1) if i read this right, it is a way for one lsr to ask another
lsr to test the first lsr's data path -- "can someone tell
me if i am working ok?". _if_ my interpretation is right
(and it is monday morning and not all the little gray cells
are at 100% efficiency yet), it seems that this is a fundamentally
stupid thing that the ietf should be standardizing, etc.
Not at all.
The point of the draft is to allow the control plane of a node to test
if it's dataplane is insync with the control plane. With large
distributed systems and hot/cool-standby control processors this is
becoming an issue that the SP community is concerned about.
The node under test needs a small amount of help from it's neighbors
just to:
upstream neighbor - reflect a packet back towards the node under test.
downstream neighbor - tell the node under test where the packet's ttl
expired and what the MPLS header looked like at that time.
2) regardless of point 1, it looks like there are various bits of
processing that have to happen here; but there is no talk
about metering/prioritizing/etcing this traffic, leaving open
a potential dos avenue. the security considerations section
seems to hint at this (recommending that loopback labels
be shared among trusted neighbors only).
The major point here was to require absolute minimal processing on the
part of the neighbors so that the node under test can be self regulating
in the load it has to bear.
This draft extends RFC 4379 which has text on metering. Didn't think it
was necessary to repeat it here.
aside, this last comment _also_ opens a huge can of worms;
who is trusted? how do i know? what filtering/etcing is needed
to make the right things happen?
There's a separate security thread. So I'll address this comment there.
...George
========================================================================
George Swallow Cisco Systems (978) 936-1398
1414 Massachusetts Avenue
Boxborough, MA 01719
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf