[j-nsp] [c-nsp] MPLS-TP on CPT platform vs IP/MPLS core on ASR with TE
Mark Tinka
mark.tinka at seacom.mu
Mon Nov 25 21:52:49 EST 2013
On Tuesday, November 26, 2013 01:33:13 AM Yham wrote:
> If a provider already have ciena as transport with
> ip/mpls core on cisco ASR, why would they want to deploy
> CPT with mpls-tp?
Pretty much every optical vendor today is implementing MPLS-
TP in their platforms, as a way to gain a share of the
"packet space".
Our side is also not standing still - the vendors have been
plugging DWDM ports into routers for a while now. The model
hasn't been successful (you can see how IPoDWDM was a total
disaster), but Cisco and Juniper are trying again. Cisco
with the NCS, and Juniper with the PTX.
So optical and routing vendors are trying to bite into each
others' space.
Some optical vendors are implementing MPLS-TP, while others
are even considering adding full MPLS switching capability
in their optical gear. The whole thing smells funny to me
(can my favorite optical vendor support p2mp LSP tunnels in
their device?).
At any rate, I never liked MPLS-TP because:
1. It doesn't support ECMP.
2. It is static (although some implementations offer
easier provisioning via an NMS).
3. It's not dynamic enough a la MPLS in the IP
world.
IMHO, MPLS-TP was just the ITU trying to find a way to admit
that Ethernet is better than SONET/SDH/TDM, but still keep
their fan-base happy by staying with the deployment
ideologies of old.
Not many folk I know run STM-256/OC-768 (40Gbps). For all
intents and purposes, after STM-64/OC-192, you need to run
Ethernet if you want to scale up (40Gbps, 100Gbps, 400Gbps,
1Tbps, e.t.c.). I don't see the ITU trying to develop higher
bandwidth beyond 10Gbps/40Gbps over legacy protocols.
Ethernet has proven its case already.
There are enough tools in vanilla IP/MPLS to make it
reliable enough for high-end data. I mean, what do you think
IPTv runs on?
Mark.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
URL: <https://puck.nether.net/pipermail/juniper-nsp/attachments/20131126/2068df8b/attachment.sig>
More information about the juniper-nsp
mailing list