Fabric of the Future

Packet Optical Transport Systems Support Transition to Ethernet/WDM Networks

Most of us know POTS as plain old telephone service, but there may be a new technology in need of that acronym — packet optical transport systems. While it is unlikely that we will switch the nomenclature, it is more than likely that service providers will be switching to this new class of data-aware optical networking gear.

Click to Enlarge

According to Infonetics Research, packet optical transport systems combine the functionality of SONET/SDH, Ethernet and WDM/ROADM. “Packet optical transport is the beginnings of that move to displace and replace SONET/SDH,” says Infonetic’s principal analyst Michael Howard, describing a migration that should continue for another 15 years. “Since it’s going to take such a long period of time, you need the transition platforms that have Ethernet and SONET/SDH and WDM/ROADM.”

Eventually, Howard says, service providers want to get to Ethernet and WDM. “Once they get everything out of SONET and onto packet, they don’t need that SONET/SDH part, so that fused Ethernet/WDM is a packet optical transport in a purer form and the eventual goal.”

There are parallel efforts by WDM and SONET vendors to attack the problem, but both are evolving to this new model, says Howard, who expects to officially begin tracking packet optical transport systems in August.

The timing is good for Fujitsu Network Communications Inc., which in early June announced its Packet Optical Networking Platform. The company engineered this solution from the ground up to integrate Ethernet, ROADM and SONET transport technologies on a single optical networking element. But Howard says the FLASHWAVE 9500, more than most others, represents a modern architecture designed from “today’s point of view.”

“What this is, is not a combination of packet box and an optical box,” explains Sam Lisle, Fujitsu’s market development director. “It’s really an optical networking box that is optimized for this packet-centric environment. We take everything from this optical networking heritage and apply it to the problem du jour.”

Today’s network challenge, Lisle says, is the stress of high-bandwidth, high-quality, multimedia packet-based services, like high-definition TV or enterprise collaboration. The FLASHWAVE 9500, he adds, allows operators to respond with “manageable scalability.”

“Since we thought about this from day one for Ethernet, ROADM or SONET, we designed the system so any slot can support any type of card,” says Lisle. Significantly, the FLASHWAVE 9500 includes a patent-pending universal switch fabric that allows both SONET-based traffic and packet-based traffic to be switched, groomed and managed in its native format without any form of circuit-emulation technology. “And, it knows which way to switch the traffic regardless of how it comes in,” explains Lisle. “Other approaches have been taking all your TDM bandwidth and chopping it into packets and switching them as packets and reassembling them as circuits on the outbound. We are not doing anything funny like that. All the traffic is switched in its native format.”

This means that all 480 gigabits of capacity can be SONET or packet, or any ratio of each in between, in 20-gigabit increments, allowing for a modular system. Because we have been able to put all this in an ASIC, it’s a low-cost fabric,” Lisle adds, and explains that it allows carriers to use only SONET and pay a SONET-like price, and use only Ethernet for an Ethernet price.

Alcatel-Lucent has a similarly architected offer in its 1850 Transport Service Switch, which has been available in European Telecommunication Standards Institute (ETSI) markets since fall 2005 and in the North America since summer 2006. In contrast to an MSPP, the 1850 TSS delivers TDM, WDM and packet natively, and does not map TDM into packet and Ethernet over SONET. A spokesperson for the company, then Alcatel, told xchange at GLOBALCOMM that the 1850 TSS stands in contrast to an MSPP, which is designed to carry SONET and is adapted for data. “It’s great for its ‘multiservice-ness,’ but you can only get a certain percentage of data traffic. That’s inefficient,” he said.

Like Fujitsu’s FLASHWAVE 9500, Alcatel-Lucent’s 1850 TSS relies on a unique “Universal Switch” technology, which can be used for either TDM or packet, as opposed to having line cards side by side dedicated to each. The result is that any ratio of packet to TDM can be provided with the capacity of the system. AT NXTcomm, Alcatel-Lucent plans to announce it has expanded the capacity of the 1850 TSS from 100gbps to 320gbps.

There are other systems with similar functions. Turin Networks Inc.’s Traverse Multiservice Transport Switch, for example, is able to support SONET or Ethernet natively. Turin will announce at NXTcomm that its platform will support 10-gigabit Ethernet.

What’s different about the new breed of packet optical transport systems, says analyst Howard, is that they also include WDM, and more specifically ROADM. ROADM delivers cost-effective bulk transport of WDM and operational efficiencies in managing wavelength capacity.

For this reason, Howard includes Cisco Systems Inc.’s ONS 15454 in the packet optical transport category. In March, Cisco added to its MSTP an eigth-degree ROADM and an MSPP on a blade. “One of the interesting things is not only can you consolidate an entire shelf into a blade, [but] the ports on this are actually software-selectable, so you can start with SONET and migrate to Ethernet,” says Mike Capuano, senior manager of service provider solutions marketing for Cisco.

Fujitsu’s FLASHWAVE 9500 supports up to an eight-degree ROADM on a shelf. At NXTcomm, Alcatel-Lucent will announce that the 1850 TSS-320 R.2.0 will support a multidegree WSS-based ROADM.

Another critical part of the packet optical transport discussion as it concerns Ethernet is that it can’t displace SONET/SDH the way it does today, says analyst Howard. “It has to have deterministic and circuit-like functionality. There’s a number of protocols that can fill that bill. We are just seeing the start of them. I think we will be seeing more and more,” he says, adding the primary contenders are MPLS, pseudowire, Provider Backbone Bridging – Traffic Engineering (PBB-TE) and T-MPLS. “The capability of the packet transport and whatever protocols are actually used need to be in the optical gear, so that’s an essential ingredient of these packet optical transport boxes.”

Many systems like the FLASHWAVE 9500 support pseudowire and MPLS technologies to deliver connection-oriented Ethernet since these protocols are standardized. Meanwhile, T-MPLS is under development in separate efforts at the IETF and the ITU, and PBB-TE is being worked on at the IEEE (see related story below).

Promoting PBT

At NXTcomm, Nortel Networks Ltd. will announce the creation of the Carrier Ethernet Ecosystem, a multivendor industry initiative with an initial goal of promoting Provider Backbone Transport (PBT), an effort the gearmaker has been driving to make Ethernet deterministic.

John Hawkins, Nortel’s carrier Ethernet marketing manager, says the initiative is necessary not only to boost the nascent technology but because it’s been inaccurately portrayed as a Nortel proprietary technology. “We want the industry to realize that not only are there multiple solutions out there, but we are working with competitors and others alike to make sure that they interoperate and make sure that they are truly solutions that can be rolled out,” he says.

And, he says, the ecosystems’ members have agreed to support the IEEE’s ongoing standardization efforts and to implement the standardized technology, whatever that may be. PBT and its IEEE alter ego, Provider Backbone Bridging – Traffic Engineering (PBB-TE), he says, basically are the same since they haven’t had time to deviate from one another.

The Carrier Ethernet Ecosystem will have two goals: interoperability testing among vendors using PBT and joint marketing around carrier Ethernet and PBT.

Hawkins says the ecosystem is meant to be an industry initiative, but Nortel is driving it for now. In addition to vendor members, he hopes to have proof of concept (i.e., interoperable products) to announce at NXTcomm, but he says the real testing is not likely to begin until the fall.

Nortel will be announcing the creation of an Innovation Center in its Ottawa labs where the interoperability testing can take place. Certificates of Compatibility will be issued to compliant systems, software and components.

So, why not call it the PBT Ecosystem? Hawkins says the organizers wanted the name to be broader in case there are other carrier Ethernet issues that come along in the future that the group can address.

Although PBT will be the primary thrust, some of the marketing and education — white papers and Webinars — will be around carrier Ethernet in general.

“PBT is a way that we devised a few years ago to make Ethernet into a deterministic technology,” says Hawkins. “We found a way to have our cake and eat it too. In other words, to be packet-oriented and flexible with the sizing of the links and the resources they consume, but also to be able to predetermine what those resources are. So we are making a packet-oriented network deterministic or predictable or traffic-engineerable.”

PBT and PBB-TE is a response to MPLS, which is considered by many to be too complex and expensive to operate in the transport and aggregation segments of the network. T-MPLS came out of the MPLS crowd as a reaction to the criticism. T-MPLS also is undergoing standards development in parallel efforts at the IETF and the ITU.

Share this article: Email, Slashdot, Digg, Del.icio.us, Yahoo!MyWeb, Windows Live Favorites, Furl
Add this article feed to: RSS, My Yahoo, Newsgator, Bloglines