[c-nsp] Conditional advertise-map

Gert Doering gert at greenie.muc.de
Thu Sep 16 17:10:52 EDT 2010


Hi,

On Wed, Sep 15, 2010 at 11:05:00PM +0100, Heath Jones wrote:
> You've got me thinking about this whole non-determinism thing and something
> just entered my brain.. The debates here and on other threads seem to be
> based on the notion that aggregating prefixes in tcam becomes
> non-deterministic due to not being able to guarantee future # of tcam
> prefixes if the topology changes.

Yep.  "Local indeterministic-ness" - you have a number of routes that
you accept, but sometimes "300.000" routes will fit into 256k TCAM
slots, and some other day, "300.000" *other* routes will overflow the
TCAM.

Or, worse, something in your IGP changes, changing BGP nexthop routing,
and all of a sudden, one completely unrelated router explodes because
it cannot aggregate his prefixes anymore (different next-hops that
took the same egress interface before and now go different ways).


> Does that mean that people think routing is completely deterministic now,
> and that we can guarantee noone will go and advertise a shitload of /24's?
> Or, does it mean that people think its more likely for a topology change to
> affect the aggregation in a tcam because the egress interface will change?

Well, today I can look at my BGP sessions and say "I am receiving 
294.712 prefixes, and so I need a TCAM that is at least that big".

Now, of course, someone could go out and add 256 /24s, and then I need
a TCAM that's 294.968 entries large :-) - so of course, you need some
headroom to cover a few thousand prefixes "mishap", but you can *see*
where the boundary is ("294.968 now, 512k is the limit, we're pretty
safe").

> It seems probability comes into play here, but my gut feeling is that very
> few networks would run into any real issues with tcam aggregation.

I have never modelled it, but indeed, it would be an interesting excercise
to run this through real-world IGP+BGP data on real-world network topologies
and see what happens.

gert
-- 
USENET is *not* the non-clickable part of WWW!
                                                           //www.muc.de/~gert/
Gert Doering - Munich, Germany                             gert at greenie.muc.de
fax: +49-89-35655025                        gert at net.informatik.tu-muenchen.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 305 bytes
Desc: not available
URL: <https://puck.nether.net/pipermail/cisco-nsp/attachments/20100916/4bc94936/attachment.bin>


More information about the cisco-nsp mailing list