[10429] in North American Network Operators' Group
Re: Another UUNET Explanation
daemon@ATHENA.MIT.EDU (Ehud Gavron)
Tue Jul 1 20:47:31 1997
Date: Tue, 01 Jul 1997 17:35:52 -0700 (MST)
From: Ehud Gavron <GAVRON@ACES.COM>
In-reply-to: "Your message dated Tue, 01 Jul 1997 16:34:02 -0700"
<199707012334.QAA23732@monk.via.net>
To: Joe McGuckin <joe@via.net>
Cc: nanog@merit.edu, GAVRON@ACES.COM
Frame-Relay, ATM, IP, X.25, DECnet, LAT, are all methods
of encapsulation of information into discrete units and
their subsequent delivery. Some are part of the PSTN,
and some are not.
Each protocol has its set of restrictions/features. None are
"bad" per se (not even X.25, Joel ;) Make sure not to confuse
the protocol with the medium it uses. For instance, frame-relay
DS3s will outperform ATM DS1s for data and voice routing :)*
Many people who "hate" frame-relay really despise the predominant
practice of overselling non Committed-Information-Rate circuits
into an overloaded central hub location.
Frame-Relay "routing" is L2, and IP routing is L3, so I won't
really go into comparing them any more than, say, discussing
the technical advantages of TCP vs. IP. They are two different things
with similar form yet different functions.
Ehud
* There are some purists who would claim that the additional overhead
of switching frame-relay packets would increase latency over that inherent
in an ATM switched network. They would say that my example above (which is
just an example, so reading too much into it is an exercise in being a pedant)
is flawed because the frame-DS3 is more latent than the ATM-DS1. Nonetheless
I claim the inherent latency of both networks to be less than that noticeable
by human ears. Nya. Nee. Nya. Nee. PooPoo.
>I've noticed that several of the larger networks use frame-relay.
>Why? Our experience with frame-relay with the local telco has had
>mixed results.
>What technical advantages does a frame-relay network have over an
>IP routed network?
>Thanks!
>Joe
