Microsoft needs to show developers it's serious about 'services'

Microsoft needs to show developers it's serious about 'services'

Summary: If Microsoft wants to encourage the creation of an ecology of development around services that use its tools and deployment scenarios, where's the beef?

TOPICS: Microsoft

While a lot of focus on how Microsoft will step up to the Internet-as-platform competitive threat has focused on MSN, let's not forget MSDN. The Microsoft Developers Network (MSDN) should be a major resource through which Microsoft binds developers and ISVs to its revamped services development and deployment strategy. But for now the MSDN portal is rich in downloads -- from code to white papers to how-tos (ie, "Create an RSS feed") -- but skimpy if not downright devoid of software as a service (SaaSer) support.

MSDN sports lots of resources and free stuff that you can bring on down to run on-premises, but what if you want to build an application or services using all of Microsoft's tools and then test them running as an open, highly scaling Internet service on the latest and greatest Windows Server System network? Not much luck at MSDN. And otherwise you'll have to pay a third-party hosting organization or Microsoft itself some pretty big bucks to try out runtime funtime at one of their executive resource centers. Last I heard, Microsoft wanted $10k a day to hang out at one of their server farms and test-run apps.

So if Microsoft wants to encourage the creation of an ecology of development around services that use its tools and deployment scenarios, where's the beef? Where is the Microsoft library of services as platform adjuncts on MSDN? Lots of kits and tools to download there for making Web services, but nothing like what Amazon, eBay,, Yahoo!, and Google are providing. So I'll believe the Microsoft new bear hug on services when we see MSDN encourage the use of an expansive library of mash-up-able services.

Of course, such a move would only amount to playing catch-up to what the Web 2.0 developers and services portals are already planning and providing. But Microsoft clearly has the resources to take the seduction of developers on services use and association a frog-leap further forward.

At the same time Microsoft could demonstrate that it understands grid computing and has the high-scaling and economically beneficial underlying infrastructure to support services, from cradle to grave. What I recommend is that Microsoft swipe a page from Sun Microsystems (they've done it before) and its grid/utility and storage computing initiatives (you know, the $1/cpu-hour-priced service to run and/or host 32-bit applications).

Microsoft in one fell-swoop could out-do Sun and the services portals by offering a Windows grid on MSDN to build, compile, test, and host enterprise and ISV apps as services. They could even afford to undercut Sun on price, to say, $0.25/cpu-hour. Nothing wrong with a loss-leader when you're playing catch-up.

Like Sun is trying to on Solaris, Microsoft could seed the Windows services hosting support ecology, and become the bottom of the foodchain (in a good way) that supports other hosting organizations upstream. This would further encourage developers to use Microsoft's latest tools (and get that great new Avalon look). Microsoft could make the entire Windows ISV and developer ecology an offer they can't refuse, and get them all to deploy to the CLR as a utility service to boot (or reboot). Maybe the telecos would be interested?

Now, the clincher ... to make it silly to build and support new Web-based applications and services on anything but Windows -- give developers free or really cheap compile time as a utility. Just throw it in as an extra. Don't even require on-premises servers to compile and recompile apps and services (as long as they use Windows standards). That's even cheaper than a Linux development stand! Provide the development cycles dirt cheap to developers anywhere in the world -- before Google does -- and that should make the ISVs and developers take Microsoft seriously on services.

After all, all those apps will have to run somewhere sometime, at scale, and that's the server demand growth engine in three years. You may see enterprise server license revenues go down, I know, but at least you're jamming with the hosts now to be their preferred infrastructure partner later. And, Microsoft, you're going to have to build out a bad-ass global SaaSer network to compete with Google anyway, so do it fast and give to it the Windows-only developers for a while. Talk about a sandbox!

May the vendor with the biggest bank account win.

Topic: Microsoft

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Some corrections and clarifications needed

    First, the $10k (thought it was actually a little higher than that), last I know includes a 24hours time in one of their regionally located 'test centers'. This also included all the consulting time (with the regionally deployed consultants) that you needed from troubleshooting networking issues, application issues, performance issues, etc. All also included the full use of their hardwares that you want/need as spec by you out from the hardwares that they have.

    Second, your 'cost' on .NET hosting is also inaccurate. Just about any ISP/Hosting company out there has that support now and their price is no different than any other 'CGI-enabled' service.

    Third, the cost of MS's development tools are cheap if you compare it with what you're getting. And with the newer .Net Express line, it's currently free (Beta) and when released, will be only $49 per Express license. So for small developers, that is not much of a cost considerating the power of the tool that one is getting. And for larger development shop, MSDN or Visusal Studio license is also reasonable.

    As for support, I dare that you try to find a better company with better Development Support than MS. My experience is that there is none. not IBM, not Oracle, not Sybase, not Sun, etc. This whole 'new shift' is too new, give it another weeks or month, you'll see the developer's resources in MSDN.
  • ..... but they aren't

  • Wouldn't work

    I don't think this would work. With SOA, enterprise IT has finally figured out how to escape Microsoft and get to somewhere it likes better. They are thinking long-term, and so they are not going to fall for any short-term seductions from Microsoft. In particular, they won't buy anything from Microsoft unless it is completely open.

    Face it, Microsoft is screwed. The world is moving to a new basic computer technology that Microsoft can't take over, and that's just the way it is.
  • Perpetual betas

    Well, before web-based services take off, companies will need to end the "perpetual beta". The majority of Google's services have been in beta for years. Google's published API's are marked beta. That means they can change at any time, and you use them at your own risk.

    You won't see new MSN services and APIs languishing in Beta for years. They will release them, and then back them with product support like all their other products.
  • You're making the assumption...

    ... that Microsoft is emphasizing the web-as-platform as opposed to web as resource to a server. (You're also making the assumption that Google is going to believe it profitable to run web as platform [SaaS]. That may be true, but it hasn't happened yet.)

    Microsoft may allow for web platform, but not emphasize it. I did a post in another thread pulling quotes from a Bill Gates interview on ZDNet that seemed to indicate that direction.

    Before considering how Micosoft can implement a strategy, I'd appreciate a discussion of whether Microsoft intends to follow that strategy.
    Anton Philidor