Cable Tech

'First Mile' Ethernet Hits Snags

Copper connections remain the sticking point for the Ethernet in the First Mile (EFM) standard, as two factions remain divided over which flavor of digital subscriber line (DSL) technology to support.

Participants in the 802.3ah task force had hoped to settle the issue during last week's Institute of Electrical and Electronics Engineers Inc. (IEEE) meetings in Hawaii, but the matter is now slated for a showdown at the group's January meeting, in Vancouver.

Other parts of the 802.3ah standard are proceeding apace. All reports say the fiber optic portions of the group have reached early agreement; and even the copper camp has found some common ground, assigning very-high bit-rate DSL (VDSL) as the technology to support for short-reach links supplying 10 Mbit/s at distances up to 750 m.

Standards for the long-reach connection over copper, however, have hit a stalemate. Most of the chip vendors in attendance would like the standard to declare asymmetric DSL (ADSL) as the technology to use here, but carriers appear to favor G.SHDSL, a version of DSL created within the International Telecommunication Union (ITU) (see First-Mile Ethernet Camps Dig In).

The dispute does not put the entire 802.3ah effort in jeopardy, but it's likely to delay ratification of the standard beyond the September 2003 target date.

Similar disputes arose on the fiber optic side of the standard, but participants simply agreed to allow two options, assigning a different port to each one. That same two-port method should apply on the copper side, says Behrooz Rezvani, CTO of chip maker Ikanos Communications Inc., which is supporting the ADSL side.

"That's the best way to go forward. It allows Ethernet to be available in the network both to business customers and to residences," Rezvani says.

Another wrinkle is that there's a separate war over which modulation scheme to use for Ethernet over VDSL -- quadrature amplitude modulation (QAM) or discrete multitone (DMT) (see Ethernet Over Copper: Now You’re Talking). Rezvani believes the G.SHDSL camp is acting simply to defend QAM.

"This other group wants to win it all -- they want G.SHDSL [at the exclusion of ADSL], essentially. But the industry now could face a problem with that. ADSL is such a pervasive technology. It's doing very well in terms of interoperability and availability," Rezvani says.

Not so fast, say G.SHDSL supporters. According to Dan Carnine, market development manager for chipmaker Metalink Ltd. (Nasdaq: MTLK), his camp is trying to adhere to the spirit of IEEE standards, where there's typically only one physical-layer chip, or PHY, that comes out of a given project -- while the short- and long-reach copper options are bundled together in a single IEEE project authorization request (PAR). Moreover, the only reason 802.3ah includes a long-reach copper option is because SBC Communications Inc. (NYSE: SBC) requested it -- and SBC wants G.SHDSL supported there, Carnine says.

It's debatable whether the IEEE would frown on a two-PHY approach. "It depends on how you interpret certain rules in IEEE," Rezvani says. "Whether they should select only one modulation technology that spans the entire copper spectrum, that's the cause of the problem right now."

In addition, Carnine doubts there's a market for Ethernet over ADSL. His argument goes like this: North American ADSL buildouts are based on ATM, not Ethernet; Europeans can't use the ADSL PHY being suggested because it would encroach on ISDN bands; and Asia isn't interested because 90 percent of the loops there are shorter than 1 km, hence suitable for VDSL.

"There is no demand for an Ethernet-based ADSL. If there were demand, the ADSL vendors would have already developed the technology, and they would be shipping in volume," Carnine says via email.

These kinds of squabbles are common in IEEE standards, as companies try to fit existing technologies into them. The 802.17 standard for Resilient Packet Rings was similarly split before the group settled on a compromise specification. (During the Hawaii meetings, the RPR group managed to complete the latest version of its working-group ballot, a major step towards getting the standard completed.)

For now, the copper issue remains in dispute, with both sides hoping to reach settlement in January. "It all really depends on the mood of everybody at the next meeting," says Rezvani.

— Craig Matsumoto, Senior Editor, Light Reading

Want to know more? The big cheeses of the optical networking industry will be discussing first-mile Ethernet at LightSpeed Europe. Check it out at http://www.lightspeedeurope.com/conference_program.asp.

TooTall 12/4/2012 | 9:18:37 PM
re: 'First Mile' Ethernet Hits Snags You'll see the Ikanos design works well on paper (especially Press Release paper) but not in reality. Check out what the Asians are doing with VDSL that exists today and does work and you'll see a successful FTTC solution already being deployed. FTTC is the interim step between pure copper, slow DSL speeds and FTTH. FTTC is also a complementary technology to FTTH, not a replacement. This is where the reality is.
Interesting how Light Reading continues to take misinformation being put out by the Ikaons types and publishes it as truth when it is far from it.
jayja 12/4/2012 | 9:18:37 PM
re: 'First Mile' Ethernet Hits Snags Ironic that the "future" solution, FTTH, is ready to go, and the "current" solution, DSL, is not.
BobbyMax 12/4/2012 | 9:18:33 PM
re: 'First Mile' Ethernet Hits Snags First Mile ethernet is a good technology for non-critical applications. The EMF technology is particular helpful as it helps convergence of voice and data. There are a lot of competing technologies over which voice and data services can be provided. These techniologies include DSL, ADSL, VDSL, and g.shdsl. Cost factors should be considered carefully before committing to a particular technology.
teebaby 12/4/2012 | 9:18:30 PM
re: 'First Mile' Ethernet Hits Snags I've been doing a little consulting work in this area recently and have been following this group from the sidelines. Its laughable to me how all of the carriers are making their requirements known and group isn't listening. Do the DMT vendors really think they're helping their cause by fighting so hard with their customer? I think Burger King should change their slogan from "Have it your way" to "Take it this way and shut up." Some people just don't get it.
lightmaster 12/4/2012 | 9:18:29 PM
re: 'First Mile' Ethernet Hits Snags "First Mile ethernet is a good technology for non-critical applications"

Just what is a non critical application? I think it something that used to exist (i.e. Internet), but doesn't anymore. In my business, I'd rather have my phone line down than my Internet connection.

Perhaps those businesses who don't depend on the Internet, but want to provide a way for their employees to shop online or surf the net for porn will spend money on a high speed Ethernet connection?

You might as well say that Ethernet is a good technology for non-revenue producing applications. The bottom line is that it HAS to satisfy concerns with reliabilty before people will pay enough for it to make it profitable for the carriers.

firstmile 12/4/2012 | 9:18:26 PM
re: 'First Mile' Ethernet Hits Snags Tee,
I'm with you. It's been one man and one vote, even when it's 7 equipment and/or chip guys voting against 6 carriers. It just doesn't make sense.
jayja 12/4/2012 | 9:18:25 PM
re: 'First Mile' Ethernet Hits Snags Tee, you are right on. Several carriers have complained to me and my company about the "Cisco crowd" in IEEE 802.3ah not listening to their requirements. In fact, the carriers actually ran "stealth" parallel meetings for a while last year to develop their own requirements that were eventually drawn up as a parallel option in the standard. I think the carriers like it better in ITU-T without those pesky vendors around. Frank Effenberger of Quantum Bridge deserves a lot of credit for trying to broker peace at these sessions.
TooTall 12/4/2012 | 9:17:54 PM
re: 'First Mile' Ethernet Hits Snags I would not flag this particular party as helping the consensus effort. This discussion has branched off from the copper track to the optics one. The GPON versus EPON issues are now being worked out (see the EFM reflector) by several leading participants. Also, I am not one to buy into the "ABC" philosophy you are promoting (ABC = Anything but Cisco) as they continue with efforts to move this ahead. They just adhere to the philosophy that Ethernet should be inexpensive and meeting very stringent requirements that the overall market may not require is adding cost.
As to copper, I would think that eventually the DMT group blocking progress, as evidenced in St. Louis, Edinburgh, and other venues will eventually be voted out and we will get some real progress made.
Sign In