Microsoft should buy Rackable instead of building custom computers We're in the midst of some very strange times. An online book seller owns the leading utility computing service. An advertising company manufactures its own servers and switches. And spots in rural America best known for being, well, rural are turning into technology heavyweights because they have access to cheap power and …
If I were Dell, HP, etc, I would see the purchase as Microsoft becoming a serious hardware competitor and I'd start to rethink my Linux positioning.
The existing OEM partner agreements have provided a great lock-in for MS on the desktop. The millions MS pays to existing hardware partners is really just grease on the OEM skids for keeping their OS position in those accounts. They would be much better served to spend a 1 billion on R&D with those OEM's to develop Windows-centric server, blade and datacenter cooling and power saving reference designs and technologies than to try and out-cheap Google.
Being cheap is just not part of Microsoft's DNA. They prefer to over-resource (see the waste that became Vista), and spend wildly because that is their biggest advantage, cash. There were thousands of IT's best and brightest programmers responsible for what is Vista. For all of it's faults, overabundance is what is killing Microsoft's ability to compete in the market today. With the exception of OneNote, what software has MS released that hasn't been an huge, bloated, mess?
It's the software version of an all you can eat code buffet. A dash of this, a pound of that, want some silverlight for dessert? It's all on Daddy Gates!
Google was forced to bootstrap it's way to market and that's still an inherent quality in Google. The backend systems were designed to be cheap and scalable because that was the only way they could keep up with growth and keep the lights on. Forget the free soda in the breakroom and free dry-cleaning, it's nothing compared to what they saved by establishing their own datacenter hardware processes.
Building reference designs and seeding their hardware partners would provide a much better return than any initial cost savings from the purchase. Rackable has a great focus today, but imagine what damage Microsoft's huge resources would inflict on Rackable's DNA and focus over the long term. In 5 years they would be forced to buy inefficent, out of date technology from an in-house subsidy rather than select the best tool for the job at hand.
... if Microsoft would make this or a similar buy, to make their own server hardware.
It could certainly lead to some very interesting innovations.
Perhaps big names like HP and Dell will put more effort into Linux and force Microsoft to become an altogether different company?
Or the big bad Apple might just come along and beat Microsoft to it, there really is no way of knowing, yet.
Depends on what services they plan to provide, but I can say that hand picking motherboard components and power supplies is going to be only a tiny part of microsoft's or msn's cost issue. The main problem is if they use IIS/dot net to perform the services, the inefficiency of the application/OS stack will destroy the 10% savings from hardware choices.
My own employer uses IIS for all web services. The have small datacenter at one location with about 50 windows server machines and by my estimate of the actual work they do, they could all be replaced by about 3 boxes if they were running on any unix-like OS from AIX to Linux. Hysterically, the main app is actually cross-platform so they could easily make this transition if they wanted to save $10k a month on a/c and power.
Apple is a hardware company with a huge online services presence, they dont really have a large server Range. A range of power efficient, highly scalable servers would be an excellent string to Apple's bow, plus Rackable play well in a specific niche, which is Apple's sort of market.
If not Apple, then Sun would be my next contender if it had the money, they like to be associated with the high end in dot com.
Previous poster is absolutely right, Microsoft will never buy a credible server vendor because it makes it look like they have aspirations to compete with their Channel, they would be mad to compete with their Channel, Microsoft doesn't sell direct, it sells through a channel, primarily by bundling.
As I remember, Microsoft's recent ventures in the hardware arena were CP/M softcards for the Apple II, and mice with two buttons (with weird communication protocols). Those were the profitable ones. Zune and X-Box rely on tie-ins to really make anything work.
I suspect that Microsoft can't be a hardware vendor, much less a service provider like Rackable. They would actually need to do some "work" and compete in a business they can strong-arm and bully people. Of course, if they use Rackable as their own, and change over to their own (IIS comes to mind) software maybe they will realize how terrible it really is. The problem there is it might take some innovation and original ideas (Bob??).
Sorry, doesn't compute for me!
Right now MS should be focusing on getting the ship afloat rather than obsessing about Google. Fix the bread and butter money earners: Vista etc. Once these are on track, then start thinking of expanding into other segments. Right now a loss of focus will only cause MS more damage.
Unfortunately (for MS) Ballmer is obsessed with Google and wants to attack Google in any way he can: buying advertising eyeballs (via Yahoo) or trying to take on Google Android (via Danger). It does not look like he has any time and energy left to take care of core business. MS has an appalling history of getting most acquisitions wrong. Purchase by MS == kiss of death.
Even when MS gets back on track, purchasing Rackable does not make that much sense. The talent is in the employees and, like Yahoos, Rackable's people would probably leave in droves if Borged.
Witht eh way M$ is going with Windows 7 and Server 2008 Core, ie back to basics and command line's I tihnk a purchase of Rackable may very well lead to modular appliances for things like IIS and SQL. That way the OS becomes negligable in the grand scheme of things, M$ use their recent cloud/utility exposure to ensure that the scalability is there and bang, easy datacentre for monkeys.
Suggesting M$ buy Rackable is a silly, laughable idea.
Rackable doesn't build anything. No factory, etc. It is just a paper business. Rackable's kit is actually built by a disti that has a custom integration group. If M$ should buy a ready made server business they should sidestep the middle man and just buy the disti's integration group. Rackable's stock value now is still higher than real value. It has little to no capital equipment, no factory, etc. It is simply a just-in-time provider using contract manufacturing resources. A baloon full of ether and happy gas that Dell 2.0 is aiming to pop with a needle.
The main story is what Microsoft is going to do with these data centres? It's investing very heavily and very quickly. What are the products/services that Microsoft is going to deliver off all this iron? I'd guess there's some big news to come that is very little about hardware.
''Microsoft is talking about getting 1,200 watts per square foot out of its new data centers.''
Errrm: no, that is the wrong metric. They should be talking transactions/web_pages per second - or something like that. The reason that MS thinks in terms of watts/square foot is because it is wedded to one architecture (software & hardware) that makes this a reasonable measure of performance.
I am wondering when Google will start using different architectures - being Linux based they could build machines based on the IBM P6 or Sun sparc chip (or whatever) without too much difficulty. MS is stuck with Intel compatible chips.
Run the Cloud Server and you Control All Virtual Metadata.
And Create a QuITE Perfect Beta Playground for AIMacrosoft Hardware Compatibles?
Yeah, of course, you could and IT would.
It is likely to Result in AIRevelation that is Revolutionary but merely Simple Evolution of Thoughts' Power. That's the High Road.
They may however choose the Lower Road and arrive to where they are going Late.
So all this rather elaborated scenario is based on a PR declaration. By MS. I believe that even the Daily Mail doesn't pay attention to these anymore. MS are just begging for "green and innovative" attention, as they were begging for "standard" attention with the whole OXML stuff, or "openness" attention with the pseudo-free pseudo-release of pseudo-useful code. I don't believe that they intend to seriously act on any of these declaration anytime soon.
That's still a good analysis btw (MS should buy Rackable and not the real makers, as I doubt they have the expertise to translate a spec into real design). As for whether the OEMs will be pissed... for sure they will be, but they spent the last 20 years advertising MS OSes and pushing MS monopole, they are kind of trapped now, it would take a long time for them to re-educate the milking cows (not to mention the migration resistance... I don't see Joe Bloggs migrating his whole small-to-medium MS-branded "integrated network and productivity solution" willfully).
But again, MS are NOT going to do what they say (it's against corporate culture or something). So the whole discussion is kinda pointless.
Rackable may have lots of Linux customers.
Maybe Microsoft doesn't want to support them.
Maybe Microsoft doesn't want to be that efficient and instead prefer to use bloated Vista and the extra power and hardware needed to run it.
Hell who cares about Microsoft anyway. They are late to the game and are now a has-been.
Someone's assuming that the owners of Rackable are only interested in leaving the table and cashing in their chips. If it's got a better solution than anyone else and it's continuing to make money, they might not actually want to, though.
Shock horror - people might be looking for a long-term investment and trying to compete, instead of selling out to J Random MegaCorporation?! They can't do that! It's... it's... un-American!!!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2023
