Monday, November 03, 2003

Ed Brill - CRN: Cracks Appear In Microsoft Kodiak 'Store Story'

Ed Brill Ed Brill - CRN: Cracks Appear In Microsoft Kodiak 'Store Story' : "Wow, this is huge news. Microsoft originally started talking about SQL Server as the basis for a future Exchange release in a leaked December, 2000 Steve Ballmer memo. I wrote about it on lotus.com in January, 2001, and MS spent most of 2001 in denial mode, even called me a liar on their website. Yet in October, 2001, at the MS Exchange conference, there was Paul Flessner, showing a roadmap slide for Exchange that included 'Kodiak' in 2003. He was, not surprisingly, light on specifics, and continued to defensively say that it wasn't important to speak about it then.
Yet here we are, end of 2003, and no Kodiak in sight. CRN is now reporting, in fact, that MS can't decide whether to wait until the SQL Server team gets their act together and ships 'Yukon', or whether to proceed with the creaky JET database that has been part of the product since its inception."

Interesting debate context -- IBM, Microsoft, and Oracle have all suggested the future of email storage is in extended DBMSs, and Oracle is already shipping an instance, albeit apparently not with much success. In the meantime the leading enterprise email products, Notes and Exchange, continue with non-DBMS architectures. Ed's reference to IBM Lotus Workplace Messaging is a consideration for a subset of enterprise email users, but, for the vast majority of users, the shift to a DBMS-based model isn't going to happen "real soon now."

The WinFS question is also worth revisiting now that we know more about WinFS; e.g., while it will certainly be sensible for MS to exploit WinFS for contacts and other PIM-ish data, it's not clear that WinFS will be the best native store for messaging, and, as Tom Rizzo noted in the CRN article, in the meantime MS has a very large enterprise messaging business to run.

In any case, let's all refresh our codename caches and move on to debates about customer-oriented features instead of "truth maintenance" on historical predictions.

No comments: