RE: [nsp] Latency on the directly connected

From: Marcio Pilotto (marcio.pilotto@intelig.net.br)
Date: Thu Mar 28 2002 - 07:02:29 EST


Try to get a better sample. From a UNIX machine do a sequencial ping to IP
on the GEIP with a one second between each ping, may be you are going to see
that each minute you have a very big response time. For the rest of the time
it´s fine.

Marcio Pilotto

-----Original Message-----
From: Jim Warner [mailto:warner@cats.UCSC.EDU]
Sent: quarta-feira, 27 de março de 2002 19:29
To: cisco-nsp@puck.nether.net; kronay@saffron.org
Subject: Re: [nsp] Latency on the directly connected

Kathleen Ronay asked:

> Has anyone seen latency like this on a directly connected link? When I
> source the directly connected interface, I see no packet loss. If the
> testing is just from the router, I see packet loss.

We have seen latency without loss on packets to a 7500 GEIP+ interface.
Our interpretation is that IOS did not properly initialize the interface
to generate received packet interrupts. Packets landed in the buffers
and were cleaned out on the next clock interrupt, producing an appearant
distribution of arrival times. Upgrading IOS fixed this.



This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:13:39 EDT