9fans - fans of the OS Plan 9 from Bell Labs
 help / color / mirror / Atom feed
From: Eric Lee Green <eric@badtux.org>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] Plan9 vs Beowulf
Date: Fri, 20 Apr 2001 07:30:03 -0700	[thread overview]
Message-ID: <3AE047EB.D8533377@badtux.org> (raw)
In-Reply-To: <20010419234118.A986@localhostnl.demon.nl>

William Staniewicz wrote:
> Is there any comparison between a Plan 9 network
> and a Beowulf cluster?

Not really. There is no such thing as a "Beowulf cluster" (and I say
that having designed and implemented one in the late 1990's when I
worked for Linux Hardware Solutions). What there is, is a set of
management tools to allow doing various system administration tasks in
parallel (rather than having to log into 1024 different computers!), and
several different hacks to the standard Unix "fork"/shared memory model
to allow that program to be "forked" to a different node on the network
and allow  that memory to be "shared" with other computers on the
network rather than just be local. It isn't really shared, of course.
Programs have to be specially compiled to use one of these memory
managers and must have the calls added to ship the memory. And it really
does not work very efficiently at all. That's why this is mostly used
for things like signal processing where the task can be broken up into
dozens of independent tasks that communicate only when they hit the edge
of their little piece of the big picture. It is not a general purpose
clustering technology.

A Plan 9 network is a set of services on a network, but the services
communicate via standard networking protocols, not via a fake Unix
shared memory manager. So Plan 9 is basically a different beast. Each
Beowulf node is in fact a standalone system, running all services that
it needs to execute programs, and the programs are what run in a
distributed fashion. A Plan 9 node may run services distributed all over
the network, but the services themselves aren't parallelized across
nodes, and programs you run on a Plan 9 network run on a particular
node.

--
Eric Lee Green  http://www.badtux.org  mailto:eric@badtux.org
     Phoenix Branch -- Eric Conspiracy Secret Labs
              Cruisin' the USENET since 1985


      reply	other threads:[~2001-04-20 14:30 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2001-04-19 23:41 William Staniewicz
2001-04-20 14:30 ` Eric Lee Green [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=3AE047EB.D8533377@badtux.org \
    --to=eric@badtux.org \
    --cc=9fans@cse.psu.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).