Hi Robert,

I agree that LDP and RSVP-based MPLS transport tunnels between hypervisors
or TORs isn't ideal as it is.  IP-based tunnels are better in this regard
with MPLS-based VPNs.

Best regards. -- aldrin

On Sunday, December 9, 2012, Robert Raszuk wrote:

> Aldirn,
>
> > Now with support for MPLS in merchant silicon, I don't see any good
> reason
> > why MPLS-based DCVPN solutions (IPVPN, E-VPN) should be held back
>
> For service demux I see no issue as well
>
> For transport I see two major issues:
>
> - MPLS requires new signalling protocol ... DC fabric and hosts which
> do act as PEs should be as simple as possible, but not simpler hence
> introduction of LDP or worse RSVP-TE to signal the labels seems not
> helpful.
>
> - MPLS FECs can not be summarized. With IP we just need information
> how to reach subnet X ... with MPLS (even if one would provide the
> relaxed match) FECs are still /32s That's a lot of them in large data
> centers.
>
> Cheers,
> R.
>
_______________________________________________
nvo3 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nvo3

Reply via email to