ATM Switch Interoperability

Webb, Tremayne TWebb at
Sat Jan 16 04:17:22 UTC 1999

Remember that the LS1010s do not shape the traffic and try to confirm
that the other end is not the cause of the cell loss.

-----Original Message-----
From: Martin, Christian 
Sent: Friday, January 15, 1999 8:03 PM
To: 'nanog at'
Subject: ATM Switch Interoperability


Does anyone have an experience dealing with cell loss between LS1010s
and Cascade/Ascend CBX 500s using VBRnrt?  It appears that either the
LS1010 or the CBX is performing UPC operations differently, as we are
seeing between .1 and 2% packet loss at the AAL, and .1 to .5% at the
cell level.  By allowing flexibility in the policing, or by removing
policing completely on the Cascade boxes, cell loss disappears.  It
appears that the LS may be sending cells too close together, but we are
unsure and cannot truly test thi without an Analyzer, of which no one
here has any experience using (even if we had one!).  The parameters are
as follows:

7513 w/AIP
PCR 30000
SCR 15000
mbs 3 (*32) = 96

CDVT 600
mbs 100 cell times
PCR 30000 CLP0+1
SCR 15000 CLP0

Cascade (As far as I know)
mbs 100
PCR 70754 cells/sec
SCR 35377 cells/sec
CDVT 600

I apologize for the vendor references, and I hope some don't view this
as off-topic!


More information about the NANOG mailing list