Re: need of sync of clocks in dup metric def

STEPHAN Emile RD-CORE-LAN wrote:
> Hi Henk,
> 
>>> The methodology (section 2.6) of the draft
refers to section 2.6 of
>>> RFC2680 which speaks about sync of clocks:
>>>
>>> Is the sync of clocks really needed for
detecting packet
> duplication?
>> No, actually this is one of the things that
suprised me as well in
>> RFC2680.  Does anybody remember why this is in
2680?
> 
> My understanding of the philosophy of RFCs 2679 &
80 is that a packet
> arriving very very late is an excellent packet lost.
Seriously, the
> definition of the packet loss requires the time the
packet is sent
> because there is not any local (on the receiver) event
available to
> start the timeout timer.
> 
> Packet duplication differs. The reference of time for
waiting a copy is
> the arrival of a first image of the packet. So it does
rely on any
> remote time.

A lost packet, should never be counted as a duplicate. These
metrics should not 
be looked at in isolation - they are related.

jeff

_______________________________________________
ippm mailing list
ippmietf.org 
https://w
ww1.ietf.org/mailman/listinfo/ippm