Hi all, I had reviewed the draft already, but I have no knowledge to comment on the metric model. As far as security issue is concerned, since this is a active measurement, it suffers the drawbacks of injection of measurement traffic which is listed well in section 9. They should see if their model complies with RFC 1262: guideline for Internet measurement activities. It can also help if they can mention what is the threshold percentage of measurement traffic that is tolerable and if they can quote an approximate percentage of measurement traffic needed to inject according to their loss metric and if it is acceptable.   Tina