out of fixed servers. Both essentially make UC/UCC something inside the network.
Then there's the activity in the service provider space. IEEE Communications Magazine's May 2009 issue has a whole section dedicated to "Innovations in NGN," and what makes the topic so interesting is the lack of focus on the IP Multimedia Subsystem (IMS) that has dominated service providers' NGN thinking for nearly a decade. In the issue there's a consistent notion that the old service layers need to be rethought, and that networks in the future will likely be based on service overlays and distributed feature components.
Smartphones and netbooks enter into the mix as well, and in fact you can see a bit of this in the HP/Microsoft announcement. The future of communications, the pair suggests, lies in a transition from an instrument-centric model to a more virtual model. Intel has taken this story up with its new Mobilon Linux version, designed both for smartphones and netbooks, and of course Google's Android clearly has the same aspirations.
So we now have three forces--enterprise UC/UCC, carrier NGN, and handset/netbook OSs--converging on a common theme of hosted service features, virtualization of services, and increasingly a view of communications applications as a set of interoperating components. Can these three things converge on a common thread without converging on each other? Hardly, so we need to look at how that ultimate convergence might work.
The current forces seem to create a world where hosted, cloud-based, or private service features could all be composed into a common structure to create experiences. This structure would consist of feature elements that might run on any convenient network or IT device or in an appliance. It could be managed by the service provider, the enterprise, or a third party--even in theory a combination of the three. That sure breaks down some traditional boundaries, but it seems like it's getting universal vendor support.
The Microsoft/HP deal, and recent Microsoft comments about public cloud computing, suggest that Microsoft is accepting that communications services in particular are going to be "outsourced" either to an enterprise cloud (running perhaps on HP ProCurve blades) or to a provider/public cloud. Given the "cloud trend," for Microsoft to fixate on running this kind of application on enterprise servers would be to lose the market. But obviously Microsoft would like some elements of communication to stay on computers, so they have to accept a componentized view of communications applications. That same view now pervades service providers, so the future of communications becomes one of connecting software components into personalized units of service and managing the result.
Anyone who's worked in the service component space will tell you that this is an application program interface (API) issue first and foremost. Software components link to each other via APIs, and we've had all manner of groups working on APIs in the carrier space, including OMA, GSMA, and TMF. A researcher in the IEEE publication I cited comments that the real API success has come from players like Google with OpenSocial, or from smartphone players like Apple or Google. Unlike the relatively "heavyweight" APIs of standards bodies, these Web APIs are based on simple HTTP and XML and so they're easy to incorporate and extend.
That brings up the second reality of converging on a universal "virtual" communications model--developers. All of the areas of communication are personalizing, and no equipment or software vendor can afford to do all the personalizing work for every user. The notion of third-party developer support has become so popular we're at risk of having more developer programs than developers. Most of the developers in the virtual communications future will be integrating some custom logic with basic features accessed through those good old APIs, just as iPhone or Android developers do today. Most of the iPhone applications that are sold create something the user could do directly with a browser; it's just easier to use an application for it. That accentuates the component model of communications services, of course. It also encourages simple APIs because that's what developers like.
This may be a new view of the communications market, but the players who want to own this new market are familiar ones. Microsoft and HP are counting on the notion that UC/UCC will migrate into the network, and that their incumbency and their alliance will give them a strong position. Given that Microsoft's service provider strategy has gelled around the notion of Microsoft working in partnership with providers to offer things like email, it's not surprising they would see the HP alliance and a network migration as something that ends up with Microsoft components within telcos cooperating with Microsoft components in enterprise switches.
The telco aspirations in the managed service space may show that telcos and their equipment vendors hope that network operators, particularly mobile operators, could use things like femtocells as outposts in the enterprise. From this position, virtualized communications services could migrate off the network and onto client systems. Verizon is now offering "cloud-based" VPN management; why not cloud-based UC management?
The user ends up the winner in this war because it finally harmonizes the two faces of communication-the enterprise-centric traditional UC/UCC vision and the operator-based view that in a mobile society every enterprise must view its communication as a subset of the public network. Components and virtual services give us the best of both worlds, and a way to tune services to match individual and business needs.
