[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: WG last call on Sampling and Filtering Techniques until March 12



Hi Tanja,
"recommended" is a rather vague term.
In my view, it should be made clear that in case someone implements the hash based filtering, he MUST (or SHOULD) implement the recommended hash function. Otherwise, the whole concept of trajectory sampling, which is one of the main reason for having hash based sampling, is broken.
From the discussion I saw on the ML I think a consensus on a MUST is probably difficult to reach, but at least a SHOULD should be in.
Maurizio


Tanja Zseby wrote:

Hi Juergen,

just a small comment: we agreed that it is mandatory to implement one of the described selection methods to be psamp-conformant. Regarding the hash-function we just give recommendations if someone decided to implement a hash-based selection. So to my understanding we will not have one hash-function mandatory but only recommended. Do you agree ?

Regards
Tanja

Juergen Quittek wrote:

Dear all,

Tanja has submitted a new version of our document on Sampling and Filtering Techniques for IP Packet Selection. Please find it at
<http://www.ietf.org/internet-drafts/draft-ietf-psamp-sample-tech-06.txt>.



The document in mature enough for entering WG last call. The call will end on March 12. This allows us bringing up and discussing comments and issues also at our session in Minneapolis.

Please read the document carefully and comment on it.

PLEASE ALSO COMMENT IF YOU THINK THE DOCUMENT IS FINE.
Positive feedback is as welcome as comment on technical flaws and editorial nits.


There is one open technical issue left that we should continue discussing and try to close during last call. The document describes the need to support hash functions for packet selection and for packet digest. We agreed already, that for each of them one hash function should be mandatory to implement. Currently it is IPSX for packet selection and CRC32 for packet digest.

However, a discussion recently started on this list considered replacing both (as mandatory to implement) by the BOB hash function.
We have not reached consensus on this issue.


Thanks,

Juergen





--
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/>