I came across this article

http://news.com.com/Windows+for+supe...l?tag=nefd.top

announcing Windows for clusters etc by this year so I'm guessing around 2007 or so those Solitaire games are going to zip along now.

A few quotes from the article from Software Architect Marvin Theimer:

Yeah, get it into the shops ready or not.

Feeling the heat?

Why on earth would anybody smart enough to run a cluster buy a crippled OS for it?

That proves it, this is just the minimum system requirements for Longhorn

Expect a few press blitzes when some fortune 500 is levered or bribed into switching away from Linux, and glowing comparisons from shills like Gartner, oh yeah, a blizzard of FUD starting any day now too - bleah

Interesting, but I bet they'll have a hard time competing:

A reduced price could still be a steep price. If it's still priced according to the size of the cluster, it could get expensive. With Linux, you can make the cluster as big as you want, and not pay licensing fees. Even at a reduced price (say, 50% off the price of a normal single-machine Windows license), you'd be paying maybe a hundred dollars per node, instead of $0. They'd have to give some pretty major discounts to make it competitive with $0 per node, especially for large clusters. IBM, for example, would be stupid to pay for Windows licenses (at any cost) for any or all of the 65,536 processors in BlueGene/L (16K nodes, if I am reading it right).

Maybe for very small clusters - 5-10 machines, say.

I think the funniest part of the article, though, is this:

Sounds to me like Microsoft is very worried about competition from Linux... strange, considering what an unfavorable light they put on Linux in their PR.

I've built Windows clusters for web hosting and email - both active/active and failover types. If Microsoft intends to cripple Windows 2003 and forward by disabling IIS on a clustered configuration, they're basically saying "Use Linux, BSD or Slowlaris for your hosting needs, because we don't want your business."

Idiotic move, Gates. Not only are you alienating users on the client side, you're now seemingly bent on alienating Enterprise users who were just beginning to trust M$ again thanks to Win2K3's being a lot more secure and flexible than previous versions of Windows Server. Smooth move, exlax!!

. . . which is fine by me since I've been encouraging clients to go the *nix route.

Um, right. Most savvy ISPs and enterprises with strong *nix backgrounds are going to go the Slackware or Whitebox Linux route, and cut distributions' licensing out of the picture. How does M$ propose to compete with Slackware?

Also, to make things worse for M$ - You used to be able to connect two PCs together in an undocumented way to share storage directly on a SCSI bus. With Win2K3 that hack is no longer possible so now you NEED a SAN or external shared-SCSI storage device. With that requirement bumping up the cost in comparison to *nix's clustering capabilities AND crippling of the exact services you WANT an active/active cluster for --- well, they've begun their corporate suicidal move. Granted, M$ can bleed money for a couple of decades before going under, but their decisions over the last few years, with their "antipiracy" (actually, anti-wine) moves, removal of WinFS from Longhorn, and this, more and more corporate and consumer-level customers are going to be taking a more-than-passing-glance look at what Linux can do for them.

Windoze for supercomputing,... that's a hoot.

In the zero-hour of an important weather algorythm,... blue screen of death!!! Yeah, great...

M$ is doing a lot of things that will seriously enhance Linux's chances of greater success,... including their new licensing initiative.