home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.dcom.cell-relay
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!ames!haven.umd.edu!decuac!pa.dec.com!nntpd2.cxo.dec.com!nntpd.lkg.dec.com!carafe.enet.dec.com!goldstein
- From: goldstein@carafe.enet.dec.com (Fred R. Goldstein)
- Subject: Re: clarifying the role of SSCOP?
- Message-ID: <1992Nov11.232255.24793@nntpd.lkg.dec.com>
- Sender: usenet@nntpd.lkg.dec.com (USENET News System)
- Organization: Digital Equipment Corp., Littleton MA USA
- Date: Wed, 11 Nov 1992 23:13:00 GMT
- Lines: 70
-
-
- In article <1992Nov11.185930.2664@sics.se>, craig@sics.se (Craig Partridge) writes...
- >Can we back up a little bit? There appears to be some contradictions in the
- >notes about SSCOP. Let me see if I can characterize my puzzlement:
-
- > Is SSCOP always to be run over AAL5, or is it really just for
- > applications that specially request it (like SS7, or telephone
- > calls over satellite links)?
-
- It's a protocol that the network won't look at except for signaling.
- You can use it if you want. You can ignore it if you want. It's free.
- You just pay for the implementation :-) .
-
- >It seems to me that different notes seem to suggest different answers.
-
- >From my point of view, as long as I can always ignore SSCOP (i.e. not
- >have it mucking up my cells for data communications), I'm happy. All these
- >arguments about what SSCOP will do for general data comm are, in my view,
- >specious. (End-to-end error recovery is better, regardless of speed and
- >error rates. One can tune the end-to-end protocols to dynamically adapt
- >to the environment presented).
-
- Well, you're free to ignore it. I personally inveighed for a while
- against ALL AALs, because I don't think frame recovery makes as much
- sense as cell recovery. Of course that was before AAL5. I could do
- cell recover in BLINKBLT in as little overhead as AAL3/4. AAL5 has
- lower overhead so it may be a better deal.
-
- The end-to-end argument is not _always_ valid. It reminds me of the
- original arguments against cell recovery. AT&T "proved" that it doesn't
- do much good across a WIDE range of loss rates, from around 10^-3 to
- 10^-8. Trouble is, the curve was crossing over at 10^-3 and cell
- recovery was much better at, say, 10^-2 loss. So how lossy is your
- network? How busy is it? Who's paying for the bandwidth? How good are
- your cell emission pacers? How big are the buffers? etc. etc. In
- other words, they proved that you don't need cell recovery with a
- moderately low loss rate.
-
- But I've used networks (TCP/IP) with a very, very high packet loss rate.
- Like 50% on a given hop. Now if you have three hops like that, then you
- will get about 12.5% of your packets across each time. End-to-end
- recovery is pretty bogus! But hop-by-hop recovery would have worked a
- lot better; each recovery would have occurred locally and with a 50%
- chance of getting through each time. Do the math yourself.
-
- Okay, you may ask, where does this idiotic network exist? Stop by
- rec.radio.amateur.packet and ask about some of the 1200 bps 2 meter
- half-duplex IP-over-AX.25 "LANs". They're basically Aloha, with some
- really bad periods of congestion caused by "hidden transmitter
- syndrome". I use this mess regularly. And of course I don't go more
- than two hops, because TCP just takes too long to recover. Note that we
- usually hold our TCP timers back to "linear" rather than "exponential",
- lest we wait weeks to retransmit.
-
- Will ATM be that bad? I hope not. I know how to keep it from being
- very lossy. But I'm not sure the CCITT does. And if it's cheaper in
- the long run to run the circuits at 95% utilization with say 1% loss
- (and selective retransmission) than to run them at 90% with 0.001% loss,
- then why not? And if you run it even lossier (hey, it may not be
- _planned_), then local retransmission begins to look good. Just like
- it does, to me at least, on 145.51. Where I end up sending my SMTP
- to an adjacent node to let _it_ pass it along, since end-to-end fails.
-
- I'm not saying everyone will need SSCOP. Just that it may be a nice
- tool when things go wrong, and not a very expensive one to keep around
- anyway.
- ---
- Fred R. Goldstein goldstein@carafe.tay2.dec.com
- k1io or goldstein@delni.enet.dec.com voice:+1 508 952 3274
- Standard Disclaimer: Opinions are mine alone; sharing requires permission.
-