[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]
... > Take it that would work a bit like Folding @ Home or Seti with each > client (of different spec) working on portions of data? You can do that sort of task without clustering - they're jobs that are easy enough to break into small tasks. One of the things that appeals to me about clustering is the idea of it looking like one machine - ie one web server or one database server or one file server (or all of them) while actually being a (possibly dynamic) collection of machines. Almost all the implementations seem to be symmetric though. I'd like to see something that would allow an organisation to seamlessly grow dynamically - ie you can start of with a one machine (say) and then just add/remove things on as you can afford them and the cluster self reorganises. Now you can do that with a lot of clustering stuff but only with symmetric hardware. Tanenbaum had some good ideas on this but I think they got lost in the Minix/Linux 'wars' Tom te tom te tom > > Rob -- The Mailing List for the Devon & Cornwall LUG http://mailman.dclug.org.uk/listinfo/list FAQ: http://www.dcglug.org.uk/linux_adm/list-faq.html