Cell Relay Archive

Cell Relay Retreat>List Archive>month:1996-Nov> msg00279



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

Re: What goes in the peak cell rate of a UBR connection?

  • From: fgoldstein@bbn.com (Fred R. Goldstein)
  • Date: Thu, 21 Nov 1996 13:39:31 EST

In article <56vvhm$k0g@bcarh8ab.bnr.ca> smiller@bnr.ca (Scott Miller) writes:

>If I'm setting up a best-effort connection, what should I use for the PCR?

I *hate* that term "best-effort".  It is taken in the ATM world to mean "worst 
effort".  That's what UBR means.

>I've been told that it should be set to the line rate, because otherwise the
>switch can use GCRA to tag based on the provided rate, and any cell other than
>the first in a burst will be tagged. (since it has no specified burst size)

>On the other hand, UNI 3.1, 3.6.2.4 says that "the user need not conform to the
>signalled PCR and the network may enforce a PCR other than the signalled PCR",
>which tends to indicate that the PCR is more or less meaningless to the switch. 

I'd go with that last clause.  In UBR, you get available, unreserved, 
low-priority bandwidth.  It's *unspecified* bit rate, so why worry about what 
rate to specify?  No promises, just spare capacity.

With VBR, this stuff really matters.
--
Fred R. Goldstein   k1io    fgoldstein@bbn.com
BBN Corp., Cambridge MA  USA         +1 617 873 3850
Opinions are mine alone; sharing requires permission.