Re: [mpls] Review of draft-ietf-mpls-residence-time-12

[Stewart Bryant <stewart.bryant@xxxxxxxxx>]

Sorry Acee, I accidentally deleted your message about paths that support 
this.

Yes, to get the best quality time transfer you would pick paths such 
that all hops had the property. If you did not do that you can undo all 
the good work you did on the other hops by one that introduces variable 
delay. Indeed if the variability has certain systematic characteristics 
it can be very hard to use other techniques to minimise its impact on 
the quality of the time transfer. An example of a characteristic that is 
very difficult to address by other means is a very slowly increasing delay.

Stewart