<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7652.24">
<TITLE>Per sesison QoS - Again</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<P><FONT SIZE=2>Have been happily using per-session QoS for a while now in 12.4M and am now forced to move to 12.2SR as part of a platform change.<BR>
<BR>
Doing so requires a new RADIUS AVP implementation<BR>
<BR>
(see<BR>
<A HREF="http://www.cisco.com/en/US/docs/ios/qos/configuration/guide/per_session_qos.html">http://www.cisco.com/en/US/docs/ios/qos/configuration/guide/per_session_qos.html</A><BR>
and<BR>
<A HREF="http://www.cisco.com/en/US/docs/ios/security/configuration/guide/sec_qos_aaa_policy.html">http://www.cisco.com/en/US/docs/ios/security/configuration/guide/sec_qos_aaa_policy.html</A><BR>
)<BR>
<BR>
Which now seems to be broken as the "shape average percent XX" now seems to set the CIR<BR>
from the PHYSICAL interface the L2TP packets arrive on (gigE) as opposed to the virtual-access<BR>
which the subscriber arrives on, which has its bandwidth set correctly by AVP (a "sh int" confirms<BR>
that the vaccess bandwidth is indeed set correctly)<BR>
<BR>
Disabling subinterfaces as a test does not fix this, s/w in question is 12.2(33)SRD, has anybody<BR>
come across this? bugtool yet again does not seem to yield anything useful.<BR>
<BR>
Thanks in advance<BR>
<BR>
Dave.<BR>
<BR>
<BR>
------------------------------------------------<BR>
David Freedman<BR>
Group Network Engineering<BR>
Claranet Limited<BR>
<A HREF="http://www.clara.net">http://www.clara.net</A><BR>
<BR>
</FONT>
</P>
</BODY>
</HTML>