(I'm also directing this to sun-nets because the discussion is more
networking oriented than system management oriented.)
From: rlg@IDA.ORG (Randy garrett)
Date: 26 Feb 91 22:48:18 GMT
I also heard of two other serial line IP programs: ppp and cslip.
cslip is supposed to be compressed slip -- presumably it does some
type of run-time compression.
ftp.ee.lbl.gov:cslipbeta.tar.Z is RFC1055 SLIP plus RFC1144 TCP header
compression, which you *really, really* want for any link slower than,
say, 64Kb. You'll notice a remarkable performance improvement in most
applications.
ppp is supposed to be a newer program altogether. Most respondents
felt these two would be superior to slip although there was some
sharp dissenting opinions on that.
If you absolutely must use SLIP (because, perhaps, it's already
running in your environment and you must interoperate) then by all
means get header compression running on top. For a new installation,
you're much better off using PPP (RFC1171 and RFC1172). SLIP was a
first cut at a serial line IP protocol. PPP learns a lot from its
example and gets it right. Read the DEFICIENCIES section of RFC1055
and the Introduction section of RFC1172 for a comparison.
In any case, I have not tried either of these two, although I plan
to do so. I recently acquired ppp from: omnigate.clarkson.edu.
That PPP was developed by Drew Perkins <ddp@andrew.cmu.edu> on a VAX
under 4.3, apparently as a design proof while writing the original PPP
RFC. It was then ported by Brad Clements <bkc@omnigate.clarkson.edu>
to a Sun386i under SunOS 4.0.1, adding STREAMS support and TCP header
compression along the way. Karl Fox <karl@morningstar.com> took that
and ported it to SPARC and SunOS 4.0.3, 4.1, and 4.1.1. If you're
planning to run PPP under SunOS 4.1.* then you'll probably find it
much easier to start with the (so far) final result, which you'll find
in tut.cis.ohio-state.edu:pub/ppp/ppp-sparc4.1.tar.Z. If you find
bugs or (heaven forbid!) SPARC dependencies when you try to install it
on your Sun-3s, please be kind enough to feed any fixes back to Karl
or me for re-integration.
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:11 CDT