Re: draft on sampling techniques

[Nick Duffield <duffield@research.att.com>]

Cristian Estan wrote:
[...]
> In the trajectory
> sampling paper Nick and Matt say that the hash should be computed on the
> entire IPv4 packet except 6 bytes of the IP header which can change hop
> by hop: the first two which contain the type of service byte and from
> the 9th to the 12th which contain the TTL and the header checksum. 

Actually, Trajectory Sampling does not require that the hash be computed
over the entire IP packet!

In general one can say the more of the packet is used as input to the
hash, the better the statistical properties of the sampling method 
(i.e. sampling appears closer to independent)

In practice, we have found that statistical properties of trajectory 
sampling are good using only fields from the IP header and some 
from the TCP/UDP header.

Nick

--
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/>