[c-nsp] BGP - Announcing routes to Internet providers.

[Ivan Pepelnjak]

Let's back a step and ask the questions we should have been asking in the first place:

* Are you an end-user or a Service Provider (somewhat reliable answer could be gleaned from Drew's e-mail address)?
* What's the size of your network?
* How many uplinks do you have?
* How far apart are your uplinks?

If it turns out Drew's uplinks are close together, all the beautiful design ideas presented here are a huge overkill.

And, BTW, I wish those of you that propose redistributing connected and static routes into BGP a huge budget you'll need to upgrade RAM and TCAM of your routers/switches when everyone decides (after reading this mailing list :) that following your recommendations unconditionally is a good idea :D

Ivan

> -----Original Message-----
> From: Scott Granados [mailto:gsgranados at comcast.net]
> Sent: Monday, January 04, 2010 10:03 PM
> To: Drew Weaver; Cisco-nsp
> Subject: Re: [c-nsp] BGP - Announcing routes to Internet providers.
> 
> Drew, network statements are for the weak.:)
> (I'm kidding of course) but there is a better way.
> You should use community tagging in combination with prefix lists and
> route
> maps.  The idea is that you announce routes according to a tag and the
> behavior of the announcements depends on the specific tag applied.  For
> example, you could tag routes as peers, transits, global announce, etc and
> formulate the type of feeds you give your customers by filtering against
> communities so a customer wants peers and customers only you could match
> the
> two appropriate community tags.  This also allows you to tag the
> communities
> you globally announce uniquely and make the announcements in a unified way
> at your edges.  If you accompany this method with the appropriate
> redistribute static, redistribute connected, etc and use route maps to
> control this behavior you can remove the need for network statements
> completely and greatly decrease the things you need to modify and as a
> result the possible mistakes.  The other upside here is you can mark your
> more specifics as do not export and better control traffic internally
> better
> directing the traffic in your example.  It also allows you to accept
> communities from your customers and have automatic actions taken based on
> the tags they apply.  Let me know if you need some configuration examples.
> 
> 
> 
> ----- Original Message -----
> From: "Drew Weaver" <drew.weaver at thenap.com>
> To: "Cisco-nsp" <cisco-nsp at puck.nether.net>
> Sent: Monday, January 04, 2010 12:35 PM
> Subject: [c-nsp] BGP - Announcing routes to Internet providers.
> 
> 
> > Howdy,
> >
> > I am trying to figure out if there is a different/newer/better(?) way to
> > announce our public IP ranges to our Internet providers, currently we
> are
> > declaring our subnets in 'network statements' in the BGP configuration,
> we
> > have static routes setup like ip route x.x.x.x 255.255.224.0 Null0 254
> and
> > then we have a extended access-list applied to each peer with our net
> > blocks listed in them.
> >
> > It appears that because of the network statements, the supernet routes
> > (/18s, /19s, etc) are being distributed via BGP to the rest of the
> network
> > which is by design(I assume). This doesn't seem ideal because if traffic
> > is sent to an IP address that doesn't have a more specific route than
> say
> > /18, or /19 it travels all the way through the network to the edge
> before
> > stopping. I might be blowing the impact of this out of proportion, but
> it
> > just seems like a waste of resources.
> >
> > Does anyone know of a seemingly more sensible way of doing this?
> >
> > -Drew
> >
> >
> >
> > _______________________________________________
> > cisco-nsp mailing list  cisco-nsp at puck.nether.net
> > https://puck.nether.net/mailman/listinfo/cisco-nsp
> > archive at http://puck.nether.net/pipermail/cisco-nsp/
>