phenomenon interstice

**************************************************** Research Assistant Computer Networking Research Laboratory (CNRL) C203,Dept. of Electrical & Computer Eng. Colorado State University, Fort Collins, CO 80523 Voice: 970-491-7067/7974 Fax: 970-491-2249 http://www.cnrl.colostate.edu http://lamar.colostate.edu/~nischal ****************************************************

Hi JP,

Objective metrics include 2 main categories:

Those for = measuring the performance (the rapidity) of a PCE entity;

Those for = measuring the accuracy of the PCE computation:

Do the quality and the performance of the path computed by PCEs entities complies with = the query?

Regarding the second point, Lei and I merged 3 = I-Ds in one which defines "Spatial and Multicast Metrics for IPPM" = (http://www.ietf.org/internet-drafts/draft-step= han-ippm-multimetrics-00.txt ). It defines 2 sets of metrics, spatial and one-to-group: =

The intend of = the spatial metrics part is clearly to provide TE and interdomain with metrics to = decompose end-to-end performance;

The intend of = the spatial metrics part and one-to-group part are to provide metrics to decompose = the performance of point-to-multipoint services;

The draft was presented at Minneapolis during = the IPPM WG. I clearly designed it to become an IPPM WG item. I take a = special care to select metrics which conform to the IPPM framework (RFC2330) and = which are directly usable for unicast or multicast TE.

This draft will be discussed in the IPPM = mailing. Mid April, chairs will decide its status evolution.

I have no doubt that there are a lot of = missing points in the discussion section such as case study, applicability… So = please read the document and send your comment in PCE and IPPM = list.

Regards

Emile

De :<= /span> JP Vasseur [mailto:jvasseur@cisco.com]
Envoy=E9 : jeudi 13 = janvier 2005 20:09
=C0 : STEPHAN Emile = RD-CORE-LAN
Cc : pce@ietf.org
Objet : Re: [Pce] contribution to charter item

Salut Emile,

On Nov 17, 2004, at 8:00 AM, STEPHAN Emile RD-CORE-LAN wrote: =

Dear JP, Adrian and = all

There is an important item in the PCE WG charter, which is the definition of MIBs and management = procedures related to the new protocols, protocol = extensions and operational elements.

I am interested to contribute to this WG item. =

Great !

Considering your involvement in IPPM, you may also be interested = by:

May 05: Submit first draft of the definition of = objective metrics

Please let me know when you will start this activity. =

As indicated in the WG Milestones ... of course, for the MIB, = this requires some time until we make progress on the protocol(s). =

Cheers.

JP.

Regards

Emile

_______________________________________________ =

Pce mailing list

Pce@lists.ietf.org

https://www1.ietf.org/mailman/listinfo/pce

> Did lost packets or reordered packets (or both) cause buffer growth? > > A key point of standardizing a reordering metric is to distinguish > reordering from loss - they have different causes and cures. > > If I look at the RBD buffer occupation frequencies, > there will be many sizes encountered as the buffer is growing. > But only one of the sizes is relevant to buffer design in this context, > the size needed to restore a reordered packet to order. > One could look at a histogram of buffer sizes after a measurement, > but the RDB frequencies alone are not sufficient to say whether a > specific buffer size can mitigate reordering and to what extent > (as a fraction of reordered packets).