Understanding Fragmentation in MPLS

jrhofman · ‎11-11-2003

We will be converting from a Frame relay to MPLS network. We will be implementing Voice on this new MPLS network. Currently we have FRTS set up between remotes and head-end to insure proper Fragmentaion sizes (among other things). How does this work in an MPLS implementation? IF my head-end is a DS3 sending 1500 byte frames out and my remotes vary from 128K to 512K how do I fragment properly to these small size remotes. The remotes will continue to us frame-relay for the last mile.

zvladov · ‎11-12-2003

Please, explain more.

As for example explain the would-be setup.

It is not clear what you would like to do.

Also you are mixing the things IMHO.

Frame-relay fragmentation?

MPLS doesn't do any fragmentations. Doesn't support it. It is just a switching method.

How you deliver traffic/voice from X to Y

is another question. Of course you can

go through MPLS core, but....

Explain more please.

jrhofman · ‎11-12-2003

I guess you answered my basic question. Voice requires fragmentaion on low speed links to reduce jitter. This can be accomplished in Frame-Relay. If MPLS will not allow fragmentation it may reduce quality in a voice deployment.

zvladov · ‎11-13-2003

If you have a MPLS core you can do this:

(1) (2) (3) (4) (5)

FR----FR-------MPLS CORE------FR-------FR

That is l2connect.

It basically encapsulates the L2 frames over

the MPLS core.

You can do FR frag at the edge (1) and (2)

go over l2connect (AToM=Any transport over MPLS)

over (3) and go to (5).

(2)&(4) are an MPLS edge device.

You can have some prioritizations, on the

l2connect.

http://www.cisco.com/en/US/customer/tech/tk436/tk428/technologies_q_and_a_item09186a008009d4e3.shtml

jrhofman · ‎11-13-2003

Makes sense to me. However, AT&T and MCI will not support FRF.12 until later 2004 in their MPLS deployments regardless of the access type.