[c-nsp] mpls/ip load balancing

[Marlon Duksa]

It actually works over a combination of tagged and untagged interfaces. As a
matter of fact it works for both, LER and LSR cases.
I have two outgoing interfaces for the same route, one tagged and one
untagged.

1) On ingress I'm sending IP packets, and they are load balanced on egress.

2) On ingress I'm also sending tagged packets and they are load balanced as
well on egress.

So, in case:

1) I guess, the hashing is done on ingress and the FIB is consulted. And as
Oli mentioned, the hashing is done on L3 info (src,dst IP)??

2) In this case, LIB is consulted - what is hash based on? IP packets may be
encapsulated few times over in this case (VLAN, another MPLS header...)
Thanks,
Marlon




On Tue, Nov 25, 2008 at 4:15 AM, Ozgur Guler <gulerozgur at yahoo.co.uk> wrote:

>
> Best is  to check "show ip cef internal" to see if the packets are load
> balanced.
> Afaik, IOS does not load balance between labelled and unlabelled paths.
>
> (Think about sending AToM or MPLS VPN traffic out towards an unlabelled
> path.)
>
>
>
> --- On *Tue, 25/11/08, Marlon Duksa <mduksa at gmail.com>* wrote:
>
> From: Marlon Duksa <mduksa at gmail.com>
> Subject: [c-nsp] mpls/ip load balancing
> To: "cisco-nsp at puck.nether.net" <cisco-nsp at puck.nether.net>
> Date: Tuesday, 25 November, 2008, 5:19 AM
>
> Hi - does anyone know how load balancing by default works over two paths,
> one using MPLS and the other one IP. I have the same route advertisement
> received over two different interfaces. On one interface LDP is enabled (and
> of course OSPF), the other interface is only OSPF enabled. Traffic
> is definitely load balanced across the two interfaces, but I was wondering
> what fields the hashing is based
>  on?
> Thanks,
> Marlon
> _______________________________________________
> 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/
>
>
>