Re: comments on draft-dcbench-def-00 (part 1)

On 7/31/13 1:26 AM, Lucien Avramov (lavramov) wrote:

> The goal is to first measure it correctly

Hi Lucien,

This implies there is One True Method for jitter measurement. I wish
that were true, but it's not.

As it is, we have at least three definitions: Two in RFC 3393 (and that
document also deprecates the term), and another in RFC 4689, which came
out of this working group. There are others too.

This reinforces Al's point about the merits of the bmwg approach of
terminology and then methodology documents.

If nothing else, I think you'd be doing testers and implementers of test
instruments a big service by defining exactly which method is in use.

The two bullet points in section 3.2 are a good start, but note that
there are different ways to measure "variation," each with its own pros
and cons. (For example, for some applications, in-sequence is critical,
for others it doesn't matter, and so on.)

Thanks!

dn