[SATLUG] Next month's presentation
tonybeau at cajunbro.com
Wed Aug 28 18:17:01 CDT 2002
Yep, the "Seti at Home" type of computing is something that I have been
experimenting with for the FDTD type software.
It is immune to a single node failure (when setup correctly). At most,
depending on how the program is written, there could be a short delay
while the missed data is replaced. (SETI sends out multiple copies of
data sets to different servers as a matter of course.)
Mosix is a good way to get the multiple copies of the client program
running on the different nodes. Just start them on the master, and let
them distribute themselves out. If a node fails, the data is sent to a
known node, and the problem node will receive another set, if it
requests it. After so many failures, the node is marked as bad,
requests from it are ignored, and the administrator is notified.
Just the results of my experimentation...
I think I'll show up for the next meeting also. Could be fun!
On Wed, 2002-08-28 at 16:35, John Ziriax wrote:
> Please keep us posted. Althought, I doubt if the AF and Navy would approve
> of a giant Quake game.......unless....Never mind.
> Yes, the death or mere glitch of a single node can be a killer. LAM at least
> good at telling you whose alive at the start. I confess to a daily check on
> the progress of the current run. At low frequencies (70 MHz) it takes those
> 60 nodes about 26 hours to simulate one wave length.
> What about grid computing? Seti at home style. It is certainly immune to single
> node failure. With dedicated systems you could keep things in sync easier.
> On Tuesday 27 August 2002 09:52 pm, you wrote:
> > I agree, this would be a very interesting presentation. Personally my
> > company (Global Gaming Innovations) is working with wide area clustering.
> > My interest would be peaked and I *might* actually make it to my first
> > SATLUG meeting. That's true that creating a cluster is easy, although to
> > be honest, you DONT need MPI or PVM to do it. The tools are already
> > present in TCP/IP protocol programing. Building a cluster doesn't have to
> > require installing or configuring anything!!!
> > However, if you wanna get into fun things like multi-threaded parrallel
> > computing, you're gonna want to use the kitchen sink of distributed
> > parallel computing (MPI). There's one major draw back in today's tech that
> > is on the market. That is, try starting up your program and pull the plug
> > on ANY of the nodes in the LAM, whops all gone, with no signal. There's
> > ways to send heartbeats and Duke University is developing methods of
> > gracefully shutting down your distributed program with error codes.
> > However, this is currently a major pain in the arse. Also for more about
> > wide area distributed processing check out the Albatras (sp?) project.
> > It's looking pretty good (4 universities making up shared clusters).
> > Oh and if there is interest, my company is about to release its first beta
> > test (will work natively in linux) of a FPS game that will use the
> > distributed processing tech to create massive multiplayer online games.
> > (read as 500-1000 or more quakster types fragging each other on one
> > server). I'll post an anouncement incase anyone is interested in seeing
> > this type of stuff work.
> They that can give up essential liberty
> to obtain a little temporary safety
> deserve neither liberty nor safety.
> Ben Franklin, Historical Review of Pennsylvania, 1759
> Satlug mailing list
> Satlug at satlug.org
More information about the Satlug