WAAS Optimization of TSM Traffic

JHaynes4 · ‎02-27-2009

I have a question regarding WAAS optimization of TSM traffic. I know there is a policy in the WAAS that applies TFO+LZ+DRE to TSM traffic on ports 1500-1502. On our network there are nightly TSM backups that kickoff and according to the WAAS its reducing this TSM traffic by 90%. However, when I look at the Netflow statics off the routers at the remote site and the core site I see a far different picture.

At the remote site the inside ethernet interface sees approximately 21GB of backup traffic between the two hosts over a two hour timeframe. The serial interface of the core site sees approximately 23GB of backup traffic between the two hosts over the same timeframe and the remote WAAS tells me it reduced approximately 20GB of backup traffic to approximately 1.75GB.

Something is obviously not right here. Does any one else have any similar experiences or things they think I might look at to reconcile this?

Thanks...

dstolt · ‎03-06-2009

I'm wondering if you are doing redirect-in and redirect-out on your serial interface at the core?

Also, are you using negotiated-return method?

Thanks,

Dan

dstolt · ‎03-10-2009

I think you may be seeing DDTS: CSCsm35350 which can create a phantom packet count for netflow when using gre-egress (negotiated return). Your interface stats would reflect seeing the actual traffic and also the return traffic (uncompressed) from the WAE, which would inflate your interface statistics.

Does that seem like what you are seeing?

Dan

JHaynes4 · ‎03-10-2009

Well, that would make alot of sense. We are using GRE Return which has fixed alot of issues. I am going to look into this further thanks for the heads up.

kelrud · ‎03-06-2009

To transfer 21GB in 2 hours you'd probably need about a 25Mbps circuit b/w the sites.

Do you have that?

JHaynes4 · ‎03-10-2009

Yes, we do. It's T3's between sites.