Cell Relay Archive

[jabley@globalbrains.com (Joe Abley)]

On Thu, 07 Oct 1999 05:18:11 GMT, nxiv@my-deja.com <nxiv@my-deja.com>
 wrote in comp.dcom.sys.cisco:
>I have been posed with an interesting management issue in our ATM over
>frame relay WAN.

Do you mean frame-relay over ATM?

>We get idiosyncratic response times to extended pings
>from our main site to remote sites.

What do you mean by "extended pings"?

>I believe that the issue is that
>the ATM/frame relay switch is buffering the traffic when it is
>congested (i.e. store and forward vs. cut through switching) which is
>causing ICMP timeouts at random times.

What switches are you using?

>Our main site has a pair of Cisco 7513s both with a SONET/ATM/OC-3
>interface.  The remote sites have predominately Cisco 2503s with an
>Adtran TSU provided by Sprint.  There are no BECNs and FECNs on the
>frame relay side of the cloud.  The CIR varies from 256Kbps to 1024Kbps.

You're sending data from the 7513 towards the 2500s? How do you
have your ingress queues configured? If this is a StrataCom network,
are you using ForeSight?

>I want to know if there's a way to measure via SNMP any congestion
>issues with the SONET interface on the Cisco 7513.  Our primary network
>management application is Spectrum 5.0rev1 with the Frame Relay module.

What does "show int atmX/Y/Z" show on your 7513?

>Is it possible to measure any congestion issues on the ATM side via
>SNMP?  If so, what are the RFCs?  Are there any white papers or websites
>on how to troubleshoot this issue via SNMP?  What I am looking for is
>some diagnostic methodology to be able to tell Sprint that their switch
>is misconfigured for our data applications.

You're not running switches? I doubt Sprint are going to be happy
to let you poll _their_ switches :)


Joe