Path: utzoo!attcan!uunet!husc6!bloom-beacon!tut.cis.ohio-state.edu!mailrus!iuvax!pur-ee!j.cc.purdue.edu!i.cc.purdue.edu!h.cc.purdue.edu!s.cc.purdue.edu!rsk
From: rsk@s.cc.purdue.edu (Rich Kulawiec)
Newsgroups: news.admin
Subject: Re: A counter-example for those who would eliminate PC binaries
Summary: Whether it's more efficient depends on how many copies are sent.
Message-ID: <3335@s.cc.purdue.edu>
Date: 2 Jul 88 13:59:22 GMT
References: <264@octopus.UUCP> <3302@s.cc.purdue.edu> <1988Jul1.043049.2418@ziebmef.uucp>
Reply-To: rsk@s.cc.purdue.edu.UUCP (Rich Kulawiec)
Organization: Purdue University Computing Center Unix Systems Staff
Lines: 34

In article <1988Jul1.043049.2418@ziebmef.uucp> becker@ziebmef.UUCP (Bruce Becker) writes:
>	it seems like the time to look at the practice of other systems -
>	in specific, I am familiar with BitNet, which send out descriptions
>	of available binaries (source, documents, etc), and issues a pointer
>	to a thing called a "listserver"...
>
>	It seems to me that this uses far less net bandwidth than the
>	broadcasting method, and serves the community equally well...

This is a good way to distribute sources, and many Usenet sites have
reply-by-mail archive servers.  However, it is not always more efficient
in terms overall network bandwidth or individual site transmission costs.

Consider the simplest possible network: two sites, A and B.  (This is
useful because in this case "overall network transmission cost" == "per-site
transmission cost".)  If someone one A posts a source of size X, then the
total xmit cost is "X", and the per-site cost is "X" for A and 0 for B.
Now assume that just a pointer is posted, and ten people on B request
it: the overall xmit cost is "10X", and the per-site cost is "10X" for
A and 0 for B.

An extreme example, but more complicated topologies and cost distributions
eventually lead one to the same conclusion: if enough people at enough
sites want the source, it's cheaper to post it.  Some years ago, I think
Chuq did some analysis of this problem, and concluded that the tradeoff
was somewhere around 100 people, in terms of overall network bandwidth.
Clearly, however, this is a huge lose for the originating site, which
must send 100 copies of something rather than 1.

I am not trying to disparage this distribution method; I simply point
out that it is not a total solution to the problem, and brings problems
of its own.

Rich