Job Ad Hints At Microsoft Move To ARM Servers 138
An anonymous reader passes along a brief EE Times note on a suggestive Microsoft job ad. ARM is explicitly mentioned, as are solid-state disk drives as an area of experimentation in the quest to reduce power consumption; but Intel does not get a mention. Here is the ad. "Microsoft is looking for senior software development engineer to help with its Bing data centers, potentially running them on ARM hardware, according to an EE Times article. Whoever gets the job 'can own the decision on the hardware that we use,' the job description said, and added that power management is a key aspect of the job. ... Microsoft was reportedly experimenting with the Intel Atom microprocessor in February 2009 with a view to creating a green low-power data center. One issue discussed then was the Atom microprocessor lacked performance compared with other Intel processors and that therefore any power saving might be negated by the need for more processors to carry a given computational load."
Product Experimentation (Score:4, Interesting)
Not necessarily dropping intel:
They are just doing expirementation (s.i.c.) !
--jeffk++
Re:Product Experimentation (Score:4, Funny)
I know, I'm weird.
Re: (Score:1)
and going with this tangent..
try springrts.com for a full 3D multi-player remake of Total Annihilation-like games
Re: (Score:2)
Continuing the off topic TA loving:
The CA (Complete Annihillation) mod for Spring is excellent and well maintained.
Netbooks! (Score:5, Funny)
Re: (Score:2)
Not netbooks. If anything, it'll have to be an ARM single board computer, something like this [igep-platform.com]. I have one of those; they're awesome in quite a few ways.
Re:Netbooks! (Score:5, Interesting)
The OMAP3530 (which is in the IGEPv2) is a cool part in a lot of ways, but it would be boneheaded to put it in a data center. Because it doesn't have any high-speed interconnects -- gigabit Ethernet, PCI Express, RapidIO, or the like -- it isn't suited for most network-intensive applications. Marvell has a variety of systems-on-chips that do have ARM cores, running at higher speeds than the OMAP parts, and with high-speed interfaces on the chip. Other vendors probably have similar offerings; those are the ones that Microsoft would probably want to look at.
Re: (Score:2)
Of course, the OMAP is more of a mobile kind of part. I wasn't suggesting that particular board. I've heard good things about Marvell's implementation of the ARM architecture.
What I want to see are "performance" offerings with proper interconnects (especially Gigabit Ethernet and SATA) and multiple Cortex-A9 cores. That'd be great for all kinds of server applications, datacenter or home alike. The thing with Marvell's fast chips is that they seem to implement just the old ARMv5 instruction set, which is no
Re: (Score:2)
Re: (Score:1)
Now, that is innovation.
Re:Netbooks! (Score:5, Funny)
They're probably consolidating Bing down to a netbook to better serve the dozen or so people who use Bing.
And most of them are related to Steve Ballmer.
Re:Netbooks! (Score:5, Funny)
Actually four of these accounts are owned by Steve Ballmer himself:
1. S.Ball (regular everyday personnal account)
2. Ballman (pr0n account)
3. MacLover (to be able to spy on Apple fanbois incognito)
4. Chairmaster (to do research on types of chairs, kinds of woods and aerodynamics of four-legged seating apparatus)
Re: (Score:2)
You forgot Stoolman, which serves double duty for kinky pr0n and three-legged, backless chair research.
Re: (Score:3, Informative)
they dont need to run the whole internet. Its Bing, not Google.
(hint: joke)
Re: (Score:2)
But... does it run Windows?
Re: (Score:3, Interesting)
I'm surprised Bing isn't better. You'd think that MS would have put a little more effort into it. It's also ugly to look at.
I wouldn't mind more competition to Google. While I like the free email and the configurable home page google offers, some really high-quality competitors would be healthy all around.
Re: (Score:2)
I use Bing.
When I'm buying something from eBay that's about $50 or more, and it's Buy It Now, that is.
Bing Cashback FTW. That's the only reason to use it, IMO. (Although, I had a friend who used it back when it was called Live Search, because she didn't want to contribute to a Google monoculture. *facepalm*)
I don't know.. (Score:5, Funny)
I don't know about "owning the decision on the hardware we use", but I'd like to "own the decision on the software they use".
Re: (Score:3, Insightful)
I wonder how true that quote actually is. I mean, what if you decide to run OSX Server, or order a shitload of iPads... or start buying Sparc boxen (they still make those, yanno...)
Re: (Score:2)
If you can make Mac servers running OS X beat their existing intel servers on price, performance, and power consumption (presumably the goal of the whole experiment) I think they would be extremely interested.
Re: (Score:2)
Um, Mac servers running OS X are already Intel servers. Apple has been shipping Nehalem XServe boxes for over a year.
Re: (Score:2)
I realize that, but it still only legally runs on Apple hardware. So the price/performance/power of the new xserve's would have to beat their existing Intel boxes ( wherever they get them from) to make it worthwhile, or even interesting. It's one thing to beat them at specific workloads with something fairly exotic (like ARM, or IBM Power chips), if you can beat them on more or less identical architectures even with Apple's significant price premium, especially when they would have to pay for OS X server
Re: (Score:1)
Eating your own dog food... (Score:3, Interesting)
The slogan is "Bing it and decide".
Microsoft, just use Bing: "Atom or ARM for my datacenter?"
*Bam* (or bing) there's your answer!
You're welcome!
Re: (Score:2)
Windows Mobile is enough of a headache on mobile devices, I don't even want to think of it running as a server. I think to handle something like Bing, they'll need Windows Server. If they've kept true to NT's platform independant designs they should be able to port it easily enough.
Re: (Score:1, Interesting)
Re: (Score:2)
Writing something like this as AC destroys all credibility, especially since it's 2nd hand.
So... (Score:5, Informative)
Re:So... (Score:5, Insightful)
NT was written [microsoft.com] as a portable OS from grounds up (remember that it had a working MIPS build before x86 build!), and much of that legacy still remains today in OS architectural design, so porting the OS itself shouldn't be hard. The toolchain (compiler etc) is already there to target ARM for CE.
Drivers (third-party ones specifically) might be trickier, though they're still mostly written in C, so for the most part it should be a straightforward recompile.
Re: (Score:2)
Never knew they had a MIPS port, I thought the first one was for alpha?
Re: (Score:1)
Re: (Score:3, Informative)
Re: (Score:2)
They had a MIPS port before Alpha, it's the reason why Alpha ARCSBIOS had MIPS style device names.
Re: (Score:2)
I doubt they ever stopped keeping back room ports of NT for Alpha and Mips.. Kind of like how apple always kept a back room, "secret" port of OSX for intel. Sure, the apps won't run, but in a datacenter, you can pick the exact apps you need, and recompile.
Re: (Score:2)
Dear lord, I can't imagine having to ensure builds would still compile for all of the dead architectures.
Re: (Score:2)
In the pre second coming of Steve Jobs era, there was a secret project called Star Trek at Apple. It was a port of the "classic" MacOS for Intel.
Surprised it hasn't turned up in a Redwood City bar.
Re: (Score:2)
I would just LOVE to see them port Windows CE. Since the thing does not even have the concept of a "current" directory. That is NOT server class software.
Since they will have to write something to run on ARM, might I suggest they take the time to create their own SSH server so it will be possible to administer one of these things like a real computer? This comment is not meant as flame bait. I take care of both Windows and Linux servers remotely. If you have to do it graphically, nonmachine is as fast as RD
Re:So... (Score:5, Interesting)
A WinCE PDA is almost exactly the same size as a data tape. With modest modifications(consisting largely of forcing the work experience kid to run lots and lots of docking cables) an industry standard tape silo can be turned into a gigantic WinCE/ARM blade farm. If a node stops responding, the robot retrieval arm pops it out, presses the reset button, and pops it back in again. Since the OS is in ROM, boot is short and downtime is minimal.
Re: (Score:1)
I can't decide whether to mod this insightful or funny.
Oh well. I guess I'll post instead.
Re: (Score:2, Funny)
Better to question whether you are clever enough to moderate, than to post and remove all doubt.
ARM Processors (Score:2)
Re: (Score:3, Funny)
Re: (Score:3, Funny)
BSD more likely.
Re:ARM Processors (Score:1, Interesting)
ARM = Automated Resource Management
And what OS? (Score:2)
Or is a version of Windows on ARM in progress already?
also a good way to hide news of porting Windows (Score:2)
So is it jus
Re: (Score:2)
Want to run Flash on an iPad? Buy Microsoft iWindows 7!!
Not going to find that one at the app store.
Maybe not for the server hardware itself (Score:3, Interesting)
ARM is severely underpowered, even when comparing to Atom. So it doesn't make *any* sense that MS is considering it as a server platform.
However, ARM excels at low power consumption and mobility. This would allow a new array of "server helper" devices that had needed quick handling of light tasks. Maybe something like packet routing or on the fly network topology auto-configuration. Another concept could be mobile cache points which would be somehow networked to the main servers and provide "smart caching" of data for light user requests.
Who knows. But to think that ARM is going to somehow best Intel's chips in the server market is crazy.
Re:Maybe not for the server hardware itself (Score:5, Interesting)
The important metric is MIPS/Watt. It doesn't matter if the CPU is "underpowered" if you can run a bunch of them in parallel to get the same performance as a Xeon, and still get better power consumption. Remember, the work they're doing is highly parallelizable, so outright clockspeed isn't very important.
However, I don't have any MIPS/Watt figures available for ARMs or Xeons, so this is idle speculation. If I were to take a guess, however, I'd guess that, given a real-world workload, the Xeons would probably beat the ARMs because of many factors: cache size, context-switching time, etc. If it were more economical to run a datacenter on tons of low-power ARMs, Google probably would have already done it by now.
As for craziness, remember, this is Microsoft we're talking about: the company that thought SongSmith would actually be a commercial success instead of a complete joke. Given their combination of big successes and utter failures, they seem to be quite neurotic.
Re: (Score:2)
Even if the ARM cores were more efficient, in terms of work done per unit power consumption, you'd still get less work per core, which means more cores, which means more network gear(whether it be network gear in the classic "1U, 48ports, ethernet" sense or whether it would be some custom thing(dozens of cores per card, with glue logic, some sort of cardcage/blade design, whatever) the interconnect silicon
Re: (Score:3, Informative)
You can get ARM blades with many chips on one blade already. This is not each little machine with its own network interface, the overhead would be massive.
Re: (Score:1)
Wolverine will be thrilled to hear that.
Re: (Score:2)
Re:Maybe not for the server hardware itself (Score:4, Interesting)
Back when I was building a Linux distro for an ARM platform (specifically IXP435) I found it to have maybe about 1/3 of the power of x86 CPUs of the time, but running so cool that the CPU didn't need a heat sink and didn't get so hot I couldn't put my thumb on the CPU. And that was after running a regression test suite for 20 hours. ARM definitely is a win for the MIPS/Watt metric.
BTW, it would be scientifically simpler to just refer to this metric as "millions of instructions executed per joule of energy converted to heat" (would roughly equates to a gain of information in exchange for a loss of information).
Re: (Score:2)
Your argument doesn't make sense. It seems to go something like this: "Because the ARM chip never did get very hot, compared to some other chip, it must be a better performer in terms of MIPS/Watt."
Re:Maybe not for the server hardware itself (Score:5, Informative)
Re:Maybe not for the server hardware itself (Score:5, Interesting)
There was some intersting reseach not too long ago which paired low-power x86 chips (Geode LX at the time, IIRC), Solid-state storage (in the form of compact flash) and a RAM-based caching of solid-state contents. About 10 of these boards were then clustered, running a distributed database application. As I recall, there were some serious performance and power-savings advantages compared to a single larger, multi-core x86 server setup. The primary advantage, in my oppinion, was that the ammount of available bandwidth, both to storage and to working RAM, combined with intelligent caching, was paired in much more favorable ratios to the power of the CPU. In short, the reseach found that a cluster of modest machines turns out to be competitive, even better-than, a single powerful server in terms of cost, power-consumption, heat disipation, and even in performance.
Microsoft is keen to realize that a modern ARM core is quite well suited to match modern I/O limitations. There's no point building a large system when the requests it's going to serve will be bottlenecked by I/O.
Re: (Score:3, Interesting)
Floating point performance does indeed lag behind even Intel's Atom (its also one of the focus areas for the Cortex A-9 core), but is not a big requirement for server or database tasks.
Atom suffers inefficiency for x86 compatibility, but it's definitely the lowest-power way to get SSE3 and it is fairly respectable at multimedia tasks as a result. It's basically a Pentium M, which was the most efficient x86 chip of its day, too.
There was some intersting reseach not too long ago which paired low-power x86 chips (Geode LX at the time, IIRC), Solid-state storage (in the form of compact flash) and a RAM-based caching of solid-state contents. About 10 of these boards were then clustered, running a distributed database application.
The problem (as alluded to elsewhere) is that you need totally custom hardware or a bunch of glue hardware to replace a single x86 blade server with a whole box of blades to do the same amount of work, and for what, power savings? We're only starting to run up again
Re: (Score:2)
I'm pretty sure that Atom is an in-order processor like the original Pentium, and completely unlike the Pentium M. Or has Intel changed the architecture in newer Atom models?
You're right that it's in-order, but it's also hyperthreading (newer ones, single core or not.) The OS is therefore in charge of keeping the two pipelines filled. The mind reels. Wikipedia says it's half as fast as a Pentium M, but the Pentium M didn't have all the features of the Atom, either. Wiki also says Atom is descended from Stealey [wikipedia.org], which is in turn based on Pentium M... so, yes, Atom is based on Pentium M.
Re: (Score:2)
Re: (Score:2)
Underpowered in which way? I'd expect ARM not to have the floating-point horsepower of Intel low-power CPUs, but this is datacenter, not HPC vector computation. I don't know enough about Atom et al to compare the rest of the feature set.
Re: (Score:2)
Microsoft Research has been looking at FPGAs in the datacenter for a long, long time.
Re: (Score:2)
Your comment makes me wonder: what, nowadays, is the real hardware bottleneck when it comes to servers?
I can imagine e.g. an SQL database server or a web server doing lots of dynamic pages needs quite some CPU horsepower and internal memory. Some multi-core Intel would be useful.
On the other hand a file server needs storage and lots of I/O bandwidth - no need for a very powerful CPU. ARM may be in place here as having enough power to keep up.
The same would account for a web server handling mainly static
Have to note as a big 'duh' (Score:5, Interesting)
One issue discussed then was the Atom microprocessor lacked performance compared with other Intel processors
Atom and ARM are great platforms when you don't need much processor in one spot. I.e. many embedded applications and a lot of consumer electronics. They need some processor, but not a lot. 'desktop/server' processors are optimized for a higher load and just don't scale down. Note that ARM isn't inherently low power, it's just the instruction set everyone in the world has rights to implement, and Intel pretty well dominated everything but an emerging low power market. You have a lot of innovation and skill at implementing 'just-enough' processors that simply picks ARM out of convenience.
In the data center context, things change. The notorious energy consumption of the low-power processors come to nothing when you can arbitrarily consolidate workload onto as few processors as possible. The economies of scale of the mainstream desktop/server platforms deliver are far greater than tiny low power devices.
In terms of MS experimenting with it, expect nothing to come of it. It will fail like Atom did in their experiment before. Assuming a long shot, expect nothing to change externally, even if MS discovers ARM is great for their data centers, they cannot readily win a market that centers around lower cost, lower energy, lower performance non-x86 compatible parts. They have a golden example of a company thinking their technology intrinsically drives the industry making a drastic change to discover they were wrong. Intel thought they dictated the terms of the industry, but Itanium simply failed to transform the market without quality x86 compatibility. This was the golden opportunity for AMD to swoop in with an alternative and make huge gains. MS is in the exact same situation, 99.9% of their clout is the environment of existing Windows apps. Microsoft has tried time after time various platforms to reach the same endgame of no success. If the new architecture in *theory* provided more performance, sufficient to emulate x86 instructions, then it would stand a remote chance, but going to lower performance platform renders this impossible. In a really long shot, MS gets a lot of really nice ARM hardware on the market, and then has to compete with Linux on its own merits rather than ecosystem of applications. It's nearly suicide to risk your largest leverage point unless the industry is imminently making you irrelevant even if you stick to your guns.
Re:Have to note as a big 'duh' (Score:5, Informative)
We found this when testing our point of sale app. So long as the POS software was running only the POS software on the terminal with the DB hosting on another machine/server, it was great. But as soon as you coupled POS + DB on the same terminal, lag started to be noticed. It was still acceptable, but it would take 3 seconds to create a new ticket vs. less than a second on a 2.8Ghz P4. Especially on the single core Atoms. The Dual Core atoms seemed to handle things just as well as their 2.8Ghz & 3Ghz power hungry Pentium they were to be replacing. And we tested both Windows XP/WEPOS & Linux (openSuSE/Ubuntu) and saw the same results.
Re: (Score:2)
This is something I've been thinking about lately. Obviously, good design dictates that we use a separate app and db server for performance reasons (as it is usually more responsive). However, what about separate purposes (app server, db server), each running in its own VM on the same host? It's something I'm hoping to implement at a not-too-distant time to test.
Re:Have to note as a big 'duh' (Score:4, Insightful)
If you didn't want to run them on the same machine for performance reasons before, virtualization will only exacerbate the problem. Virtualization does not speed up anything, it does not magically make multitasking better (in fact, worse). Virtualization can be used for some security separation without having to think at all about it (I would prefer people do the thinking to make it ok to coexist in the same OS instance, the VM thinking just leads to effectively static linked binaries for everything, but practically speaking, people are lazy) or for running disparate OS apps (i.e. Windows and Linux) concurrently with reduced hardware.
Re: (Score:2)
Put simply: which would you rather have servicing millions of small, distributed requests?
a) thousands of 2.8GHz multicore Intel platforms using large amounts of electricity and heat
b) tens of thousands of 800MHz multicore ARM systems using much less than 1/10th the power of the Intel platform, producing significantly less heat?
The answer - to me - is pretty clear. The bottleneck to implementation is the software.
Re: (Score:3, Insightful)
The answer - to me - is pretty clear. The bottleneck to implementation is the software.
I'd rather have whichever one is more cost-effective, and I sincerely doubt it's going to be the ARM solution.
I'd like to see some evidence that an ARM CPU provides two orders of magnitude better processing power/watt than a Xeon CPU.
Then you might want to consider how much power consumption the order of magnitude greater supporting electronics (motherboards, RAM, switches, etc) is going to add to the ARM solution.
*T
Re: (Score:2)
Why would there be more machines?
You just have more CPUs per machine.
It's not magical (Score:3, Insightful)
You don't just plop down a bunch of ARM processors on a board and magically get suitable performance without scaling out memory architecture and such. The only way in the x86 space that very many core systems get acceptable results is by increasingly sophisticated memory architectures that demand more memory modules in aggregate to allow direct, lightly loaded paths between compute and memory. Those memory modules draw more energy, as does various strategies that put more memory controllers down to lighte
Re: (Score:2)
One thing others haven't pointed out yet is there are many other *fixed* power consumers. Like NICs, switches, motherboards in general.
Even with a CPU that is 1/40th the speed, but 1/100th the power is you still have to use 40 times more computers, which also means 40 times more network ports.
You also have to remember a mostly fixed amount of overhead from the DB/OS standpoint. It might only be 1/40th the speed, but you may take a relative performance hit from XXX amount of mips just to calculate indexes an
Re: (Score:2)
However, the ARMs are much smaller physically, and you could probably get a 16-core Cortex-A9 MPCore setup in the same space as a quad-core (or maybe even dual-core) x86 CPU.
Re: (Score:2)
and Intel pretty well dominated everything but an emerging low power market.
I know what day it is, so are you high?
ARM cpus are sold by the container ship full, these things are in everything from coffee makers to cellphones. Intel dominates x86, in the embedded world they are just starting to get a userbase.
Re: (Score:2)
As I said everything *but* an emerging low power market. ARM is the place where they don't have a hold.
Re: (Score:2)
The low power market is not emerging and they do not have a place in several other markets. They own x86 that is all. Embedded is a market far bigger than PCs and servers.
Re: (Score:2)
Fine 'emerging' might have paid a disservice to the rest of embedded, but you have to admit embedded sensibilities are meeting more front-and-center consumer interaction with more general purpose set-top boxes and cell phones going around. Sure, special purpose embedded well behind the scenes has been larger volumes, but the more general purpose in a single device aspect is changing the landscape of embedded.
Re: (Score:2)
Nobody here remember RAQ2? or Acorn? (Score:2)
RISC can do a surprisingly heavy lift, pound for pound...
My old RAQ2 didn't even have heatsinks on the CPU...
Re: (Score:2)
My old RAQ2 didn't even have heatsinks on the CPU...
How amusing that Cobalt went to the K6 because they couldn't keep going forward with MIPS. Further, how amusing that K6 and Pentium are both internally-RISCy.
"Edge" systems? (Score:4, Insightful)
Here's a thought...most pure data retrieval tasks don't require a huge amount of compute power on the device making the request. If I were operating a datacenter with thousands of hits a second, I'd want to optimize for the ability to hold a session open, then offload the request to either a monster data layer or a midrange layer that brokers requests and caches frequent search results.
Something like a single-board computer (or a really scaled-up thin client :-) ) running a low-power processor dedicated to driving network interfaces that also have their own offloading processors would allow them to scale the access layer way up for less power costs. Reliability would be less of a concern too, because you could have tons of cheap devices for the same costs as a fraction of full servers.
When you scale out, you often don't need the overhead that full servers would give you, because you're limiting the tasks that layer of access has to do.
Or...they just want to see how many smartphones it would take to replace layer one of Bing. :-)
I'm waiting for the announcement of Windows Embedded CE 2011 Datacenter Edition.
"own the decision" (Score:5, Insightful)
MBA-ese for "take the blame"?
Re: (Score:2, Informative)
Re:"own the decision" (Score:4, Insightful)
This is meaningless (Score:2)
Even if Microsoft got involved in a low-power-computing market, it will not replace or supplant its core i386 product. Microsoft has a long record of attempting to support non-intel platforms and a long history of dropping that support even if all it required was minimal effort to cross compile and test.
In the end, other devices will compete and win. Most massive hosting activities, especially cloudy ones, are looking to Linux and similar/related technologies to operate them. Microsoft won't be able to c
Re: (Score:2)
Re: (Score:2, Troll)
Re: (Score:2)
Why not use Linux or BSD or ARM?
OSX is pointless on the server.
Re: (Score:2)
Second or should be on.
Damn slashdot not letting you edit your posts.
Re: (Score:2)
Trying to do home automation on an iphone based architecture will save you neither ... even if they released it in a headless mac mini format it would still come with all the existing limitations, to expose the I/O at all you'd have to jailbreak it, trading an ultra cheap plug computer running Linux for a jail broken Apple monstrosity for embedded programming is insane. The clients wouldn't even be touching (or wanting to touch it).
Allowing an iPod touch or iPad as an interface to the home automation system
Re: (Score:1)
Not quite as beefy as a mac mini, but for my own home automation I've found these to be excellent:
http://en.wikipedia.org/wiki/SheevaPlug [wikipedia.org]
http://www.plugcomputer.org/ [plugcomputer.org]
I built a small box with a USB hub IC and a number of parallel port USB chips for IO, which hangs off the plug computer. The thing fits in ceiling tiles and walls, really anywhere you have power and ethernet.
Unless of course you are going for the video output instead of usb, in which case the mac wins hands down.
Re: (Score:1, Interesting)
I'd like to see Apple ship an ARM-based Mac Mini server. That would come in handy for a couple of home-automation projects I'd like to do.
-jcr
Why not look at an OpenRD Base or OpenRd Client? Or a Sheevaplug or GuruPlug Server?
On the other hand if you wait a few months their will be systems based on multi-core ARM SOC running at 2 GHz.
Re: (Score:2)
Re: (Score:2)
I'd like Apple to ship an ARM-based, iPhone OS-based Apple TV with app store support.
Uh-oh, now what? (Score:2)
Translation: "Will somebody from Google please come work for us? We'll give you the keys to the executive washroom!"
I wasn't aware (Score:2)
I wasn't aware that the servers needed protection. Why don't they hire security guards instead of arming the servers? Seems like it would be safer tha way.
logic (Score:2)
ARM NETBOOKS ? (Score:2, Funny)
Oh great... (Score:2)
new hardware expirementation [sic] including SSD and ARM,
Oh great, Microsoft isn't content with their proprietary OS and office suite; they now wants to sell us hardware, but not just any hardware. This hardware is so special that it expires after a period of time.
Re: (Score:2)