BT Likes Nortel's New Ethernet Flavor
BT's backing for Nortel's technology comes just as the carrier awaits responses from the vendor community to its major Ethernet equipment Invitation to Tender (ITT), a critical stage in BT's decision-making process about which Ethernet technology it will use in its £10 billion (US$19 billion) next-generation network, the 21CN. (See BT Issues 21CN Ethernet RFP.)
Nortel's approach is to simplify Ethernet by turning off features, such as the Spanning Tree protocol and MAC address flooding, and enabling network operators to manage their Ethernet service delivery by using a set of standards-based management tools. (See TelecomNext: Even More Ethernet.)
"We're taking Ethernet and turning off all the bits that carriers don't like," says Simon Parry, Nortel's Carrier Ethernet product line manager for EMEA. "It dumbs down existing Ethernet switches and treats them like SDH crossconnects."
Nortel's claim -- and this is where the controversy comes in -- is that this approach gives carriers the control and scaleability they want from their Ethernet networks without relying on the expensive MPLS technology many carriers use to support carrier Ethernet today.
The cost factor is clearly at the front of BT minds. Some BT executives visiting Light Reading's recent Ethernet Expo in London said the carrier was supporting PBT because "we are concerned about the cost of MPLS."
Nortel's Parry adds: "BT loves it, and we've been pushing it together" at standards bodies such as the International Telecommunication Union (ITU) and Internet Engineering Task Force (IETF) .
BT hasn't been hiding its admiration. The carrier's group technology officer, Mick Reeve, spoke about the BT's support for PBT during a recent interview with Light Reading's TV service, LRTV.
"There are inherent limitations in the way the Ethernet protocol works that limit the scaleability on the wide-area network," Reeve told LRTV. "Things like self-discovery [and] broadcast packets work well in an enterprise, [but once you] start to have many, many thousand end points or a big-scale network, some of those things start to run out of steam effectively.
"So we proposed this idea that the standards world calls 'provider backbone transport,' which is a way of reducing and removing those limits, but still trying to keep the inherently low-cost nature of the Ethernet hardware, which is why we would like to use it."
(See the LRTV video that includes Reeve's interview by clicking here.)
Reeve also made a recent presentation to a group of MEF executives about PBT, where he praised the technology for being standards-based, scaleable, resilient, secure, providing OA&M (operations, administration, and maintenance) and simple fault-finding, and having traffic engineering supported by default.
"It's as easy to operate on a carrier scale as SDH, with the costs and flexibility of Ethernet," he told the MEF gathering. "It gives SDH-like services with the costs, granularity, and plug-and-play capabilities of Ethernet."
Reeve also identified Ethernet as the "de facto Layer 2 transport for backhaul" between the MSANs (multiservice access nodes) and metro rings in the 21CN architecture. Triple-play and wireless backhaul are two of the key areas Nortel has identified as being relevant for PBT deployment.
Even rival vendors, talking under conditions of anonymity, grudgingly admit that Nortel has done a good job in developing a different Ethernet story and gaining the backing of BT.
One executive noted that Nortel was making "a good case for PBT" in ITU and IETF groups, supported by BT, adding that the U.K. carrier's SDH history has no doubt attracted it to PBT, which he described as "an Ethernet flavor of SDH, in very basic terms."
But he added that PBT has its downsides, too. First, it is focused on point-to-point services, whereas carriers are now seeing demand for point-to-multipoint -- demand that has been driven by the uptake of IP VPNs.
The demand for Ethernet VPNs can't be met with PBT, says the source, adding that an MPLS-based VPLS (virtual private LAN service) approach is already the standard choice for such services, though he added that large carriers such as BT are likely to deploy multiple approaches to enable different flavors of Ethernet service.
The second downside is that the PBT technology has to be fully integrated into the existing network management and provisioning systems, and that it's not yet clear how that would be achieved. PBT might simplify the network, says the executive, "but it pushes the complexity up into the management layer, and that makes it even more complicated."
Another Ethernet industry source adds that, while PBT is "promising end-to-end QOS [quality of service] and centralized management" for Ethernet circuits, such capabilities are already on offer from the MPLS-focused suppliers. He suggests that BT and any other carrier looking at PBT should take a closer look at some of the technologies already deployed by other carriers, and not be taken in by Nortel, which is trying to manufacture a position in the market "to make up for its lack of an MPLS story."
But with such vocal support for PBT from the likes of Reeve -- a well respected and key technology influencer at BT and in the industry in general -- Nortel looks well placed to find its way into the 21CN process. The Canadian vendor missed out in the first major round of 21CN infrastructure contract awards when BT named eight preferred suppliers for access, IP, and transmission infrastructure. (See Vendors Sign BT 21CN Contracts and BT Closes 21CN Deals, Touts IPTV.)
However, as former Marconi staffers know, public support and favored technology doesn't guarantee a 21CN purchase order -- the price and conditions have to be right, too. (See Marconi in Turmoil and Ericsson Buys Bulk of Marconi.)
— Ray Le Maistre, International News Editor, Light Reading