Microsoft boots Savvis from data centers

Need some details, but quite interesting...

Some details are lacking in this comparison. For example, how much data center capacity does MS have en-toto? Is the Savvis acquisition significant in the way that a "Googleplex" is?

The "interesting" part of all this, and it needs some corroboration with Microsoft's total server farm acreage: is MS getting the same "mileage" out of its data centers as Google? In other words, is MS getting more done (in proportion to Google) with less hardware, or are the Redmond boys running behind in Internet services the same as they are with Xbox gaming?

This is a VERY significant item to research. Microsoft is making big plays for the data centers of major businesses and governments. If they can demonstrate that they have better ROI and TCO on a 100% MS product based data center, then the rest of the world can just roll over and die.

No one (except Gartner, maybe) really looks into how the efficiencies of a complete data center – or even the performance of a functional area within a data center – work out with Microsoft or IBM or Open Source solutions anchoring a majority of the services. And it is incredibly significant, especially in the TCO arena. If a MS monocluture in the server room means that you can get equal or better work loads as a mixed platform data center, you immediately reap the benefit of uniform staffing, volume discounts and lower overhead than maintaining "home-grown" or OSS applications.

It doesn't matter where all the “bright young lights” from college go: the business world only cares about getting value from its assets. Since MS will eventually assimilate any good ideas that get proved in the wild by these same smart kids, all a business has to do is patiently wait for new products to come in from MS.

This is a business model that is well proved and works to generate immense profits for the vendor as well as giving customers a reasonable performing and stable information environment. IBM did essentially the same thing with mainframes up to the point where sheer costs of hardware made microprocessor-based networked systems a better ROI than mainframes. Even today IBM continues to be able to drive massive revenue from this model...it's just updated with AIX boxes and blade servers - as well as Z-class mainframes and big AS/400 clusters still doing useful work.

Finally, we must remember that very few organizations need the specialty hardware and software that Google uses. Most companies can do just fine with a fairly finite number of platforms and applications; indeed, this is the message that most vendors carry to their customers (“use our stuff, get rid of the other vendors, and reap the benefits of doing more with less”). If Microsoft is running its (admittedly 3rd. class) web services and search applications on less than the same per-search fraction of data center power as Google, then Microsoft has a VERY strong message to tout.

And if I could s**t monkeys, I'd be famous.

Picking up scraps after Google

Google has just moved *out* of these same datacenters.

So not only isn't Microsoft capable of building their own DCs, they are effectively picking up scraps after Google.

"Gain energy and operational efficiencies" my ass.

'Gaining energy' by opening up a DC in California? Rrrrrrright...

Roadkill in the sun

that smell is the first sign of decay of the old business model.

commercial software will survive in the long term only if it has compelling value to its target customer base. the rest will become commoditized, and will either go OSS, or move to another business model.

MS stated years ago that it will embrace SaaS (software as a service). so far, there is little to show for it. if licensing revenue were to collapse tomorrow, the company would have to scramble to purchase a viable SaaS vendor. they certainly have the money to do this, but first, they'd have to admit there's a problem.

this seems unlikely. MS has always been slow to learn from mistakes. more of the same is much more probable.

It may not matter now

where those "bright young lights" go, but it sure will matter where they are in about twenty years, since they will be in a position to actually make the strategic decisions.

For having worked with University graduates, I can say that Linux is very much in favor over Microsoft, viewed as an inherently flawed and inferior OS. When these bright types hold managerial positions, I think Microsoft is going to have it's rear handed to itself in so many ways it'll make Gartner dizzy trying to keep track.

Microsoft as its own customer

I don't think there is any significant evidence to suggest that commercial software is on its last legs. Firstly it already does offer compelling value to its customer base - I don't think Microsoft, Adobe or Oracle are worried that they'll become open source yet. The truth is most open source software is developed commercially anyway, the idea that it's mainly hobbyists working on Linux or Apache or MySQL (or whatever) is a myth. There are huge commercial interests in making sure that open source is successful for a whole raft of reasons, the central one is to ensure these companies make money. Fundamentally there is only one business model and that is to make money.

As far as I can see the non-Microsoft companies embrace of open source has nothing to do with open source ideology and everything to do with not being Microsoft; companies don't naturally cooperate with each other (especially when they're in the same industry) and the best you can say is they've entered a period of détente to avoid being obliterated by Redmond. The problem is though it seems unlikely to me that this kind of technology will become a commodity as none of it is fundamentally good enough to become transparent. The usability and reliability of most software, be it commercial or open source, is too poor for that to happen and software development is currently too fluid and lively for that to happen.

An additional problem with entirely commoditising software is development will inevitably stagnate, ideas become very conservative but basically commodities tend not to make much money. Just look at the PC market, hardly anyone is making money at the white-box PC market. Go into any supermarket and the hyper-commoditisation of food has driven down prices to the point that farmers are literally giving their crops away. The car industry make so many very similar models there are literally millions of unsold cars on manufacturers inventories. You make money out of scarcity, not plenty. You need some kind unique quality you can't get anywhere else, that's just basic to capitalism. The problem with nearly all kinds of commoditisation where prices collapse because of surplus is it depends almost entirely on either massive subsidy or someone losing their shirts; be it farmers losing money to keep food prices down, Ford losing billions because of excess capacity and no demand and right through to many open source applications being developed because of indulgence of rich companies like IBM and HP who know they can afford to subsidise it.

Given Microsoft's size and wealth I'd say they have been very quick to turn itself around, whether to the Internet or to making security much more fundamentally important. People have short memories, both Apple and IBM have almost went bust because they couldn't turn their business around fast enough. After ten years of open source hype there still hasn't been a turning point because it still isn't compelling enough - conquering the cheap web server market seems like a Pyrrhic victory. Even Google use Linux simply as a means of reducing cost, it's just a means to an end but the real nuts and bolts of Google, the really fundamentally bits of their search engine code is all closed source and given that's their most valuable asset it'll always remain so. Google may well be software as a service but their business model is as old fashioned and conservative as Microsoft's.

Re "bright young lights"

I used to be one of those students happilly thinking that Linux was going to eat M$ (and proprietary UNIXs) just as soon as my generation got in the driving seat. Then I got into real enterprises, met the previous generation of which many had thought the same of BSD and Linux, and soon came to the realisation that it would be very hard to completely remove Windows from just about any business simply because the change is too big to make it commercially attractive. I predict even Vista will not kill M$ as others have said it will - exactly the same was said of WIndows 95, 98, 2000 and XP. The Linux acceptance curve is definately ramping up, I see more of it in commercial settings every day, but it almost 100% in servers and not the desktop, which means the browsing and other web services are being driven from (usually) Internet Explorer. Those same employees then go home and use their very similar Windows PCs to do even more of the same.