Re: [Re: [Re: [Re: [Re: [Re: [OSPF doubt [9:2754] posted 09/03/2003
- Subject: Re: [Re: [Re: [Re: [Re: [Re: [OSPF doubt [9:2754]
- From: "Curtis Call" <curtiscall@xxxxxxx>
- Date: Wed, 3 Sep 2003 14:18:09 GMT
"Alok Dube" wrote:
> thanks for the reply,
> What i am trying to actually do is do a bit of a "diff" in the logic
> between OSPF and ISIS.
> In the case of ISIS we have a orthogonal address space to check for loops
> etc (router-id need not be an IP etc).
> In OSPF the router-id is an IP, which should ofcourse logically be
> orthogonal to the address space. The router-id is only to identify the
> router, the source of the LSA and basically the way i see it work is:
> prefix x.y.z.w comes from router routerid1.
> so incase we get the prefix xyzw twice from different interfaces etc or
> there is a link flap etc, we know that routerid1 is associated with the
> prefix and hence avoid poision reverse/split horizon and those scenarios.
> However, i agree the router-id not being = IP address is an grey area, it
> still dont quite comprehend the "stub".
> fine if its numbered and P2P we need to send around the network, but if
> its unnumbered and P2P,....the whole purpose of unnumbered seems beaten (
> i.e we need switch from interface A to interface B to go to end point X,
> if the link has an IP, it would be a part of reachibility, if not why the
> stub...even in the "all P2P unnumbered cases").
> its more academic curiosity than rebuilding the protocol :-)
> -thanks again,
It might be worthwhile to take a look at OSPFv3 in order to remove your
confusion regarding the RouterID. It is geared around routing IPv6 but still
uses the same ID, that might make more sense for you.
Regarding unnumbered interfaces, the advantage they provide is the
conservation of address space, having the peer advertise the address it's
neighbor is sourcing their packets from does nothing to hurt this goal, it
just provides reachability information to other routers in the network.
Message Posted at:
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/juniper.html