Ethernet in the 'Hood
Potential members of the group include startups like Alloptic Inc. and OnePath Networks, as well as a few seasoned giants, such as Cisco Systems Inc. (Nasdaq: CSCO) and possibly Corning Inc. (NYSE: GLW).
PONs can be built using relatively inexpensive telecommunications equipment that splits a single strand of fiber and allows the bandwidth to be shared among multiple users (see PONs: Passive Aggression). It's been touted by carriers as an economical way to roll out fiber optic technology to homes and businesses.
Most of the handful of PON products now available use ATM (asynchronous transfer mode) technology, based on a suite of so-called Full Service Access Network (FSAN) specifications. These specifications were created over a ten-year period by vendors and carriers worldwide and adopted last year by the International Telecommunication Union (ITU).
Sources say it's time for a new approach, particularly when it comes to residential access.
"ATM was the right solution ten years ago, when FSAN was formed," says one vendor representative, who asked that his name and company be withheld. "But nothing's happened. The specs took ten years to write and they're still not finished. Trials haven't resulted in many new services. Meanwhile, Ethernet offers a faster, cheaper way to solve the problem."
It's a logical leap. Ethernet has gained wider acceptance in recent years, especially for use in metropolitan area networks and homes. Basing PONs on Ethernet instead of ATM could widen the market because standard Ethernet components are cheaper than those based on ATM. Gigabit Ethernet technology also provides speeds as high as 1 Gbit/s today, while FSAN technology only reaches 622 Mbit/s.
"The edge is ready and waiting for Ethernet PONs," says William Markey, cofounder of RelevantC, a consultancy specializing in residential access. "Users have Ethernet-compatible PCs. Ethernet is univerally supported as an IP protocol. From our perspective, Ethernet is the latest no-brainer in residential access."
Organizers backing Ethernet PONs plan to meet early in January to discuss the official setup of a group. Reportedly, the goal is to create a consortium, in which participating vendors will pay membership fees. Once the format is decided, the group hopes to begin work immediately on specs that can be fed to official standards-setting bodies, such as The Institute of Electrical and Electronics Engineers Inc. (IEEE) or the ITU.
"We're starting a consortium to spin out and speed up PONs for home access and work through the issues," says Dave Stehlin, CEO of OnePath , an Ethernet PON startup.
The IEEE created a separate Ethernet PON project in November, including a study group called "Ethernet in the First Mile," which is focusing on Ethernet PONs as well as xDSL. Interest in the new IEEE group was reportedly overwhelming, and a meeting will convene at the IEEE P802.3 interim committee meeting in Irvine, Calif., the week of January 8, 2001, to discuss its agenda.
But potential members of the new Ethernet PON consortium say the industry shouldn't wait for the IEEE. The IEEE could take years to standardize new specs, and it's important to get industry consensus on a standard spec as quickly as possible.
"We've seen this movie before," says RelevantC's Markey. He says the efforts to promote an Ethernet PON spec remind him of efforts to create an Ethernet-based spec for cable modems, which ultimately resulted in the data over cable systems interface specification (DOCSIS -- see Last Mile Lexicon). "The two years it took for DOCSIS to form resulted in vendors losing key opportunities in the access market," he says. He hopes the lesson isn't lost on the PON vendors.
On the downside, some analysts warn that Ethernet PONs may not meet everyone's expectations. "The PON market will grow, but not necessarily because Ethernet is used instead of ATM," says Alan Bezoza, access analyst at CIBC World Markets. He says the cost of laying fiber, which can run carriers $300,000 to $700,000 per mile (approximately 70 percent of which are labor costs that aren't likely to come down over time), is what's holding up PONs in general. And he argues that the price difference between ATM PONs and Ethernet PONs "won't be all that significant on a per-subscriber basis."
Nevertheless, would-be members of the new consortium are intent on their goal. "We just want to show the ATM folks that other options make sense," says OnePath's Stehlin.
--Mary Jander, senior editor, Light Reading http://www.lightreading.com