PBB-TE: What Now?

In summer 2006, as Nortel Networks was announcing its new metro Ethernet division, new executives — at the time working to quickly redefine the vendor's strategy — boasted about how quickly the unit had been formed in response to an emerging market opportunity.

“Six weeks ago, we said, ‘Hey, this metro Ethernet thing looks really interesting,’” George Riedel, chief strategy officer for Nortel, told Telephony at the time. “We didn't spend six months debating it or analyzing it. We had a one-day session where we had a couple plans come together. We had a follow-up discussion. The next week, we decided to go do it.”

But as Nortel found out, the market can shift in the opposite direction pretty quickly, too. In the time between last year's publication of Telephony's Guide to Carrier Ethernet and the one you're reading now, the vendor ecosystem surrounding the new technology that propelled the creation of that metro Ethernet division — the connection-oriented transport technology first called provider backbone transport that standards bodies later renamed provider backbone bridging - traffic engineering (PBB-TE) — has been hit hard. One of the first vendors to follow Nortel's lead in promoting PBB-TE, Hammerhead Systems, recently ceased operations. And another, Soapstone Networks, is weighing strategic alternatives after admitting the market did not develop as hoped. In March, Nortel cut investment in the carrier Ethernet switch/routers based on PBB-TE after disappointing market penetration.

Later that month, Philippe Morin, the president of Nortel's metro Ethernet networks division — insistent that PBB-TE simply represented too dramatic a technological shift for carriers at a time when the macroeconomy made them excessively cautious — told Telephony the vendor was still planning to bring PBB-TE to its OME 6500 metro core optical platform (though he wouldn't say when). But pushing PBB-TE may become even more of an uphill battle going forward.

One of PBB-TE's limitations since its inception has been its point-to-point nature. Billed as a simpler alternative to MPLS, it is comprised of manually provisioned “tunnels” of dedicated bandwidth. Soapstone's control plane was designed to help automate provisioning in PBB-TE networks as the technology evolved to fit multipoint-to-multipoint networks (also called E-LAN networks), but the company said the technology never quite got there.

“Because [switch/router vendors] simplified the technology to such an extent, some switches ended up taking out too much of the information you needed to support an E-LAN service,” said Esmerelda Swartz, senior vice president of marketing and business development for Soapstone.

Soapstone is still working on that problem with partners, possibly leveraging virtual private LAN service (VPLS) technology or extending to E-LANs the information used in point-to-point networks to set up IS-IS link-state protocols. Nortel even developed its own control plane last year based on provider link state bridging (PLSB), a link-state protocol based on existing Layer 2 standards that populates forwarding tables to construct network pathways for Ethernet services.

But there's another problem that continues to nag PBB-TE, said Tom Nolle, president of CIMI Corp., a consultancy. Using carrier Ethernet tunnels to transport services around the metro requires tight integration between the management of the tunnels, the Ethernet underlying them and the services within the tunnels, Nolle said.

“That kind of service management integration is simply not available today,” he added. “If they could solve the management problem quickly, then [PBB-TE] offers a lower [total cost of ownership] in both metro and even national infrastructures than IP/MPLS would, as long as we're talking about legacy and premium services. The problem is: Operators have ceded this [management task] to the standards bodies, which are proceeding at an order of magnitude slower pace. As a result, it's my sense that it's not possible to fix the management problem in time for [PBB-TE] to be as successful as it could have been. So it's becoming increasingly likely that — for reasons that have nothing to do with Ethernet or [PBB-TE] and everything to do with management — [PBB-TE] will end up losing out to MPLS-TP.”

MPLS-TP, or multiprotocol label-switching - transport profile, is a connection-oriented transport technology based on Layer 3 MPLS routers as opposed to Layer 2 Ethernet switches. It was proposed last year as compatibility concerns grew over T-MPLS, a transport technology being standardized by the ITU-T. Like PBB-TE, T-MPLS was accused of disabling too many important features in the name of simplicity. But shifting focus to the newer technology, MPLS-TP, means waiting longer for industry standardization, potentially giving PBB-TE a better shot at mass adoption today.

However, with capital budgets constrained by a harsh economy, waiting for MPLS-TP to mature might not seem like such a bad idea to many carriers. For example, Verizon talked about plans to deploy MPLS-TP at the OFC/NFOEC optical trade show in March. And in fact, existing vendors of T-MPLS gear are working hard to convince service providers to deploy T-MPLS today with the promise that eventual upgrades to MPLS-TP, when it's ready, will be seamless and inexpensive.