OTN? Don’t be gauche, this is the MPLS conference

OVUM VIEW

Summary

At the Marriott Paris Rive Gauche, the venerated MPLS & Ethernet World Congress 2011 conference was held again on February 8–11. A lively debate unfolded about whether MPLS or OTN switching would be the most cost-effective way to manage bandwidth in the core of carrier networks. Optimizing bandwidth management in the network core is a Layer 0–3 challenge. Carriers can continue to utilize MPLS or SONET/SDH for this purpose or introduce connection-orientated OTN switching; their choices will depend on their projections of future traffic flows and their operational preferences. Ovum is undertaking a comprehensive research effort on bandwidth management in the core, with a report slated for 2Q11. 100GE was also on display at the MPLS conference and continues to demonstrate market progress.

Bandwidth management in the core: MPLS vs. OTN

Optimizing bandwidth management in the network core is a complex problem. It is a forecasting and resource allocation problem, because network traffic varies from average to peak and from source to destination. It is also an internal carrier organization challenge. Large incumbent carriers have traditionally conducted network planning on a layer-by-layer basis. Optimizing across network layers requires a new generation of tools, skill sets, and internal organizational flexibility. The specific bandwidth management pain point that irks carriers is paying for IP router resources that are utilized for transit traffic. One optical network (ON) vendor offered, “We have done the carrier studies; 80% of traffic in some routers is for transit.”

The ON vendors have the solution: “router bypass.” The pitch is to utilize ON to bypass router resources that are used for processing transit traffic. This is not new. Optical network vendors have been advocating router bypass via full wavelengths for years. What is new is the introduction of an OTN switching capability that offers the promise of managing at the more refined sub-wavelength level. Within the ON community, the consensus is that router bypass is a good thing, eliminating the need to put transit traffic through the purportedly high-margin IP routers and thereby saving the carriers a significant amount of network capital expenditure.

Say “router bypass” at the Paris MPLS conference, though, and the reaction is viscerally opposed: “What! You would give up on statistical multiplexing?” “Why would you add another network layer to create an opportunity for more network planning complexity and inefficiency?” “If you add OTN, you will have OTN pipes that are only partially full.”

The carriers will have to weigh in here of course. Every carrier’s organizational situation and network starting point is different. Carriers may elect to conduct very detailed multi-layer network planning studies, with a major challenge being projecting the future traffic patterns. To set up a traffic study though, carriers will bring in biases and preferences based on their organizational histories.

Ovum’s view is that the market will support both OTN and MPLS switching. Also, many carriers will not likely make a dramatic break from the past and will continue in an evolutionary manner. If a carrier has always had a multi-service network with a SONET/SDH switching capability, then the adoption of OTN switching would be a “natural” evolution. MPLS-centric carriers will likely view OTN has an unnecessary additional layer and continue with MPLS-centric solutions. The vendors with OTN strengths will align well with OTN-orientated carriers. Vendors with predominantly MPLS strengths will align well with MPLS-centric carriers. And the vendors that have both MPLS strengths and OTN capabilities will articulate the full toolkit capability.

100GE progress continues

100GE interconnecting with 100G WDM was on display at the European Advanced Networking Test Center (EANTC) interop. Some vendors reported shipping “hundreds” of 100GEs. Other vendors have product ready, but their customers have not deployed because of the high line-card cost, which is due to the very expensive 100GE short-reach interface component. Nevertheless, Ovum expects 100GE momentum to continue to build due to the strong desire from the datacom community. With momentum building on the 100GE client side, Ovum expects 100G DWDM to keep pace and gain market favor.

APPENDIX

Disclaimer

All Rights Reserved.

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the publisher, Ovum (a subsidiary company of Datamonitor plc).

The facts of this report are believed to be correct at the time of publication but cannot be guaranteed. Please note that the findings, conclusions and recommendations that Ovum delivers will be based on information gathered in good faith from both primary and secondary sources, whose accuracy we are not always in a position to guarantee. As such Ovum can accept no liability whatever for actions taken based on any information that may subsequently prove to be incorrect.