Slashdot Mirror
Windows Cluster Edition
eth8686 writes "Microsoft is aiming to have its first cluster version of Windows ready in time for a supercomputing conference this fall." From the article: "The next version of the Compute Cluster edition will extend to Microsoft's .Net programming infrastructure, letting developers write software using the C# programming language, he said."
a thousand blue screens a thousand times faster!
/* oops I accidentally made a comment, sorry */
What is the fundamental difference with the "cluster version" of Windows? OS X clusters just fine and there are no "special editions" other than a few software additions that hardly count as a different OS. And Linux requires very little to get it in a cluster compute configuration.
.Net programming infrastructure, letting developers write software using the C# programming language, he said.
.Net is easier and more secure.
However, Theimer said the cluster version will include some restrictions on how the version can be used to prevent companies from performing standard Web hosting or other functions.
Wow. When you compare this to the standard capability of OS X, it seems like a real rip off. You get reduced functionality. Why?
The first version will reproduce many basic features of Linux clusters, Theimer said.
Then why not use Linux?
The next version of the Compute Cluster edition will extend to Microsoft's
Ah, I see why now. But what impetus is there to use the first version if this is coming in the second version? Kinda like Windows 1.0 I guess.
Although such code runs more slowly than C programs running directly on Windows,
Aauuummm........
writing programs in C# that run atop
Says who? It certainly is/will be easier but more secure is something that has yet to be proven. To date, the track record is not impressive.
Often, Theimer said, it's more important to have a program as soon as possible than to have it running at peak performance, he said.
Ah, the fast food approach to software design. Don't you know that stuff makes you code obese and causes an early demise necessitating frequent checkups?
A third version will include developer improvements to ease programming on clusters. It also will include high-level management tools and will help customers integrate their high-performance computing equipment with the rest of their infrastructure, he said.
This is going to be in the third version of the release? I guess they have been looking at Xgrid, Pooch and other software and it will take them two versions to integrate what others have already got.
Seriously, Microsoft. Please come up with some innovative features and give us something that no other vendor offers or in a package so slick that we cannot help ourselves, but to purchase the Microsoft solution. This is nothing that is not offered elsewhere in the market, but has the appearance of locking us further into a Microsoft paradigm.
You guys have the right idea in that cluster computing is going to be a bigger market than it currently is, but you have to be more hungry and learn again how to ship software that creates desire and meets your customers needs in a timely fashion.
Visit Jonesblog and say hello.
I didn't "see it coming", but it's not unexpected. When was the last time you heard "Windows" and "cluster" in a sentence without some vulgarity attached? Meanwhile, Apple's been in the news with its clusters and is catering to the distributed computing with software like Xgrid and Xsan, not forgetting support for distributed compile in Xcode.
Microsoft is behind on this, and they're now playing catch-up. I suspect we'll see a few cluster-related items from them in the next year.
That what was all this school was for... to teach us how to solve our own problems. -- janeowit
Windows Cluster Fuck Edition?
There, it's been said.
"Look Lois, the two symbols of the Republican Party: an elephant, and a fat white guy who is threatened by change."
Imagine having to reboot a whole cluster after the BSOD.
... Beosloth cluster of those.
Imagine a Beowulf cluster of........
No, wait, it's just too terrible to comprehend.
Ripping an new rectum in the fabric of spacetime.
does it run Linux?
If you're spending $millions on a cluster, it's more useful to spend the money you'd spend licensing MS software on more computers for your cluster.
cost benefit analysis.
oh come on someone was gonna say "beowulf cluster" eventually, and i have "Karma: Excellent" so might as well be me getting flamed
my karma will be here long after I'm gone
If I had mod points, I'd mod myself redundant. We all thought we were soooo clever.
--dv
Insert witty saying or aphorism here.
Every edition of Windows I've ever tried has been a pretty reliable cluster-f*ck, where's the news here ;)
At least, that will be the corporate management perception.
This move could put clustered computing in the mainstream.
Agile Artisans
Imagine a Beowulf cluster of ... oh never mind.
On second thoughts, should this be called a "Grendel" cluster?
Let's see... half the resources, twice the security risk and ten times the price???
Where do I sign up to throw my IT budget down the drain?
This is my sig. There are many like it but this one is mine.
...which is what they built DBSOD (Distributed Blue Screen of Death) on top of. You can't call it windows without BSOD. DBSOD over MPI finally gives windows customers what they've been waiting for.
Too bad there aren't any "I'd hate to be the guy calling up Microsoft to activate all those copies of windows" jokes.
But if there were, "I'd hate to be the guy calling up Microsoft to activate all those copies of windows"
I work in the field (sysadmin for a 800 node cluster), and this is pretty laughable. Microsoft is desperate for the "street cred" of being able to handle high performance computing. Sun, IBM, Dell, HP, Apple all have it. Microsoft doesn't.
If they want so much as the proverbial foot in the door, they must 1) release all (as in *ALL*) of the source code under a GPL or BSD license, 2) make it available for free to all comers, 3) have user's 3rd-party apps (ISE-TCAD, CFDRC, etc) ported, and 4) provide a knowledge base equal to (All Linux + BSD hackers) * Google.
And that only gets their foot in the door.
LOL! No but I'll be designing the T-shirts. :)
:)
Place your orders today.
This is my sig. There are many like it but this one is mine.
Serious scientific computing falls into one of two implementations: Either a tried and true "dusty deck" implementation, or a coding to some new fangled architecture. If Microsoft is really peddling C# as a distributed high performance distributed computing environment, they probably won't win the hearts of the dusty deck people. It will be interesting to see if Microsoft ports something like a parallel Fortran. C# may provide access to a lot of system internals and the .NET framework, but scientists will have to spend more time porting their code then they might want.
In theory, since this is most likely based on a .NET framework and will be a variation of either WindowsXP or Windows Server 2003, and development language supported by the .NET CLR would work for parallel distributed computation.
As someone who has run code on a Cray and who has 45K SETI@home units I say that choice in the clustered supercomputer OS market is good.
It references the original "coming out" article which states:
So now I'm curious... Are they selling to managers, who use the windows i/f and want to think they can "get the job done" on the new server cluster? Or are they trying to suggest that no one in corp uses un*x systems?
I think what Microsoft really needs to do is come up with a line of kitchen appliances. I for one would buy them. I mean, hey, maybe then I could learn how to cook! Imagine having the same interface on the fridge and the coffee maker! Oh sure, some whiny liberal will probably complain that they don't NEED the percolate button for their ice cream, but this is America! Choice is what made this country great!
Note: /. may edit out the
tags.This tagline brought to you by 1500 monkeys in just under 17 years.
We just got our copy in the mail to test it. We been talking to MS about this for about a month now. We're going to do a comparison of it over clustering Linux boxes. Same hardware for both clusters. See how they perform.
DarkMantle I been bored, so I started a blog.
From the article: Theimer said; "We want to be competitive with something like Red Hat."
That shouldn't be a problem. At these prices Windows 2003 is already cheaper. It's only when you start adding CALs that Microsoft gets more expensive and people won't be buying a lot of CALs for a supercomputing cluster.
Looking at the MS licencing, Bill Gates would perhaps be the only person able to afford to run this...
I thought it said Windows Cluster Fuck Edition ..
.. then i realised i was right.
I thought i was wrong
But did i read "Custard Edition"?
HAD
C# is a winner for these kinds of applications: it is far simpler and less error prone than C++, yet it offers crucial features for compute-intensive jobs like value classes, multidimensional arrays, efficient genericity, overloading, an efficient and simple native code interface, and some other language improvements.
Sun really screwed this up with Java: if they had taken the scientific and numerical communities seriously and added the necessary features to Java, Java could be the undisputed winner in this market. Instead, Sun kept Java proprietary, played politics with it, and ultimately turned into a bloated web applications platform.
Sun has been claiming that they will be coming out with a separate Java-like numerical language, but that will likely be too little too late.
*Sits back*
:(
In Soviet Russia, M...ahh fuck it.
Imagine a Beow...bah never mind
BAHAHAHAHAA - Cluster...baaahhahah - oh you guys are serious - Sorry
You could set up a Beowulf cluster of linux boxes to automate that task. Or, if you feel like doing an infinite regress, set up a cluster of Windows for Clusters boxes to automate the process.
After all, I am strangely colored.
Here's a top 500 server that runs windows. Buy a clue thanks.
Have you ever been to a turkish prison?
FTA: "Even Microsoft's Excel can benefit, he said, noting that some businesses have worksheets that can take hours to calculate."
Ah, I knew Microsoft would come up with a new way to force new hardware. Now companies will need a server farm to run Office BWE 2010.
(BWE: BloatWare Edition)
Ruby on Rails Screencast
1's and 0's should be free.
Imagine running Linux under VMware on each node of one of these Windows clusters and using that to implement a Beowulf cluster.
Quoted: "We have developed Windows for Clusters for those computers with significant processing power, but not enough to run Longhorn" a Microsoft spokesperson said.
First of all, scientists are no dummies; C# is a much nicer language than Fortran 77 and (with the right class libraries) nicer than Fortran 9x, and it's not a very complicated language, so they'll probably just use it.
Furthermore, a lot of scientific libraries are now written in C and C++, for which there are already compilers with CLR backends.
But there is no need to recompile: unlike Java, C# and the CLR have very fast and easy to use native interfaces, so you can just keep your existing binaries and call them from C#. This is also important for calling things like PVM and MPI.
But, yes, you probably will also see Fortran-to-CLR compilers.
Technically, I think C# is a great language for scientific and cluster computing, unlike, say, Java. Whether you want to use Microsoft-designed languages, APIs, and/or software is another question.
Cluster != huge expensive computer engine. With software like clusterKNOPPIX (I was just playing with this today), it's really easy to take all of the computers in a research lab (that are already connected to the internet) and turn them into a load-sharing cluster. This is different from a super-computer, although you can presumably get some of the benefit if you are also running MPI (I haven't tried that locally yet). In case you don't understant the purpose of such a cluster it's so that when I want to launch 100 simulations (say to do a parameter sweep, which is embarrassingly parallel), I can launch them all from my local computer and openMosix will automagically distribute the workload across all computers in our lab.
Personally, I'm glad Windows is getting in the game, just like I'm glad when the US gets competition in the space program. Competition, it's a good thing (tm). :)
Ben Hocking
Need a professional organizer?
...at the development lab already running it!
:%s/Open Source/Free Software/g
YTARY!
I followed the link, then downward several levels without seeing any claim this was ever in the top 500. Perhaps I just missed the text. But there were pricing and offers of help, nonetheless, the claims seemed to be circumspect given the performance level of your claim.
Is this my failure, or is your link just FUD too?
You can make a turtle fly at speeds breaking the sound barrier aswell.
Is the turtle designed for that?
Yes.
Is it capable of that?
It's in the Top500 list, isn't it? If it wasn't capable, it wouldn't be doing it. Simple as that.
Please answer those questions
Just did.
So, basically, you implied Windows isn't good enough to run as a Top500 server, someone pointed out that it already does, and now you're defending it by saying, "Even though it is, it's still not good enough?"
This kind of crap really makes the community look immature.
Then use what many in the high performance compupting field do: Fortran. There is at least one advanced C++ development project I know of that has Fortran as its core deep in the bowels of the FFT routines... for efficiency reasons. It's just plain faster.
Plus, how many buffer overflow exploits have you seen recently on Fortran programs? :-)
-- This
Well... they didn't exactly "license" anything from DEC... more like they hired away developers from them (Dave Cutler, et al.)
From The Windows NT article on Wikipedia:
"Microsoft hired a group of developers from Digital Equipment Corporation led by Dave Cutler to build Windows NT, and many elements reflect earlier DEC experience with VMS and RSX-11"
Here's to the crazy ones
Someone is going to spend a lot of time on the telephone dictating license numbers to activate an entire cluster...
If you take a look at the worlds largest supercomputing project, it uses a distributed computing system, and it's for the most part os agnostic, but, the target software has not been compiled for _all_ availble platforms. For reference, check out seti@home. Granted, this project is of a scale that it deserved a customized message architecture, so it's quite unique overall.
Clusters in general are utilized to solve problems in a distributed manner. In the scientific community, MPI is used, and in the web hosting world, clusters are used for load balancing and high availability. The reality is, both of these tasks can be very application specific, and operating system agnostic. In theory, there is no reason a properly written MPI application cannot be deployed on a cluster consisting of half linux, and half windows machines. In reality, such applications tend to rely on artifacts of having identical nodes, and it would be a lot of extra work maintaining a code base such that it can be arbitrarily launched on both platforms. Its far more efficient to tune it up for a single platform, and just use the same platform across the entire cluster.
In the load balancing world, same issues will surface. There is really no reason you cant use a mix of windows and linux based apache systems to back a load balancing cluster. Again, it would be a LOT of extra work managing the mixed configuration, and ultimately, that gets kind of pointless.
Out in the real world, clustering did focus in on linux rather early in the game, because it's open source, hence the folks doing clustering had the option to actually make changes to accomodate thier clusters. There are numerous models to choose from, ranging from a really simple MPI implementation where each machine is virtually independant, and simply passing messages via some high level api, all the way down to the OpenMosix implementation where each machine in the cluster just has the appearance of 'yet another processor' on the overall host. In the former case, applications need to be custom written for the cluster, in the latter case, no modifications are required to applications. Two vastly different architectures, that both fall within the buzzword 'cluster', but are so far removed from each other, there is no similarity other than the fact both use a lot of computers.
A move by microsoft to produce a 'cluster centric' variation of windows actually validates the linux cluster more than anything else can in the marketplace. It demonstrates clearly that the cluster buzzword is gaining enough traction in the management mindset that microsoft needs a presence in that area.
It'll be interesting to see what the final form of the product really is. If it's just a set of gui configurators to manage an MPI system, it's really nothing that couldn't have been done as a third party add-on, and an admission that no third parties were interested in tackling this high end portion of the marketplace on the windows platform.
If the clustering system turns out to be a full process/thread level migration system, akin to the mosix implementation, it'll have a lot of potential, simply because applications do not need to be re-written in order to take advantage of the cluster, assuming ofc, the application already has enough smarts to distribute it's workload amongst multiple processors. the last time I checked (and it's been quite a while), excel is not smart enought to distribute it's calcs amongst multiple processors, something to do with the single threaded nature of serial calculation.
The final proof of technical issues will come over the next few years, and it's going to be an interesting thing to watch. There is going to be a significant amount of support business generated in migrating clusters from one platform to the
Data Center edition goes to eight nodes with up to 32 way SMP - more than needed for just about any database clustering, for example, but not designed for high performance number crunching of dozens or hundreds or thousands of nodes.
Plus, how many buffer overflow exploits have you seen recently on Fortran programs? :-)
:)
Oh, we have them. We work them the same way as unix mail viruses: on the honor system.
hawk
"C has a really crappy track record of being secure"
C doesn't kill people, sloppy programmers kill people.
C is just a language; it is neither secure nor insecure. It depends on how lazy the programmer using it is. I'm thrilled there's a thing called C# that helps sloppy programmers. Warn me if anybody writes an OS in that bloated pig.
But C by and of itself dosn't mean code is insecure.
Need Mercedes parts ?
While true, an MS operating system license doesn't cover support. So you would be paying up the nose for software, and paying up the nose for support as well.
After all, I am strangely colored.
Am I the only one that read this and think Theimer and MS have no clue what clusters are designed to do? Using a spreadsheet to do lots of number crunching is an application problem. You are using an application in ways that it was not intended. That's far different than an OS performance issue where your application is limited by what the OS can do and need some way to tweak the OS to perform better. How is MS Cluster going to solve this spreadsheet problem? Is that spreadsheet going to run across hundreds of machines? My solution would be not to a use a spreadsheet in the first place. There must be some other more suitable application.
I may be cynical but I see this as another ploy by MS to expand their revenue by bloat. As their software gets more bloated, the average consumer will be forced to use newer and more powerful OS to run it. I'm sure that MS will be tickled in 2010 when you need a 64 way processor just to run BackOffice 2009 so they can charge your company for a license for each of the the 64 processors.
Well, there's spam egg sausage and spam, that's not got much spam in it.
I am a physiker (Physicist?) and worked with a lot of mathematiker (mathematics?). Until recently I programmed on big irons & clusters. We all used only one thingas : Fortran. We programmed on cray, On Origin, on cluster. And we (many groups over many countries) used Fortran. Why is that ? 40 Years worth of mathematical libraries. Extremly optimised compiler code, generating machine code as packed as direct assembler code. And a community you can talk to to get code snippet or code solutions hand on. MPI with years over years of experience.
... And forcing another language... Only a manager would buy such a system for scientifics, and the scientific would snicker while formating all HD and install an OS where High performant Fortran does exists.
Fat chance in hell any of us use anything *BUT* fortran. Why should we care for new language which do NOT BRING ANYTHING to us and force us to port many many libraries, debug again what is mostly now bug free, and start over ? Man. Get a clue. And Make a MS-FOrtran for that cluster and MAYBE if for the same price we get better performance you might get a chance. But for worst performance and same price
As for security... What security ? By the time your are on the cluster you should already have been thru with security, through a front end. The Cluster is to be used for high performance calculation NOT for securly checking bounds of arrays.
Frankly I think this offer is only directed towward marketing/enterprise which use their cluster for ANYTHING but mathematics.
C. Sagan : A demon haunted world:
http://www.amazon.com/gp/product/0345409469/
visit randi.org
Imagine a beowulf cluster of BSODs.
There. I said it. Happy?
Give me my freedom, and I'll take care of my own security, thank you.