7

I am looking for technical clarity on why someone should / shouldn’t use ISIS p2p topologies on Ethernet media.

Ethernet is a broadcast medium , however if it’s configured as a p2p link with /31’s the broadcast domain is minimised , therefore as I understand it the advantages are the reduction in the amount of information routing protocols must carry and manage and DR/DIS election can be omitted. ( ref here draft-ietf-isis-igp-p2p-over-lan

I am able to find little evidence to support this theory in practice therefore I am looking for sound technical reasoning to justify the above and ideally some information on how this affects convergence.

DrBru
  • 1,609
  • 3
  • 16
  • 32
  • Did any answer help you? if so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you could post and accept your own answer. – Ron Maupin Jan 04 '21 at 00:55

2 Answers2

5

I configure all my core OSPF interfaces as p2p. I would do the same for integrated IS-IS. This stops the DR/DIS election process from happening which can take a fair amount of time in a re-convergence scenario.

You do not HAVE to have them running as /31. It could be any mask as long as you only have two active speakers.

An added advantage on the OSPF side is that you get no type2 generated for the link.

From Cisco's site as well in regards to integrated IS-IS (http://www.cisco.com/en/US/docs/ios/12_2s/feature/guide/fissp2p.html#wp1037322)

Benefits of Integrated IS-IS Point-to-Point Adjacency over Broadcast Media

Configuring an IS-IS point-to-point adjacency over broadcast media provides performance improvements to the network convergence times of the customer's network because the feature saves the system from electing a designated router (DR), prevents flooding from using CSNPs for database synchronization, and simplifies shortest path first (SPF) computations.

mellowd
  • 3,824
  • 19
  • 24
3

I think best technical rationale for P2P is that you get reliable LSP delivery, in LAN it is unreliable thus it's done periodically.

Also pseudonode is additional resource use and during DR failure convergence is hurt.

ytti
  • 9,776
  • 42
  • 53