[177944] in North American Network Operators' Group
Re: mpls over microwave
daemon@ATHENA.MIT.EDU (Eric Louie)
Mon Feb 9 15:41:21 2015
X-Original-To: nanog@nanog.org
In-Reply-To: <54D5CAD5.4000909@seacom.mu>
Date: Mon, 9 Feb 2015 12:41:13 -0800
From: Eric Louie <elouie@techintegrity.com>
To: Mark Tinka <mark.tinka@seacom.mu>
Cc: NANOG <nanog@nanog.org>
Errors-To: nanog-bounces@nanog.org
because I have a partial implementation of MPLS routers. Whether the
routers support MPLS or not, the routing on an OSPF level doesn't depend on
MPLS being enabled. Eventually everything will be MPLS-capable. The MPLS
network is not multiple-path. The OSPF network is.
On Sat, Feb 7, 2015 at 12:20 AM, Mark Tinka <mark.tinka@seacom.mu> wrote:
>
> On 6/Feb/15 00:31, Eric Louie wrote:
> > I work for a fixed wireless provider, and our mpls-capable backhauls are
> > all running mpls with 9200 MTU with no problem. The only weirdness I
> > encounter is if I have multiple equal-cost routes to the same location,
> one
> > over MPLS and one not, end up having ping/unreachable issues from my
> > monitoring equipment. The solution has been to cost one path (the MPLS)
> > lower than the other. The only other problem I had was with radio's that
> > didn't support larger 9000+ byte MTU packets - we've phased that radio
> out
> > for now. if you run MPLS with 1500 byte MTU, you'll have issues with
> 1500
> > byte packets with the DF-bit set. That was a nasty discovery in the
> > production network, your mileage will not vary with that problem.
>
> I'm curious why you'd have multiple paths in your network (equal-cost to
> boot) where some support and others don't.
>
> Mark.
>
