[HZB2]First: This document does not require that all affected traffic be
directed to "one" neighbor node, and traffic is still balanced on all different
neighbor nodes.
[SLI3] Again, purely theoretical, it depends on the network topology, where is
the traffic coming from and number of
Thanks Gyan!
In the next revision, I'll add more information on the requirements and
comparisons with the other approaches.
Best,
Haoyu
From: Gyan Mishra
Sent: Sunday, January 30, 2022 9:41 AM
To: Haoyu Song
Cc: Greg Mirsky ; IETF IPPM WG ; Tianran
Zhou ; spring@ietf.org
Subject: Re: [ippm]
Tianran,
Your summary is basically correct.
But for the first point, I don't think using UDP is a new protocol. And
extending an existing protocol beyond its original scope in many cases is like
designing a new protocol. The pros and cons all need to be carefully
considered.
Best,
Haoyu
[HZB2]First: This document does not require that all affected traffic be
directed to "one" neighbor node, and traffic is still balanced on all
different neighbor nodes.
[SLI3] Again, purely theoretical, it depends on the network topology, where
is the traffic coming from and number of neighbor