[ns] Re: Single Polic{y,er} Entry for all BE sources?

[Vinay A. Mahadik]

Replying to my own question, sorry. Source/Destination ids of "-1" worked
for me to aggregate flows; got this off the list. (It was somewhat
difficult to query for this specific question on Google)

Thanks,
Vinay.

--
Vinay A. Mahadik
http://hickory.csc.ncsu.edu

>
> Hi,
>
> I have just begun using NS2. I want to study the behavior of a specific
> AF/EF flow in a DiffServ environment under a selfsimilar background BE
> traffic.
>
> My approach for generating a selfsimilar traffic is to use about 40
> UDP Pareto on/off sources generating together a BE traffic that is fed to
> the ingress node of the DiffServ cloud. My immediate problem is that NS
> asks me to provide an "addPolicyEntry" and an "addPolicerEntry" for every
> source-destination pair I have for the BE traffic. Then it runs out of
> bounds (with errors like - ERROR: Policy Table size limit exceeded.) that
> can be corrected only by reducing the number of pareto sources.
>
> Is there a better way to do this? (Add a common Polic-er/y Entry for the
> entire BE aggregate traffic entering the ingress DS node?)
>
> Also, how many Pareto sources could be considered 'sufficient'?
>
> Any other problems with my script(attached)? (I will be increasing
> testTime to over several hours eventually hence I figure start/stop times
> needn't be randomized?)
>
> Thanks,
> Vinay.
>
> --
> Vinay A. Mahadik
> http://hickory.csc.ncsu.edu