home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.dcom.cell-relay:461 comp.dcom.lans.fddi:591
- Newsgroups: comp.dcom.cell-relay,comp.dcom.lans.fddi
- Path: sparky!uunet!sun-barr!ames!sgi!rhyolite!vjs
- From: vjs@rhyolite.wpd.sgi.com (Vernon Schryver)
- Subject: Re: >>>>Future of IP routers
- Message-ID: <pbgavto@rhyolite.wpd.sgi.com>
- Organization: Silicon Graphics, Inc. Mountain View, CA
- References: <Btx5JH.2t7@usenet.ucs.indiana.edu>
- Date: Wed, 2 Sep 1992 13:06:06 GMT
- Lines: 37
-
- In article <Btx5JH.2t7@usenet.ucs.indiana.edu>, KUMQUAT%SMCVAX.BITNET@mitvma.mit.edu (Gary C. Kessler, +1 802-879-3375) writes:
- > If a station on an FDDI ring releases the token because its own TRT has
- > expired, it does NOT necessarily follow that anyone else who follows that
- > station will also see a "late" token. Since we all have the same target
- > token rotation time (many feel that 8 ms. is an ideal value) and independent
- > TRTs, it really depends upon the amount of synchronous bandwidth that is
- > allocated and used. Thus, it is possible that I will see an "early" token,
- > use up my entire TRT time (i.e., the entire amount of the earlyness), pass
- > it on to you, where you will also see an "early" token.
- >
- > Yeah, this is hard to do in just prose...
-
-
- I don't see how any synchronous bandwidth allocation can make the
- token appear to arrive earlier with respect to the local TRT than
- otherwise.
-
- The Token Rotation Timer of all stations downstream of the station that
- is currently transmitting and about to have its own TRT expire are
- "earlier" and closer to expiring than the transmitting station. That
- is because the transmitting station saw the token most recently (it was
- the last station to see the token) and so reset its own TRT most
- recently. All other stations on the ring are certain to think the
- token is even more "late" than the transmitting station whose on TRT
- expired.
-
- Some of us do not think much of an 8 ms. For example, an 8 ms TTRT on
- a ring with 1 ms latency limits the maximum possible bandwidth between
- any pair of stations to about 77Mbit/sec. I think maximum bandwidth
- goes something like (TTRT-latency)/(TTRT+latency) with a correction for
- the frame being transmitted when TRT expires. A maximum sized ring
- with a 1.67 ms latency and 8 ms TTRT permits only 65 Mbit/sec between
- any pair of stations even when the rest of the ring is completely
- idle. That is just plain silly!
-
-
- Vernon Schryver, vjs@sgi.com
-