On 7/19/2010 6:16 AM, Stewart Bryant wrote: > I have always been assuming that we would do this over a timing FEC - > not reserved labels - not special types - ordinary MPLS with a timing > semantics attached to the label(s). > > What I would also like to see is an encapsulation that works for both > NTP and 1588 - i.e. put the corrections in a well known place at the > front of the packet (we have to pick a single TS format), followed by > the timing payload (NTP, 1588 or foo), then fix it all up at the edge of > the network, either via a slave timeserver, of via correcting the packets.
I'm not clear here whether you are suggesting diddling with the contents of the packet or this is some sort of add-on. If you modify the packet then all attempts to add any authentication is off the table. If you are changing some other part, any changes cannot be authenticated so you render the protocol subject to interference. Of course I could be totally misunderstanding this as I'm not really knowledgable about MPLS except for the basics. Danny _______________________________________________ TICTOC mailing list [email protected] https://www.ietf.org/mailman/listinfo/tictoc
