I have some clarifications to seek regarding RIP-2 opeartional modes.
I come tounderstand that generally most platforms support two different
modes of operation of RIP-2, aggregated and non-aggregated mode. In
aggregated mode a boundary RIP-2 router behaves more as a boundary RIP-1
router, wrt outside networks, filtering non-natural network adverti*ts
across the network boundaries. While in RIP-2 non-aggregated mode it
makes no special distinction between the natural network boundaries.
I suppose, RIP-2 aggregated is mainly to prevent containment of routing
table explosion. ( Also RIP-2 RFC doesn't any special reference to these
different modes and I presume this is
more of an implementation feature than a standard. )
Now here are my questions:
1) Is it realistic that an *adjacent* RIP-2 routers ( assume that both
are boundary routers with a common net ) be configured such that one is
acting in an aggregated mode and the other in a non-aggregated mode.
( This "mixed mode" doesn't seem "logical" )
2) Will the following ( implementation ) policy decision on RIP-2
receive side have an "adverse impact": If I am configured for
RIP-2 aggregated mode and if I am receiving adverti*ts for a
non-natural network on a network different from the advertised network,
don't create a routing entry. ( ie, if an adverti*t for
10.1.1.1/ff.ff.0.0 is received on 184.108.40.206/ff.ff.0.0 ignore the the
10.1.0.0/ff.ff.0.0 adverti*t but if I receive 220.127.116.11/ff.0.0.0
on the same interface accept it. )
3) Also on the RIP-2 send side would it be required to give a
"gratuitous advertsiement" for a natural network. ie., if I have an
interface with 10.1.1.1/ff.ff.0.0 and 18.104.22.168/ff.ff.0.0 and I am
running in *non-aggregated* mode should I advertise 10.0.0.0/ff.0.0.0
through the 22.214.171.124/ff.ff.0.0 inteface.
My understanding is that, questions 2 and 3 arise only if "mixed mode"
in question 1 makes sense !
Will be happy to hear your insights on this !