Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Supercomputing Software Linux

Linux Distro turns PCs into Night-time Clusters 200

renai42 writes "An Australian security firm is about to launch a clustered Linux distribution based on openMosix that aims to utilise the unused nightly processing power of corporate desktops. Dubbed CHAOS, the distro is able to remotely boot a computer and run it on Linux without affecting the local hard disk. CHAOS is designed to provide dumb node power to a cluster run by existing full-featured clustering distributions such as Quantian and ClusterKnoppix."
This discussion has been archived. No new comments can be posted.

Linux Distro turns PCs into Night-time Clusters

Comments Filter:
  • by Soko ( 17987 ) on Friday April 08, 2005 @03:56AM (#12173866) Homepage
    No.

    Most entreprise level desktops have Wake On LAN and PXE boot capability. You send a magic packet to each desktop to wakr it up, and then tell the PXE BIOS to boot ClusterKNoppix via TFTP.

    It's not that hard to do, even for lazy sysadmins.

    Soko
  • by Jessta ( 666101 ) on Friday April 08, 2005 @04:20AM (#12173953) Homepage
    We call the solution PXE booting. Never trust users to do anything.
  • by davedx ( 861162 ) on Friday April 08, 2005 @04:23AM (#12173957) Homepage
    Not sure what kind of distributed computing you can really do over latency measured in milliseconds. One of the big bottlenecks for today's supercomputers is bus/shared memory access time. I can't really see this being useful for much more than we already do - SETI@Home and so on, where you send packets to be processed and after a few hours the node sends them back.

    So yeah not sure if we could ever have a true supercomputer distributed over the net (as it is now, with the light speed as it is!) that's parallel in real time.
  • for information (Score:4, Informative)

    by cotyx ( 699460 ) on Friday April 08, 2005 @04:26AM (#12173963)
    Hello For information this kind of stuff already exists, from long time. I invite you to visit this webpage : http://www.lri.fr/~fedak/XtremWeb/introduction.php 3 [www.lri.fr] Regards
  • Re:Seriously?? (Score:3, Informative)

    by benjamindees ( 441808 ) on Friday April 08, 2005 @04:42AM (#12174018) Homepage
    allow their PCs to be booted remotely

    The actual booting could be controlled locally.

    used for a task outside their control

    Yeah, I'd want to see some security measures in place, like running it in User Mode Linux or something. A dedicated client program like SETI@Home is one thing. A full OS with the capability to fsck with your hardware is another.

    which doesn't make them any money.

    But it could help save them money. Lots of OSS users have no viable way to contribute back to their favorite projects. Lots of projects could be helped by a vast pool of computing power "on tap". Surely somebody could come up with some interesting applications for a ridiculous amount of free CPU time?

  • by evilviper ( 135110 ) on Friday April 08, 2005 @04:44AM (#12174027) Journal
    Most entreprise level desktops have Wake On LAN and PXE boot capability. You send a magic packet to each desktop to wakr it up, and then tell the PXE BIOS to boot ClusterKNoppix via TFTP.

    It doesn't sound like you've tried this. W.O.L. doesn't power-up the system when it's been shut-off, so it's really not of any use in this situation.

    PXE should be almost all you need... Set the machines to boot from the NIC first, and HDD second, but leave the Bootp and TFTP server off during the day... At night, turn on the netboot servers, and just reboot all the machines. You could either reboot them remotely, set them to automatically reset at a certain time, or just have employees hit the reset button at the end of the day.
  • by Gollum ( 35049 ) on Friday April 08, 2005 @04:52AM (#12174054)
    Here is a suggestion that would allow computers that are not in use to be "co-opted" for use in the cluster.

    Identify the PC's that COULD theoretically be used, and collect their MAC addresses. Also, configure them to try netboot first, then fall back to booting from the hard drive.

    When you want to perform computations, send a WakeOnLAN packet targeted to each of these computers. Wait for netboot solicitations, then, if you have recently sent a WOL packet to that computer, respond with an appropriate netboot directive, booting the PC into a cluster node configuration, with all details loaded from the cluster director.

    Otherwise, allow the netboot solicitation to time out, and the computer will boot into its normal configuration.

    Not sure how OpenMosix handles nodes that simply vanish, but users could simply reboot the PC when they arrive in the morning, if the computation is still ongoing. Otherwise, the cluster director could remote shutdown/reboot each node prior to the user arriving at work.

    Unused PC's would not consume power, cluster node PC's could be instructed to immediately drop the monitor into Power-save mode, etc.

    The cluster director could decide how many nodes to start, or the location of the nodes, to optimise the comms between it and the servers.

    An idea with potential, I think!
  • by Anonymous Coward on Friday April 08, 2005 @05:00AM (#12174080)
    WOL can definitely remote-start ATX machines that have been shut down. Requires support from LAN card/chip and motherboard, but most corporate desktops support the feature.
  • by pe1chl ( 90186 ) on Friday April 08, 2005 @05:20AM (#12174143)
    W.O.L. doesn't power-up the system when it's been shut-off, so it's really not of any use in this situation.

    It doesn't sound like you've tried this.
    When configured correctly, it works. We do weekly maintenance and nightly installations of software that way. In some scheduled job, all systems get a wake-on-lan packet and they start, and run some install. The users are never bothered with it, unless their systems are offline at that time (e.g. laptops).
  • Re:Go DownUnder! (Score:1, Informative)

    by Anonymous Coward on Friday April 08, 2005 @05:20AM (#12174145)
    "Australian inventions from all decades of the twentieth century are included. Many are well known -- like the Ute (1934), the Hills Hoist (1948), the IVF freeze-thaw method for storing embryos (1983), the Triton portable multipurpose workbench (1976), the Wiltshire Staysharp knife (1970), and anti-counterfeiting technology for banknotes (1992)."

    "Other Australian inventions are more surprising. Almost every office in the country has a wall-mounted Miniboil (1981), aeroplanes worldwide carry black box flight recorders (1961), and the concept of Racecam TV sport coverage was also developed here (1979)."


    From

    http://www.abc.net.au/science/news/stories/s421099 .htm [abc.net.au]
  • Re:Useful? (Score:4, Informative)

    by dario_moreno ( 263767 ) on Friday April 08, 2005 @06:19AM (#12174337) Journal
    search for "Warewulf" clusters (turning into a Beowulf at night)...it's quite old news ahref=http://warewulf.lbl.gov/ [slashdot.org]http://warewulf.lbl. gov/>
  • by Hast ( 24833 ) on Friday April 08, 2005 @07:08AM (#12174484)
    It isn't a software thing. It's done in hardware.

    You need a network card which supports it as well as a mainboard which supports it (or with built in networking, that usually supports it).

    To start it up you send a "magic" package to the NIC which tells it to boot. AFAIK it's just MAC level package with all FF in the data field or something like that. The NIC will then boot the computer just as if you had pressed the power key.
  • Re:Useful? (Score:5, Informative)

    by TamMan2000 ( 578899 ) on Friday April 08, 2005 @07:19AM (#12174524) Journal
    Pratt & Whitney [sc-conference.org], one of the big 3 jet engine makers, has been doing that for over a decade. It is there primary means of supercomputing.

    They have been at it so long that they had to write their own message passing system (PROWESS) because MPI was not there yet.

    I used to work for them as a computational fluid dynamicist, we were the main consumers of this "cluster".
  • by yecrom2 ( 461240 ) on Friday April 08, 2005 @07:55AM (#12174664)
    Unfortunately, CHAOS isn't one of them. There was an article on CHAOS in Linux Magazine in 1996-97 somewhere. It stood for CHeap Array of Obsolete Systems. The author put together a set of 386, 486, and Pentium boxes that he bought bundled on a pallet. I think he used slackware and beowulf, but in the end, it actually had some pretty significant computing power. The computational power/Kwatt hour ratio wasn't very good though. I wonder if he ever had to run his furnace in the winter?

    I've got the Cheap and Obsolete part of his setup already, but not setup in an array.

    yecrom2
  • Re:WTF??? (Score:4, Informative)

    by Professeur Shadoko ( 230027 ) on Friday April 08, 2005 @09:02AM (#12175010)
    This is at night. Elecricity is cheap at night.
    Well, at least in my country, where nuclear power plants like to have a steady load.

    Computing on workstations at night is probably waaaaaaaaaaay cheaper than on a supercomputer during the day, then ;-)

"The one charm of marriage is that it makes a life of deception a neccessity." - Oscar Wilde

Working...