Please send any corrections.
Routing Area WG (rtgwg)
73rd IETF - Minneapolis, MN
TUESDAY, November 18, 2008 1710-1810 Afternoon Session III
CHAIR(s): Alex Zinin <alex.zinin at alcatel-lucent.com> (not present)
John Scudder <jgs at juniper.net>
Alia Atlas <akatlas at gmail.com>
Scribe: Bob Salmi <bsalmi at routingdynamics.com>
existing work item update
a) expect a last call on the 2 framework drafts after ietf.
ordered fib draft progressing
revived is-is extensions draft
no ospf extensions draft need to investigate this.
c) longer term
d) drafts which are on hold pending demand
What things need a home now
dmitry: propose what about multicast frr should this be part of
charter update ?
jgs: Mcast frr would be reasonable to add
dward : 1) igp scaling architectural issues
2) composite transport groups
(Not clear if there are deliverables now/yet)
Danny: transport layer protecction
align with work going on in other groups
Stuart Bryant: you mean layer 4 right
Dward: some of this may happen in opsec
Loop Free IP fast reroute using Local and remote LFAP's presentation
Extensions needed to local lfap
X-hop neighborhood parameter
route tables of nodes within X-hop neighborhood are locally
using lsdb and calculation of SPT without exchanging information
no impact on convergence
failure notification mechanism
How it works
1) receive lsdb from ospf
2) calculate LFAP's
3) verify interfaces via fea
4) on failure send notification via fea & istall lfap's to rib
5) on receipt of fail notification install lFAP's to rib
have implementation using "WISER" emulation tool
Fedora core 8 and XORP
Convergence results in slides
loop free discussion
Mike Shand: Are you saying that for unicast you won;t get micro loops ?
Mike You may get remote loops outside the diameter of
the repair area
Alia: Inconsistencies in draft regarding how LFA's are computed
some examples in draft would cause forwarding loops
Alia: Why is the notification not applicable to the IGP.
I.E. Why not just tune the IGP instead of shorting it.
everything is pre computed so you can just trigger the install
George Swallow: do you need to precompute all the failures for your
neighbors links as well
George: so that lots of state I have to maintain right. I have to
strategy for all failures
Stewart in a realistic topology is this order k neighbors to the power
x hops the number for the number of strategies we need to
Stewart: what about competing solutions
also should look at some of the work with frr tunnels
what is wrong with an encapsulation based solution
disadvantage is overhead of additional header and processing
jgs: perhaps we should take this discussion offline.
jgs: 2nd or 3rd time this draft has been presented what do you want
to do with it
jgs: based on room poll we will pass on this work
rtgwg mailing list