[c-nsp] OSPF design (danger will)

[Christopher J. Wargaski]

While conventional routing practices say to run the link state routing
protocol internally, and to use the path vector protocol (i.e. BGP)
externally, I agree with Danger Will. It is quite a pain to run OSPF
and BGP simultaneously, especially when you have redistribution to
employ and administrative distances to play with for secondary paths.

IMHO, run BGP (iBGP + eBGP) end to end and life will be much simpler.

cjw


> Date: Sat, 23 Oct 2010 09:58:17 -0700 (PDT)
> From: danger will <myniuid at yahoo.com>
> To: "Robert Crowe \(rocrowe\)" <rocrowe at cisco.com>, Heath Jones
>        <hj1980 at gmail.com>
> Cc: cisco-nsp at puck.nether.net
> Subject: Re: [c-nsp] OSPF design
> Message-ID: <946936.1564.qm at web55008.mail.re4.yahoo.com>
> Content-Type: text/plain; charset=iso-8859-1
>
> First of all most of the mpls providers dont even support ospf as routing protocol between the CE and the PE (and the CE most of the times is managed by the ISP too contrary to the popular belief) . The reason why is very debatable and there are good arguments on both sides both the ISP and customer but wont get into them.
> You probably could run ospf between your router and the CE but you will have issues with the backup. if you dont have any backup then its fine.
> I have also given this a lot of thought and the answer is just run bgp thats the only real solution at the moment.
>
> O.
>