IIS 7.0 and "integrated innovation"

As you may have noticed I have been pretty quiet on the blogging front during this PDC. There are a couple of reasons for this:

  • There are so many interesting sessions here, including lunch sessions, that it is very hard to find the time to blog if you can't or won't blog during sessions. The program extends far into the evening and in order to be awake the next day, it's necessary to catch some sleep.
  • I don't want to blog just about what I saw because so many people are already doing that.
  • It takes time to digest all the new stuff and write a good blog post about it. This post alone took 40 minutes.

That said, let's go the the subject of the title of this post. This was something that I didn't want to hold back on.

As you may have seen already at the PDC, on blogs or on Channel 9: the new IIS 7.0 rocks! The complete modularization of the IIS functionality and ASP.NET like distributed XML based configuration system, makes this a much better and manageable web server than IIS 6.0. IIS 6.0 is of course already a great web server in terms of performance, stability and security.

Now the bad news. Microsoft's current plans are that IIS 7.0 will only be available on the Longhorn Server platform and not on Windows Server 2003. This means that we will have to wait till at least 2007 before we can use IIS 7.0 for production web sites. It will be released as a crippled pre-release in Windows Vista. I say crippled because it will be limited in some way (connection or bandwith limit) so we won't be tempted to use Vista (the client platform) as a server platform.

Of course I understand the business reasons for doing this. IIS 7.0 is a good driver to sell more Longhorn Server licenses. But linking releases together like that means that it will take way too long to ship IIS 7.0. It looks pretty well baked already. Of course PDC demoers know which things to avoid to give the impression that a product is stable and almost finished. But I don't believe it will take another two years to ship it.

IIS 7.0 as it stands today goes way beyond the capabilities of Apache. But where will it stand in 2007? The development of other web servers is not going to wait until Microsoft is ready to release the shipping vehicle that makes money.

This is a good example of where Microsoft's concept of  Integrated Innovation really hinders Microsoft's ability to ship software and stay ahead of the competition. Shipping version 2.0 of the .NET Framework and related technologies has taken already way too long.

Of course I can appreciate the benefits that Integrated Innovation brings when aligning a lot of products on one version of the .NET Framework and Visual Studio. But in the case of IIS 7.0, I really see no technical merit whatsoever. When I asked a Microsoft employee at the IIS booth at the PDC about the availability of IIS 7.0, he first started by claiming that there were technical reasons why IIS 7.0 could not be released on Windows Server 2003. He also stated, quite honestly, that there were business reasons for doing this. Then I told him I heard that IIS 7.0 was being developed on Windows Server 2003 and not on Longhorn Server. He admitted that was true and that the technical reasons are not really there yet.

I am afraid that customers will not migrate immediately after the release of Longhorn Server just to get IIS 7.0. That means we wil have to wait till 2008 and beyond before adoption will take off.

This is why I hope Microsoft will rethink its strategy and make IIS 7.0 available for Windows Server 2003. This will not be a backporting effort as IIS 7.0 is already up-and-running on Windows Server 2003. Microsoft, please release IIS 7.0 as early as possible and include an even better IIS 7.1 in Longhorn Server if you need a business driver!

Leave a Reply

Your email address will not be published. Required fields are marked *