Please create an account to participate in the Slashdot moderation system

 


Forgot your password?
Close
typodupeerror
×
IBM

IBM (Offically) Launches Linux Box Clustering 136

Posted by Hemos from the the-official-story dept.
Neumsy writes " IBM has offically announced that it is releasing Linux-Based Server, Software Packages. Yahoo! News has the story . According to IBM, this will expand the use of Linux. It's a good overview article. Not too much in depth, but still nice to see Linux getting out there more and more."
This discussion has been archived. No new comments can be posted.

IBM (Offically) Launches Linux Box Clustering More

IBM (Offically) Launches Linux Box Clustering

Comments Filter:

  • The commercials weren't lying? (Score:4, Funny)

    by FortKnox ( 169099 ) on Tuesday November 13, 2001 @02:30PM (#2559764) Homepage Journal
    The commercials weren't lying about IBM selling Linux servers?!?!?!?!

    Maybe I outta start believing marketing people? Nahh.... ;-)
    • Maybe I outta start believing marketing people?

      I have this bridge that's going on auction, cheap.
    • I'm not fully positive (I overheard them just two or three times) but I sort of remember that initially the commercial ended with something like (translating to english):
      A : "what is Linux?"
      B : "it makes you save money"
      Now it says:
      A : "what is a server?"
      B : "it makes you save money"

      Can anybody confirm/deny that ?

      (BTW this sort of 'ads fixing' seems to be a standard practice: I've seen it happen lots of time. They sort of analyse people reaction and 'fine tune' the ad accordingly)

  • Since Quake III Arena is multi-threaded, would it be possible to run it on some sort of cluster? If not, or if there is, is there any other game that is capable of playing on a cluster?

    Mind you, it would need a fair bit of bandwidth I'd bet.
    • I doubt it, unless its a cluster made by SGI ;)
      (You could do that with the Onix)
      Because most of the stuff is done by GFX card ..
    • I'm not sure if it would work. But it damn sure sounds like a fun (albeit expensive) project.
    • Well...
      The main factor in determining your q3 performance is not processor speed, but the bandwidth between the processor and video card (and the video card and the actual display buffer).

      Even if you could thread quake through a cluster, it would make no difference.

      As for threading.. these clusters are not hte same thing as a multirpocessor machine... unless they implement shared memory and such across the network.. which would be slow.
      • Just throwing some ideas around here.

        I thought Q3A game quality depended on the CPU to video card bandwidth as well as the 3d processing power of the processor. I assumed as well that the original poster's idea was to offload all the 3D computations off of the video card and onto the system (which he planned to cluster. Let's assume that his cluster had high bandwidth between them)

        In an entirely 2D application, how fast can the system update the ENTIRE video buffer (all 1024x768 pixels say) in a given amount of time?

        If the processing was offloaded to the system and all one needed to do was to reflect the changes in the video buffer, wouldn't that help to alleviate the 3D processing bottleneck?
      • quoth mindstrm:

        The main factor in determining your q3 performance is not processor speed, but the bandwidth between the processor and video card (and the video card and the actual display buffer).

        er... iirc, q3a is actually directing calls to the machines opengl subsystem, so it stands to reason that if you moved that opengl processing from the video subsystem back to the cpu, you might be able to reap some type of clustering benefit.

        otoh, i'd be surprised if that type of setup could work; q3a is built so carefully that a major change like that would probably be detrimental.

        maybe The Carmack will post the answer.
        • Latency would still be too high, I'm thinking.
          Yes, you could take the hugest cluster in the world and do massive computations, really fast...
          but the time between feeding the equation to the machine and getting the answer is what's at stake.

          It may reduce a 100 year problem to 10 seconds.. but it can't reduce a 1 second problem to a hundredth of a second.

          Part of the reason that modern 3d video cards get such performance is the high-speed bus between the video processors doing all the 3d work and the video ram itself.
    • Q3 itself isn't really multithreaded in the way that you are thinking. There is a way to let the video work in a seperate thread, but the actual game itself (AI, physics, sound, etc.) are not. But... the great Carmack has said that the upcoming Doom sequel will be fully multithreaded, and will be terrific on multi-processor systems.

      steve

    • The Holy Grail of CLustering (Score:1, Funny)

      by Anonymous Coward
      The Holy Grail of CLustering is being able to migrate an open socket from one machine to another without disrupting it. I have a trivial solution but there is not enough room in this Windows message box.

  • Hopefully... (Score:2, Funny)

    by nob ( 244898 )
    ...this business move won't put IBM too much into "debth."

  • Linux Myths too - full page ad (Score:5, Informative)

    by ethereal ( 13958 ) on Tuesday November 13, 2001 @02:37PM (#2559807) Journal

    Interestingly enough, just this morning I saw a two-page ad for IBM servers running linux. I haven't found the actual ad online, but it showed the famous "bigfoot" photo, labeled as fake, and then a penguin walking through the server room in the same pose, labeled as real. The other page of the ad was an abbreviated list of the usual Linux myths that we all know and love, with IBM-specific arguments as to why these were no longer true. This is the real fruit of the $1 billion campaign from IBM, and a great answer for your hesitant management.

    IBM's main page for this, aimed at upper brass rather than engineering, is at http://www.ibm.com/linux/cio2 [ibm.com], and the myths seemed to come from this brochure: http://www.ibm.com/linux/Demystifying_Linux_Brochu re.pdf [ibm.com].

    Maybe this isn't entirely on-topic, but I thought it was a great example of some more of that good mindshare. And this time IBM isn't going to have to scrub off any sidewalk paint :)

  • Price and performance less than Unix... (Score:5, Funny)

    by Skater ( 41976 ) on Tuesday November 13, 2001 @02:38PM (#2559814) Homepage Journal

    "Linux is appealing because of its price and performance, which can be less than with a comparable Unix-based system, Quandt said."

    Sounds great...oh, wait, no it doesn't.

  • The beauty of clustering (Score:4, Interesting)

    by Anonymous Coward on Tuesday November 13, 2001 @02:38PM (#2559815)
    IBM has a winner. For organizations which can use more computing muscle, particularly in scientific computing, clustering is true blessing. The real beauty is that an organization can "test the waters" without an outrageous investment in hardware. When budgets are thin, clustering allows you to start small, and then ramp up as the benefits warrant it financially. Here is a lucrative niche for which IBM's expertise is ideally suited.

    • Three types of cluster (Score:2, Informative)

      by Macka ( 9388 )


      Clusters are clusters, but not all clusters are equal.


      You have high speed compute clusters like Beowulf.

      Then there's the first type of High Availability clusters -- the hot standby/failover configuration, where services and storage on one system are reassigned and restarted by another if the first goes away. Most Unix and Linux implementations haven't got beyond this yet.

      Then there's the more grown up version of High Availability clusters, where all the clustered systems have concurrent access to the same storage, cluster wide lock management, and can run multiple cooperating instances of the same application on all systems. Like Oracle Parallel Server. VMS pioneered this ; Tru64 Unix has it now, and Linux is working towards it with GFS (the Global FileSystem).

      GFS has got to be one of the most exciting current Linux developments in my book. I've had a taste of this kind of clustering on Tru64 Unix, and believe me, once you've experienced it you don't want to go back.

      • GFS [sistina.com] is now commercial-only (fairly "cheap" from a business point of view - 1000 US$/node before rebates and special offers, but completely unreasonable for home hackers who'd like to try it for the heck of it - might be fun trying to combine it with iSCSI).

        OpenGFS [opengfs.org] seems to have taken off in the free software side of the camp.

        Cluster-wide locking requires applications understanding it, so it's not easy. I'm not sure, for instance, what would MySQL (a popular app that might benefit from this) if two processes tried to access the same storage read-write concurrently - even if the locking semantics were perfectly implemented by the filesystem.

        Where I work I'm trying to set up a mixed active/standby+active/active configuration (shared Fibre Channel-connected storage, applications that can run independently do so, and those who cannot run in hot-standby). I'm almost ready to go live (glee).
  • Its always refreshing to hear corps stress the benefits of this OS ..

    http://www-4.ibm.com/software/is/mp/linux/softwa re /

    • > Its always refreshing to hear corps stress the benefits of this OS ..

      Maybe old news, but yesterday I noticed an IBM/Linux commercial on the telly for the first time. I wasn't paying attention when it came on, but the gig was that some execs walked into their datacenter and were shocked to see a big empty room -- thought someone had stolen all their servers. But IBM had merely replaced them all with a single Linux box.
      • But IBM had merely replaced them all with a single Linux box.

        I've seen the commercial too, but couldn't tell if that was a rack full of Linux boxen or an IBM mainframe. I'm guessing the former, but the use of the singular caused me to question my assumption. Anybody know by lookin'? We've got IBM boxen here, but they're all AIX and Windows (except the one on my desk....).
  • Well acording to this article, it sounds like Corp America isnt tech savy enough to run linux without nice gui point and click conf interfaces. Figures...

    • That is definitely not news. These are people who have worked long and hard to be able to turn on and off their Windows and Mac machines :)

      • Linux is appealing because of its price and performance, which can be less than with a comparable Unix-based system, Quandt said.

        Then Windows NT must be the IDEAL choice. Less expensive with less performance... No wonder Linux is having a hard time ;)
    • Nice attitude. Gee, I wonder why people are so reluctant to switch to Linux! Is this the kind of "support" that so many Linux based companies are going out of business trying to sell?

  • The official news link (Score:3, Interesting)

    by Anonymous Coward on Tuesday November 13, 2001 @02:42PM (#2559839)
    Here is the link to the IBM site. Unfortunately, not much details. However, it seems IBM follows a clustering strategy which includes Linux. It is not a Linux only strategy.

    http://www.ibm.com/news/us/2001/11/13.html [ibm.com]

  • I wonder whether IBM will bring out bewoulf clusters, It is too hard to get any good software for bewoulf. Maybe IBM may finally give up unix and come home to linux. Ps : I am looking for a 3d Rendering farm using clusters any idea ?

  • ms fighting linux expansion (Score:3, Interesting)

    by Anonymous Coward on Tuesday November 13, 2001 @02:56PM (#2559918)
    It's good to see more linux successes. MS has really been perking up though. Take, for example, the recent link here to the leaked memo to theregister where linux is stated to be the threat. http://slashdot.org/articles/01/11/12/136243.shtml [slashdot.org] MS has really been perking up. Check out this little tidbit; looks like MS is now dumping copies of winxp at universities to try and combat the fact that intelligent people are using other systems. http://www.browndailyherald.com/stories.cfm?S=0&ID =5668 [browndailyherald.com] The MS drones are already out in force; you should leave a little feedback...

    • Not only that, they're calling it "free software" all over the place. Those bastards! :)

    • The article about Brown points out that the students do not get the MS-Office package. MS-Office being the main reason most people keep Microsoft products on the desktop. To top it off, they don't even get any support.

      It looks like Bill's now desperate to hook students into paying for MS-Passport.

      As an example, the Nov. 13 issue of PC Magazine points out that even Microsoft's online privacy statement Generator requires hooking users first to MS-Passport. In contrast, IBM's generator does not.

      So, it is good to see IBM using positive examples. If enough new people give up on pathways that lead to guaranteed vendor lock-in, then we'll see even more Open Source and Free Software and more useful applications.

  • And the obligatory comment... (Score:3, Funny)

    by jht ( 5006 ) on Tuesday November 13, 2001 @03:01PM (#2559944) Homepage Journal
    Boy, I'd really like to see a Beowulf cluster of those!

    What? Oh, okay - never mind...
  • Linux is appealing because of its price and performance, which can be less than with a comparable Unix-based system, Quandt said.

    Less performance!! Woohoo!!

    Reminds me of the Simpsons when Homer starts his own (internet?) company: "I'll make myself president!! No.... VICE president!!!"

  • Tivoli on linux? (Score:1, Insightful)

    by Jon_E ( 148226 )
    Isn't this article more about Tivoli on linux [tivoli.com] and less about clustering?

    Keep in mind it's IBM software on linux that IBM promotes more than linux (partly because AIX is so lame and was never that widely adopted) ..
  • This is great. Hopefully it'll help to bring Linux into the corporate world more as a mainstream solution. I find that with about 90% of the solutions I come up with, I'm met with hostility, disdain or disbelief at my suggest of using 'An unproven, unsupported OS'.

    Thank god! Mabye my job will be a bit easier now.
  • To the term clusterf**k.

    At least in Redmond...
  • Its insteresting how the author chose to avoid the obvious fact that linux competes with windows.

    Linux is free open-source software, which means that all code is public and can be adapted by companies and individuals. It competes with other operating systems, such as the popular Unix operating system.
    • Its insteresting how the author chose to avoid the obvious fact that linux competes with windows.

      Not according to Linus.
    • I wouldn't say that Linux was a direct competitor against MS. Instead, I'd say it was an alternative.

      Competitor implies that they're working towards a common goal, just using different methods.

      • Competitor implies that they're working towards a common goal, just using different methods

        But isn't that exactly the case? MS just doesn't actually say their goal is world domination. Both camps are "working towards the same common goal, just using different methods".

  • IBM Clusters at NCSA (Score:5, Informative)

    by dlapine ( 131282 ) <lapine@illinois.eREDHATdu minus distro> on Tuesday November 13, 2001 @03:33PM (#2560093) Homepage
    We already have two here. Check out:
    http://access.ncsa.uiuc.edu/Releases/011018.Titan. html

    Top500 org (http://www.top500.org) has the latest rankings out and Platinum is ranked #41 in the world with a 594 rating and Titan is at #34 with a 677 rating. These are not slow systems, but they do require a knowledgeable support staff. Both systems are IBM "out-of-the-box" clusters, running RedHat Linux versions.

    I'm just glad to see linux advertised in the mainstream media. Name recognition helps.

    More information is available at the main website: http://ncsa.uiuc.edu

  • What most people fail to understand about this type of cluster (high-performance, rather than high-availability) is that your apps don't *magically* become faster because you have N more nodes to run them on. Cluster applications require coding to an interface such as MPI or PVM to reap the benefits of the hardware. I don't know of any MPI-aware Quake or Apache servers though ;-)
    • Well, if it's something like Apache, then you just put a traffic director (like the Linux Virtual Server) in front of the cluster, and yes, it is as simple as plugging in N more machines, and up to the limit of your traffic director, it will scale pretty much linearly. And if you can't get your hands on a machine or two to use as traffic directors, even round-robin DNS can do somewhat effective load-balancing.

      Not magic, but darn nice.

      steve
  • Clustering is such a fascinating area on its own and the article is so shallow that I'm curious why they've published it at all. They could mention the potential benefit [port.ac.uk] of cluster computing [nec.com] as well as examples of some working clusters like Beowulf [beowulf.org] or Mosix [mosix.org] or even the famous fact that there is a cluster [theregister.co.uk] among top 500 supercomputers.
    • Why is IBM important? Mainly, because when a company like IBM trusts and supports Linux, corporations are much more willing to listen. What does that mean for Linux? It means a lot more people working not only on Linux itself, but on all Linux-related software. Kernel enhancements are only the beginning.

      Look at what happened with SGI - they start working on porting Linux to their machines, we get their coders porting XFS to Linux. IBM ports Linux to their machines, we get AFS, and more is in the works. IBM is also active in helping Linux run efficiently on "big iron". Companies whose business model revolves solely around Linux have been dropping like flies, but companies like IBM and SGI have the ability to keep the Linux side running, and give a lot back to the community.

      steve
      • I agree as long as you talking about Linux. Nonehteless, apparently, this article was designed to project it on truly exciting field of cluster computing, where it, IMHO, fails.

        As for Linux it's simply a tool of choice being among a small number of operating systems able to facilitate such an extensive research that clustering (especially the branch of global computing [gridforum.org]) necessitates.
    • Actually, your reference is a little old. There are at least 3 clusters in the top 50! We have 2 of em! :)

      I like the fact that one of our clusters is composed dual pentium PIII at 1.0GHZ- nothing you couldn't buy yourself. Ours are in a 1u form factor so that we can fit 30+ in a rack, but otherwise the basic node isn't that far out.

      We do use some fancy networking equipment to pump up the message passing speed, and those NIC's aren't cheap, but what we have in production is inexpensive and effective for a whole new level group of potential users- those whom CRAY and SGI ignore.

  • Am I the only one who noticed that throughout the article, Linux was only compared to UNIX, and there wasn't a single mention of MS?

    Really? I was?!

    woof.

    Every two weeks (max) I have to explain to some "Administrator" how to make Microsoft Clustering work with various software. I haven't ever in my life had to explain it to someone working in UNIX or Linux. Any ideas why?

    • Maybe it's because IBM makes so much money from selling Windows hardware and software? AS Linux supporters go, they are stunningly pro-Microsoft. I've noticed they insist on making constant distinction between Linux and UNIX when, compared with everything else, Linux and UNIX are blood brothers.
  • Hey everyone, check out this [ibm.com] Redbook, published by IBM. I've loved their Redbooks for years, and now that I work for IBM e-business, I've been asked to study this guide before I start deploying HPC clusters.

  • Uhhh... IBM has been selling clusters for a while (Score:5, Informative)

    by Hedonistic BOFH ( 27550 ) on Tuesday November 13, 2001 @05:53PM (#2560736) Homepage

    The announcement may make it 'official', but IBM has been selling turn-key style clusters for over a year now.

    How do I know, and why do I care? 'Cause I work for IBM, and design and build Linux clusters for a living.

    We rack purpose-built 1U's (x330's) built for easy large-scale mangement (built-in daisy-chain KVM capabilities, integrated service processor network, cool blinky lights, etc, etc), have internally developed mangement software and system imaging solutions, and ship them to customers at a point where all they need to do is plug them in and log on.

    The only real issue is almost no one has a 'standard' cluster application. Almost every one uses a different IPC mechanism, and usually an app is only validated against one very specific software image. So to get the most out of a cluster, customers still need to spend time tweaking.

    Maybe someday soon, we'll all be able to 'apt-get install damned_big_cluster', but until then... <shameless>just call IBM.</shameless>

    --Matthew
    slashdot at sigalrm dot com
  • What I find interesting about IBM's Demystifying Linux Brochure [ibm.com] is that it seems to be wholly honest. I couldn't see any attempt to mislead. It paints things positively, but nothing beyond that. The facts speak largely alone.


    Compare this with Microsoft's advertising, which is sometimes grossly misleading, occasionally even other-worldly. I feel good about IBM being on our side. Let's hope lots of Slashdot'ers forward this to their top IT management.

  • If they keep promoting these servers as in this commercial [ibm.com] then at least we'll get a couple of good laughs out of it :)
  • that a utility to check your cluster's file system is called:

    Clusters.fsck?
  • It's definately refreshing to see Linux getting some mainstream advertising, and it surely raises some hopes as to the future of Linux distribution. Unfortuately, whether or not this is significant in any way and whether or not it is representative of any sort of development can only be seen through time. Breaking open the OS market is something that Linux isn't going to do anytime soon. Possibly in some years, we might see this happen, but I'm afraid it's the sort of thing that, barring any absolutely huge events which completely revolutionize the OS market, will take a very long time and I remain unconvinced as to how important this actually is.
  • Well the reason that company's want a point and click based OS is because they can hire cheaper workers and there is less chance of major problems since most people use point and click OS's at home.

Slashdot Top Deals

If it wasn't for Newton, we wouldn't have to eat bruised apples.

Close