Microsoft's software + services strategy takes shape

The news emerging from this week's Convergence conference and elsewhere shows that Microsoft's servers+services strategy for surviving the shift to on-demand computing is crystallizing. Whether it will work is another matter. But here are the strands that seem to be coming together:

Office running on web services (ie XML and Windows Communication Framework) is the recommended platform for enterprise users to interact with cloud resources (the word 'cloud' here is intended to encompass both external on-demand services and internal SOA services. Users don't need to know where the services reside. They'll let the IT department worry about that.)

Keeping Office itself on the client rather than letting it get sucked into the cloud is the critical element. Quite simply, Microsoft can't survive unless it finds a way of maintaining a healthy stream of license revenues from both Office and Windows. It's prepared to go as far as virtualizing the client so that users aren't tied to specific machines. But it's going to make damn sure that licenses are tied to individual users (or at worst, concurrent users).

Duet was just a learning experience and SAP is welcome to it. The strategic product for Microsoft is Dynamics Client for Office, which as Mary Jo Foley observed yesterday, "in addition to functioning as a new user interface layer — also is a new Client Access License (CAL) for Dynamics ERP users." What's interesting about this is that partners can use it as a platform for building new interfaces into the data and processes that execute on a Dynamics back-end. This is a perfect legacy migration strategy that allows Microsoft to keep its partners and customers hooked on Office and Windows without being shackled by the constraints of Microsoft's ERP software. Crucially, by decoupling the user interface from the back-end, it also makes it easier to implement hosted ERP services without impacting the user experience.

It's SOA on the client that matters, not on the server. Microsoft says it has all but abandoned its Project Green plan to converge all its different Dynamics products onto a single SOA code base because customers didn't want it. Well, customers aren't that keen on Oracle's Fusion platform or SAP Netweaver either, but those vendors are pressing ahead full steam nonetheless, and if Microsoft adjudged Project Green to be in its best interests then it would do likewise. It has found a better plan, though: migrate customers to an Office client and then help them transition their server platforms at a later date.

Hosted servers aren't strategic. Microsoft will build a roadmap for hosted ERP, just as it has for hosted CRM, but it's not wildly excited about the prospect (indeed the forthcoming multitenant version of CRM seems to have slipped some more). Why spend time and energy developing hosted services that compete with its installed base of established applications when there are so many unexploited new service opportunities emerging?

Now's the time to start investing in services. The impending TellMe acquisition shows that Microsoft is willing to spend large sums of money where it sees big potential for delivering network-hosted services. (TellMe uses an entirely hosted infrastructure and as such counts as one of the largest on-demand application vendors currently operating). I wouldn't be surprised to see TellMe-developed services becoming available on Office clients in the future as well as on phones.