Packet-Optical Transport Takes Manhattan

We are gearing up for Light Reading Live's second annual Packet Optical Transport Evolution conference, to be held May 19 at the Marriott Marquis in New York City. With packet-optical products and technologies maturing and with the US economy on the mend, there will be much to discuss, define, and debate at this year's event. Here's a look at the major trends that we expect to see in New York.

100G Transport Standardizes
Industry momentum continues to solidify around 100-Gbit/s transport. At the end of last year, we saw the first commercial deployments of 100G gear in long-haul networks, when Verizon Communications Inc. (NYSE: VZ) announced the deployment of Nortel Networks Ltd. 100G DWDM systems in Europe. In recent months, we've seen two significant 100G trial completions announced. Verizon announced completion of a 100G Ethernet field trial using T1600 core routers from Juniper Networks Inc. (NYSE: JNPR), the SpectralWave DWDM system from NEC Corp. (Tokyo: 6701), and 100G CFP MSA short-reach optical transceiver modules from Finisar Corp. (Nasdaq: FNSR). A day later, AT&T Inc. (NYSE: T) announced the completion of a 100G Ethernet trial involving the new CRS-3 Carrier Routing System from Cisco Systems Inc. (Nasdaq: CSCO) and client-side and line-side 100G optics from Opnext Inc. (Nasdaq: OPXT), in a 900km link between Louisiana and Florida. Both the diversity of suppliers and the focus on standards-based systems are highly encouraging for the future of this fledgling technology.

Packet-Optical Transport Moves to the Core
With the packet-optical transport system (P-OTS) evolution well underway in the metro, we are starting to see P-OTS move into the core, with a new class of products that combine DWDM transport, OTN switching, Sonet/SDH, and Ethernet switching in a single device. Verizon has put its stake in the ground with an RFI/RFP defining its requirements for the next-generation core. Vendors including Alcatel-Lucent (NYSE: ALU), Ciena Corp. (NYSE: CIEN), Tellabs Inc. (Nasdaq: TLAB; Frankfurt: BTLA), Huawei Technologies Co. Ltd. , Hitachi Ltd. (NYSE: HIT; Paris: PHA), and startup Cyan Optics Inc. are moving aggressively to define the right combination of functions for their prospective customers and get their products to market quickly.

Packet Optical Transport Targets Mobile Backhaul
Mobile backhaul for 2G and early 3G networks has been good to Sonet/SDH equipment suppliers over the past few years, but those good times are quickly coming to an end, as Ethernet over Sonet/SDH (EoS) can't scale to handle the advanced 3G and 4G network traffic of the future. To date, Layer 2/3 switch-based networks have provided the only long-term solution for high-capacity Ethernet backhaul (where fiber is available), but operators have had to contend with what to do with their TDM traffic. Synchronization in an Ethernet network requires a new set of technologies to be introduced, and maintaining separate networks for packets and TDM is operationally inefficient. Enter packet-optical transport: The P-OTS value proposition – migrating networks from TDM to packet over time – applies just as much to mobile backhaul networks as to wireline networks. The opportunity for P-OTS in mobile backhaul is immense, provided that suppliers can differentiate their products and their marketing from the "legacy" MSPPs that backhaul operators are moving away from.

We will be discussing these and many related issues at the Packet Optical Transport Evolution event on Wednesday, May 19. Click here to register!

— Sterling Perrin, Senior Analyst, Heavy Reading

DCITDave 12/5/2012 | 4:38:18 PM
re: Packet-Optical Transport Takes Manhattan

A lot of the P-OTS boxes certainly do sound like MSPPs of five or six years ago. I wonder how much development overlap there is between the two types of devices?

Sterling Perrin 12/5/2012 | 4:38:15 PM
re: Packet-Optical Transport Takes Manhattan


I agree there is a lot of confusion on these distinctions - due, in no small part, to MSPP and WDM equipment suppliers that want to call their boxes "packet optical" in order to get included in RFPs, as well as trade press articles and analyst reports. Fundamentally, MSPPs and P-OTS have different missions - MSPPs were built for predominantly Sonet/SDH networks and P-OTS are built for predominantly packet networks.

To be more concrete: MSPPs and P-OTS will look very similar in TDM functionality but very different in packet functionality. If it only does Ethernet over Sonet, it has to be MSPP. P-OTS can function as a full Ethernet switch in applications. Connection-oriented Ethernet is another good distinguishing feature for P-OTS. Packet switching capacity relative to TDM capacity is another. If the box has 60 G of TDM capacity but only 10G Ethernet, it's not P-OTS. In the core, large scale OTN switch fabrics are big defining feature. No MSPP will have this.

If a vendor says their MSPP can do all of the functions I've listed above, then I'd say their MSPP is actually a P-OTS in a counter-intuitive disguise.


torivar 12/5/2012 | 4:38:07 PM
re: Packet-Optical Transport Takes Manhattan

I look at the boxes out there today and to me what separates a true P-OTS from the others is the ability to statmux multiple packet data flows originating on different client ports into a single fixed TDM channel without the use of external equipment.      So it requires having at least an Ethernet fabric which can take multiple client VLANs and map those to a transport signal or signals.   

I know some boxes will use ODUFlex to pack multiple GE ports into a wave and call itself a P-OTS, but people have been doing that for years using proprietary mechanisms, it's not packet aware. 


Sterling Perrin 12/5/2012 | 4:38:00 PM
re: Packet-Optical Transport Takes Manhattan


I agree that a packet fabric is required to be considered true packet-optical transport. There is a lot of momentum globally right now around ODU0/ODUflex for core transport - which I would consider a component of packet-optical transport (though packet fabric would still be needed). I'm curious though about your comment on proprietary mechanisms for ODU0/ODUflex functionality. My understanding is that this is very new. Can you elaborate on the proprietary options?


Sign In