A fundamental flaw in Microsoft's new-found commitment to on-demand services (it calls it 'live' software) is that its licensing policies are driving on-demand ISVs to open-source alternatives — what's known as the LAMP platform (Linux, Apache, MySQL, PHP/Perl/Python). Don't take it from me, this is what Sam Ramji of the Microsoft Emerging Business Team just wrote in his blog (my emphasis added):
"I've been working hard to develop a strong Microsoft-based offering for startups building SaaS companies, because the economics are with LAMP right now."
The licensing problem is described at length by Shannon Clark (who in turn was responding to Robert Scoble's heart-on-sleeve list of 12 reasons why Web 2.0 entrepreneurs aren’t using Microsoft software):
"I have yet to see an easy to obtain quote (and explanation) for how to build an unlimited number of users web 2.0 application based on MSFT servers in a legal and fully licensed manner," Clark wrote. "... we either have to invest a lot of time (and thus money) into understanding the nuances of MSFT's licensing/versioning schemes or we can avoid all that by going with much cleaner and more scalable licenses from open source applications."
Microsoft's licensing is still far too rooted in the enterprise software mold of per-seat client licenses and single-instance servers to adapt to an on-demand, multi-tenant environment where the number of registered users could fluctuate from moment to moment (and that's assuming users are registered; in an ad-funded service you wouldn't even be able to reliably count the precise number of unique users).
OK, that's not going to hold back Microsoft from delivering services of its own, but if no one else in the on-demand space is using its platforms, it drastically restricts the pool of innovation Microsoft can draw on. In fact, the first of the many acquisitions Microsoft will be making to pad out its on-demand offerings perfectly illustrates this point: the FolderShare site is a LAMP site. If Microsoft can't even find a .NET-based file-sharing service that's good enough to buy, then its own acquisition preferences fatally undermine any pretension that it offers a superior platform for on-demand services.
Kudos to Scoble and Ramji for owning up so publicly to Microsoft's shortcomings in this area, but it's a sorry indictment of the company's leadership that these problems are only now being acknowledged, let alone addressed. Microsoft may boast that it's been thinking about on-demand services since 1998, but in all that time it's consistently swept the on-demand licensing problem under the carpet, leaving it facing huge headaches that top Microsoft executive Eric Rudder earlier this year admitted are "making my head hurt. It is a very complex issue, I don't have the answers."