OSPF performance/SPF calculations

[Joakim Tjernlund]

Ondrej Zajicek <santiago at crfreenet.org> wrote on 2010/04/23 11:52:26:

> From: Ondrej Zajicek <santiago at crfreenet.org>
> To: Joakim Tjernlund <joakim.tjernlund at transmode.se>
> Cc: bird-users at trubka.network.cz
> Date: 2010/04/23 11:46
> Subject: Re: OSPF performance/SPF calculations
>
> On Fri, Apr 23, 2010 at 10:50:38AM +0200, Joakim Tjernlund wrote:
> > > Yes, i got the idea. Our algorithm (for PTP) is to search for a ptp
> > > iface with a full neighbor with given Router ID and choose the cheapest
> > > one. This would lead to the same results as your idea, but a slightly
> > > less efficient, but probably not important unless you have a hundreds
> > > of PTP ifaces on a router.
> >
> > I got a lot of PtP I/Fs(some 20-30) :)
> >
> > But this won't fix multiple ptp I/Fs between the same two routes and
> > I don't think it will work if one end is unnumbered and the other one is not.
>
> Our algorithm (in the git tree) works well for multiple ptp ifaces between
> the same two routers (because both the SPF and the calc_next_hop() chooses
> the same (cheapest) ptp link.

Even if all local PtoP I/Fs have the same IP address or no IP address?

What does this comment in calc_next_hop mean?
 /*
     * Remaining cases - local neighbours.
     * There are two problems with this code:
     * 1) we use IP address from HELLO packet
     *    and not the one from LSA (router or link).
     *    This may break NBMA networks
     * 2) we use find_neigh_noifa() and does not
     *    take into account associated iface.
     *    This breaks neighbors connected by more links.
     */

>
> Our algorithm works regardless of whether it is unnumbered or numbered, as
> we don't use the information from the 'Link Data' field of router-LSA for
> ptp networks.

Yes, the Link Data field is useless for unnumbered I/F as it contains the
local interface ifIndex and neither of Option 1 and Option 2 in
section 12.4.1.1.  Describing point-to-point interfaces