Just to summarize this discussion:

(1) RTR1 originates the AS-external-LSA, which is flooded
throughout all areas (except stubs and NSSAs).

Then if 1583-backward-compatibity is turned on:

(2) RTR4 prefers the shortest path to RTR1, which
is through Area0
(3) RTR4 advertises RTR1 in ASBR-summary-LSAs to
Area2 and Area1
(4) As a result, packets from RTR5 to 48.0.1.0/24
go RTR4-RTR1

If instead 1583-backward-compatibity is turned off:

(2) RTR4 prefers the longer path to RTR1 through the non-zero
area Area 2
(3) RTR4 advertises RTR1 in costlier ASBR-summary-LSAs to
Area0 and Area1
(4) As a result, packets from RTR5 to 48.0.1.0/24
go RTR2-RTR1

John

> > > > > This is the scenario.
> > > > >
> > > > > 48.0.1.0/24 (External Route)
> > > > > |
> > > > > |
> > > > > RTR1------Area 0 ---RTR2---Area1
> > > > > | \ ---- |
> > > > > | \ |
> > > > > Area 2 Area0 RTR5
> > > > > | \- |
> > > > > | \ |
> > > > > RTR 3--Area2----Rtr4---Area1-|
> > > > >
> > > > > >From the perspective of RTR5
> > > > > the cost to 48.0.1.0 is 9 via RTR4-RTR1
> > > > > 10 via RTR2-RTR1
> > > > > 13 via RTR4-RTR3-RTR1
> > > > >
> > > > > Is it correct for RTR4 to advertise a metric
> > based
> > > > on
> > > > > a path via RTR3 because this subscribes to the
> > > > > non-backbone intra-area rule from its
> > perspective
> > > > or
> > > > > should it advertise the backbone path/metric
> > via
> > > > RTR1
> > > > > ?

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]