home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!pipex!unipalm!uknet!strath-cs!str-ccsun!ccsun.strath.ac.uk!craa85
- From: craa85@ccsun.strath.ac.uk ( D.W.Stevenson)
- Newsgroups: comp.dcom.cell-relay
- Subject: Re: Future of IP routers
- Message-ID: <1992Sep3.094148.21738@ccsun.strath.ac.uk>
- Date: 3 Sep 92 09:41:48 GMT
- References: <1992Aug24.184659.1879@gandalf.ca> <22264@venera.isi.edu> <JTW.92Aug26203449@pmws.lcs.mit.edu>
- Sender: news@ccsun.strath.ac.uk (News account )
- Organization: University of Strathclyde
- Lines: 23
- Nntp-Posting-Host: coll
-
- In article <JTW.92Aug26203449@pmws.lcs.mit.edu>, jtw@lcs.mit.edu (John Wroclawski) writes:
- |>
- |> This problem will be overcome when people start using ATM based
- |> backplanes internal to the router.
- |>
- |> I don't understand this point. You appear to be describing the
- |> difference between a time-division "bus" backplane and a parallel or
- |> "switch-fabric" backplane. ATM has nothing to do with this; either
- |> design can be used to switch either IP or ATM.
- |>
-
- You are right in saying ATM has nothing to do with this. I'm just pointing out
- the current limitations of high end commercial routers, which use time division
- backplanes whereas in the future, in order to get the increased backplane
- throughput required to match the increase in packet processing power of each
- interface, they will use switch fabric backplanes instead. ATM is the most likely
- candidate because ATM silicon will soon be available, in quantity and at low cost.
-
- --
- Dave Stevenson d.w.stevenson@strath.ac.uk
- Computer Centre Communications Tel : 44 41-552 4400 ext 3461
- University of Strathclyde
- Glasgow, Scotland, U.K.
-