Re: PSAMP open issue 18: export rate limit?

[Andrew Johnson <andrjohn@cisco.com>]

Hello Benoit and all.

Benoit Claise wrote:
> Dear all,
>
> In [PSAMP-FMWK], the following statement is still un-addressed in 
> [PSAMP-PROTO]
> PROTO-18 “The exporting process must have an export rate limit, 
> configurable per Exporting Process”.
> See section 8.3 for the details 
> http://www.ietf.org/internet-drafts/draft-ietf-psamp-framework-10.txt
>
>
> First of all, I have a question regarding the requirement itself. Isn't 
> it supposed to mean: “The exporting process must have an export rate 
> limit, configurable per Collecting Process”? Otherwise, I don't 
> understand the exact requirement when we have a PSAMP device exporting 
> to multiple collectors!

I would say that each Exporting Process exports to only one collector,
but can be fed by multiple Metering Processes.  Also, each Metering
Process may feed Multiple Exporting Processes.  With this concept we
can rate limit the Exporting Process and it will only affect one
collector.


> Then, how to address the requirement?
> Shall we simply update [PSAMP-PROTO] with a sentence such as: "The 
> exporting process MUST have an export rate limit, configurable per 
> Collecting Process"?
> And that's it?

If we agree that one Exporting Process deals with only one collector
then this simplifies matters.


> Do we keep the export rate limit as a generic term? Or do we define 
> precisely what it mean? Example: number of IPFIX Messages, number of 
> bytes/s, number of records/s?

I think we should pick a MUST, such as we MUST be able to rate limit
based on the number of data and option records per second, but you MAY
rate limit on other criteria also, such as bandwidth.  This will allow
some common ground between implementations, useful for people who buy
from multiple vendors.

Regards

Andrew

--
to unsubscribe send a message to psamp-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/psamp/>