[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[safnog] Prefixes in IGP vs iBGP
-----BEGIN PGP SIGNED MESSAGE-----
On 5/12/2015 7:41 AM, Graham Beneke wrote:
> On 08/05/2015 14:24, Mark Tinka wrote:
>> A bit disappointing that IS-IS in Quagga/Zebra has been very
>> poorly implemented, and that has not changed since I started
>> testing this back in 2011. So we use OSPFv2/OSPFv3 for Anycast,
>> and redistribute that into our IS-IS core.
> This raises an interesting question for me:
> Why do you choose to distribute this in your IGP rather than as a
> iBGP prefix?
> The basic rule is that all customer prefixes should be in iBGP. I
> have typically extended that to be any hosts and services on the
> network are 'customers' of the network (even if they are
and then I would (like Mark) exclude "internal anycast" - ie prefixes
only seen/anycasted inside your network, typically v4 /32s (as opposed
to the "eBGP-anycast"). So these go into IGP as well.
If you put your service address (say for the DNS resolver(s)) in iBGP,
what would be the next-hop for that prefix?
something specifically added to IGP for this purpose?
(that otherwise wouldn't be in the IGP)
With IGP you get to the closest server (by metric).
With iBGP, you have to ensure several other parameters are identical
to get load distribution (as opposed to all-or-nothing). and it comes
back to IGP metric of the next-hop...
PS: also doing things as Mark described.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
-----END PGP SIGNATURE-----