Stuart Yes, the network issues have to be addressed as well.
I don't believe that we are going to say that there is a requirement for IID or perfectly symmetric in the mean, or stationary or white or Gaussian or anything like that, as that would rule out use on any real network. What we need to understand is what the influence of lack of nice behavior on the results we need to get. For example, unknown asymmetry in the network obviously gives a strong limit on the accuracy of time, while nonstationarity and low-pass behavior of the PDV give strong limits on the frequency recovery. Right now they are dealing with PDV issues at the SG15Q13 meeting. Y(J)S -----Original Message----- From: STUART VENTERS [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 23, 2008 5:36 PM To: Yaakov Stein Cc: [email protected] Subject: [TICTOC] Requirements draft - iid or not Yaakov, I suspect another issue in the requirements draft is the description of the packet network that the thing is supposed to run over. In particular, we need to deal with the requirement that the probability density functions of the inbound and outbound packet paths be iid (independent and identically distributed processes). (See Dave's book page 181.) Maybe we could say that the paths are iid when they have no other loads, but it seems likely that the paths will have different distributions once they are loaded. (Or perhaps some applications could guarantee equal loading on both paths.) Perhaps someone on the list has some thoughts in this area. Regards, Stuart -----Original Message----- ---------------------------------------------------------------------- Message: 1 Date: Thu, 18 Sep 2008 08:20:38 +0300 From: Yaakov Stein <[EMAIL PROTECTED]> Subject: [TICTOC] Requirements draft - call for assistance To: "[email protected]" <[email protected]> Message-ID: <[EMAIL PROTECTED]> Content-Type: text/plain; charset="us-ascii" Hi all, Yesterday the chairs and Silvana, who is co-editor of the requirements draft, discussed the progress of this draft. The requirements draft tries to cover quite a number of areas for which timing is relevant, and while there has been significant participation from the cellular and TDM PW communities, there has been inadequate input on most of the other areas. In particular, for the areas * industrial applications * test and measurement * electric power and teleprotection we have no data other than that "pulled out of the air" at the Paris meeting, and no one has expressed interest in providing insights into their requirements. On the one hand TICTOC needs to consider all of the applications that will need to distribute timing, but on the other hand if groups with interest in particular applications do not join in the effort, we will not be able to produce a meaningful document. Please consider this email as a call for participation for the aforementioned areas. If you are working or interested in them, please send an email to the list. If you know people who are, please forward this email to them. If we do not receive input that we believe reliably reflects the requirements of these areas, we will have to drop them from the document, and thus remove them from consideration while working on solutions. Y(J)S _______________________________________________ TICTOC mailing list [email protected] https://www.ietf.org/mailman/listinfo/tictoc
