Re: [bmwg] draft-ietf-bmwg-acc-bench-meth-03.txt

["The Poretsky Family" <s.poretsky at verizon.net>]

Jay,

Item 1 makes sense.  Consider it done.  As far as item 2, at the wise
suggestion of the BMWG, co-chairs we have made the Methodology document a
"Genral Methodology" with the generic procedure for performing accelerated
stress testing.  The strategy is to create separate Methodology documents
for specific applications such as EBGP Peering, Operational Security, etc.
These already exist and have been posted.  LDP VPNs over MPLS-TE could make
a very useful Methodology document for this work item.

Scott

----- Original Message -----
From: <jakarthi at cisco.com>
To: <sporetsky at reefpoint.com>; <shankar.rao at qwest.com>
Cc: <bmwg at ietf.org>
Sent: Sunday, September 25, 2005 7:13 PM
Subject: [bmwg] draft-ietf-bmwg-acc-bench-meth-03.txt


Hi Scott/Shankar,

Excellent work ! I just had a few comments on
draft-ietf-bmwg-acc-bench-meth-03.txt.

1. I see that in section 3.3.1 you describe
   Number of Ingress Tunnels Tunnels
   Number of Mid-Point Tunnels Tunnels
   Number of Egress Tunnels Tunnels

I think it is more appropriate to call them "Number of tunnels as
Ingress, Number of Tunnels as Mid-Point,  Number of Tunnels as Egress
etc".

2. I would like to see a test case added to have a grid (say 20 x 20) of
simulated nodes (with TE enabled on the virtual nodes) advertised from
the tester, since this will REALLY add stress to the device under test
in calculating the constrained shortest path first (CSPF) from the
experience I have had with stress testing. There may be N*(N+1)/2
different paths to reach a point A from B. This is highly CPU intensive
and adds to the overall stress of the device under test.

Thanks,
Jay





_______________________________________________
bmwg mailing list
bmwg at ietf.org
https://www1.ietf.org/mailman/listinfo/bmwg


_______________________________________________
bmwg mailing list
bmwg at ietf.org
https://www1.ietf.org/mailman/listinfo/bmwg