Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 9 May 2007 15:12:21 -0300
From:      AT Matik <asstec@matik.com.br>
To:        "Kirk Davis" <Kirk.Davis@epsb.ca>
Cc:        Freebsd-ipfw@freebsd.org
Subject:   Re: Policy Routing natd+ipfw
Message-ID:  <200705091512.22501.asstec@matik.com.br>
In-Reply-To: <DB9A31C316524A4A83E54A2C0D2065570240029E@Exchange24.EDU.epsb.ca>
References:  <33910a2c0705041812s2aaf0b62t785e16abc0decee6@mail.gmail.com> <200705090647.31588.asstec@matik.com.br> <DB9A31C316524A4A83E54A2C0D2065570240029E@Exchange24.EDU.epsb.ca>

Next in thread | Previous in thread | Raw E-Mail | Index | Archive | Help

On Wednesday 09 May 2007 14:05:52 Kirk Davis wrote:
> >
> > I do not know enough about quagga but if you really run BGP
> > and quagga does
> > what BGP is supposed to do I wuold say you shoudl use policy
> > route-map
> > filters for that purpose
>
>    We are probably getting a little off topic for the ipfw list now ;-)
>

well, maybe we will be forgiven :)

>    BGP route-maps will not do what I need.  I am not trying to change
> the routes advertised to my peers (or change the incoming ones that I
> receive).  What I really need to virtual routing tables that I can then
> control how they are updated from the BGP.  Since FreeBSD only has one
> core routing table then I seem to have to use the firewall rules to
> modify the routes.  It works but it is a kludge and doesn't scale well.
>


bypassing bgp with policy forwarding rules does not change route advertisin=
g=20
to the bgp neighbour and vice-versa. You can do "redistribute static" if yo=
u=20
are an endpoint but would not be wise eventually. Anyway the advertised=20
routes need to be announced by your bgp router upwards and not by any=20
artificial routing scenario otherwise there is no way to say that you get t=
he=20
traffic back over the same route, even if you frame bgp and they go out ove=
r=20
path 1 you may get them back over path 3,4,5 or any other bgp may decide. A=
nd=20
that is the point at the end, bgp does the routing decision when you are=20
running bgp. So it does not matter which routing capacities your OS has=20
because it comes after bgp did it's job.

Jo=E3o

>    I haven't played with them yet but the changes to ipfw may get me
> closer to what I am looking for although ipfw probably isn't the best
> place to do the full routing solution.
>
> ---- Kirk
>
>
>







A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura.
Service fornecido pelo Datacenter Matik  https://datacenter.matik.com.br



Want to link to this message? Use this URL: <http://docs.FreeBSD.org/cgi/mid.cgi?200705091512.22501.asstec>