home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.dcom.cell-relay
- Path: sparky!uunet!sun-barr!ames!sgi!rigden.wpd.sgi.com!rpw3
- From: rpw3@rigden.wpd.sgi.com (Rob Warnock)
- Subject: Re: Computers dont like ATM?
- Message-ID: <sjdveq8@sgi.sgi.com>
- Sender: rpw3@rigden.wpd.sgi.com
- Organization: Silicon Graphics, Inc. Mountain View, CA
- Date: Fri, 20 Nov 1992 09:01:16 GMT
- Lines: 63
-
- finn@dalek.isi.edu (Greg Finn) writes:
- +---------------
- | ATM is politically very popular. Arguments that ATM brings
- | about end-to-end compatibility do not impress me. Most LAN traffic
- | stays within the LAN. Forcing LANs to comply with WAN data-link
- | standards is putting the cart before the horse.
- +---------------
-
- Especially when end-to-end "ATM compatibility" can be achieved *without*
- ATM-sized cells in the LAN! I have been harping on an idea I call "fat cells",
- which is basically that you switch AAL_PDU-sized packets in the LAN, with
- exactly one ATM header on the front of the AAL PDU. When the packet is to be
- routed out of the LAN via "real ATM", as it chops it up into real ATM-sized
- 53-byte cells, the gateway switch can replicate the prototype header onto the
- front of each cell. (For AAL5, it must also set the "end-of-frame" indication
- in the header of the last cell in the PDU.) Naturally, it is also the respon-
- sibility of the gateway switch to reassemble incoming cells into PDUs (leaving
- a prototype ATM header on the front), and switch these "fat cells" (complete
- AAL PDUs) into the LAN.
-
- Think of it as transparent bridging between "cell relay" and "ATM-formatted
- frame relay".
-
- Since AAL1 & AAL2 traffic don't have natural "PDU" boundaries, you could
- set an "aggregation ratio" at call-setup time, that says how many cells
- should be included in each "fat cell" on the LAN.
-
- The fun thing about "fat cells" is that a native-mode ATM terminal at the
- far end of the circuit cannot tell that you are sending/receiving "fat cells"
- instead of 53-byte cells!
-
- Another fun thing is that we could easily "tunnel" or encapsulate fat cells
- inside other protocols such as IP [the *inverse* of using ATM to encapsulate
- packets], to provide access to "native ATM" for many, many hosts without ATM
- interfaces! So much for "end-to-end ATM compatibility" requiring hardware
- changes... ;-}
-
- +---------------
- | I have great reservations about what ATM costs us in LANs.
- +---------------
-
- Likewise, but more for the dampening effect it's already having on alternative
- technologies (such as PTM, or gigabit "frame relay", or "fat cells", etc).
-
- +---------------
- | When ATM-based LAN performance exceeds PTM-based LAN performance/cost,
- | ATM proponents will have a much better argument than they now have.
- | Frankly, I don't forsee that happening in the electrical domain.
- +---------------
-
- Well, I do, actually. Sliding down the price/volume curve, and all that...
- Two-chip solutions that go from AAL PDUs to ATM cells to SONET line coding
- (and back) are not that far away, at least at 155 Mb/s.
-
-
- -Rob
-
- -----
- Rob Warnock, MS-9U/510 rpw3@sgi.com
- Silicon Graphics, Inc. (415)390-1673
- 2011 N. Shoreline Blvd.
- Mountain View, CA 94043
-
-