At SNW I met with Claude Lorenson, Ph.D., Group Product Manager for Windows Server Marketing. If you can decode that title and triangulate where he sits in the MS corporate huddle you are a more informed man than I.
I’d met Claude several years ago when MS was just getting ready to roll out their first version of Windows storage server – NFS and CIFS – maybe five years ago. I wasn’t impressed then because it just seemed like another uninspired MS effort to leverage their Windows monopoly to extract a few more bucks from the OEM in their thrall. Big woop.
With the HP All-in-One storage announcement Microsoft is playing, through their OEMs, in their home turf of SMB markets. Claude said there will be more announcements in the next few months as well.
What a difference five years makes
Yet as I’d analyzed the Boxwood project (a few posts down) it gradually penetrated that Microsoft could actually be a significant change agent in a moribund storage industry. MS doesn’t have any hardware storage revenue to lose by doing something creative. They could seriously PO HP, IBM and Dell, but what the hey: what good is a monopoly if you can’t jerk people around? Besides, those companies would all be happier, with the possible exception of Dell, if EMC took it in the shorts.
If MS wants to kick Linux where it hurts – and despite the Novell deal I’m sure they do – getting serious on the storage side would be a great place to start. Boxwood demonstrates that it is entirely feasible to build a scale-out infrastructure like Google’s using Windows. While customers would have to buy Windows machines, they wouldn’t have to buy RAID controllers and FC SANS and they could use the storage functionality that Claude’s group has come up with over the last several years.
Claude acknowledged that several big Microsoft OEM customers would be upset if MS developed an internet data center-style infrastructure without expensive RAID boxes. Yet they have to balance that against the fact that lots of big corporate customers, who are facing increasing competition from web-based services, would learn to love it. And pulling in a couple of billion more of 98% gross margin revenue would be nice for MS stockholders, who’d like to see some growth.
Will MS take the plunge?
The StorageMojo.com take is that, as usual, MS will wait for someone else to prove there is a market for internet-scale clusters before they step in, so the folks in Hopkinton can breathe easy, for now. Which is too bad, since the demand is obvious. Claude, you and your troops should be working on this stuff since when the day comes that MS needs a clear differentiator from Linux servers, storage will be your ace in the hole.