According to Light Reading, Ethernet in the First Mile (EFM) could be ready to roll this summer, if participants in the IEEE 802.3ah task force vote on a proposed standard this July.
Even though formal ratification would be at least six months away, the standard would be solid enough to build products on say industry observers.
EFM extends the reach of Ethernet beyond the LAN. DSL modems, for example, currently use Asynchronous Transfer Mode (ATM) for the backhaul. “Getting rid of all those protocol conversions and transitions is what we’re really after here,” says Craig Easley, president of the Ethernet in the First Mile Alliance and a technology director at Extreme Networks.
Here’s a rundown of what’s inside the 802.3ah spec:
Point-to-point copper:
- Short reach: 750 meters at up to 10 Mbit/s
- Long reach: 2,700 meters, but at only 2 Mbit/s.
- 1-Gbit/s speeds on single- or dual-strand fiber. In dual-strand, one strand is used for upstream traffic and the other for downstream. In single-strand installations, both directions share the strand and are kept separate by frequency-division multiplexing.
- 100 Mbit/s for cheaper switches
- For Ethernet passive optical networks (EPONs). It uses time-division multiplexing to combine users’ upstream traffic. “That whole TDM upstream is seen by IEEE purists as Not Ethernet.
- Provides the ability to do remote bit-error-rate testing and other remote management. Ethernet right now is a link-layer technology. It has no intrinsic management.
The standard is closed to major additions, so no new copper or fiber variants will be added between now and IEEE ratification.
For short distances, vendors bickered over which flavor of DSL should be used to transport Ethernet across long distances. They settled on VDSL for transport. Other standards bodies might follow the T1E1 group, which is expected to decide on a standard this June.
The next 802.3ah meeting will take place in July during the IEEE plenary meeting in San Francisco. If the working group agrees on VDSL standards, they’ll submit the current draft for a working-group vote, putting 802.3ah on pace for full ratification six months later.
Products will likely appear long before then. First-mile Ethernet is a big deal in Korea and Japan right now. It may be less urgent for North American incumbent carriers because of the weakened CLEC competition.
Still, point-to-point or multi-point Ethernet at 100Mbps, may be exactly what 802.16a proponents like Wi-Max and Wi-Fi switch proponents like Vivato need to provide inexpensive backbone connections.
| Incumbant carriers have a problem; how can users get high-speed DSL past 18,000 feet? DSL requires twisted pair to the CO. Or does it? According to Network Magazine, the key to massive DSL buildouts by the incumbant carriers is Passive Optical Networking. It lets RBOCs build a network not unlike cable’s hybrid fiber/coax. PONs let carriers deliver fiber closer to the customer. A CO can run a single fiber stand to a cluster of end users. A passive optical splitter directs traffic to the customer’s premises without active optical components, lowering cost. The residential terminal, however, must have a box that converts the optical signals to twisted pair or coax for the last 100 feet. PON gear at the head end uses lasers strong enough to handle the attenuation that takes place at each splice point. When traffic moves upstream, towards the CO, branch traffic can collide with other users so ATM plays traffic cop. EFM could lower costs dramatically for the backhaul. SBC’s $7 Billion Project Pronto planned to extend DSL service to 77% of its users by 2002 using PONS. But the expense was too great. Video delivery using VDSL is the carrot. Phone companies may roll out Passive Optical Networking to neighborhood pedestals fed by inexpensive, 100 Mbps EFM. Then twisted pair could be used for the last 2,000 feet for phone, high speed internet or Video On Demand.
Dropping a multithousand-dollar DSL access multiplexer (DSLAM) cabinet in the center of a community, fed by fiber, is fine when there are enough users to justify the cost. Symmetricom has another idea. Their GoLong extends DSL range from 14,000 feet to 30,000 feet. Although the extender can cost between $500 and $700 per line it’s still faster and cheaper than ISDN. IDSL - ISDN Digital Subscriber Line, works up to 36,000 feet from the Central Office. Speed and cost is comparable to ISDN but circuits carry only data - no voice.
|
For more information check out Network Magazine (11/00) and PONs case study article as well as Lightreading.com.
The economic viability of VDSL must be questioned. Satellite delivery of video is cheaper. POTs users could get stuck with the bill for underutilized plant. Your local Public Utility Commission is SUPPOSED to watch out for cross-subsidies.
Nevertheless, RBOCs may be compelled to rollout 10Mbps-50 Mbps with Video On Demand trials next year. VDSL is a natural evolution of DSL.
But new fiber plant (especially FTTH), is not necessarily cost/effective. What’s wrong with a coop, anyway? The phone company isn’t going to get rich quick with VOD. RBOCs like Qwest are already $20 billion in debt. It’s fancy dancing and baroque politique.
I’ve said this before and I’ll say it again. I think the solution for RBOCs is 802.16e/802.20. They could deliver 1 Mbps DSL/voice. Anywhere. That attacks their main enemy - cellular - head-on and lets them deliver DSL at less cost over longer distances. It’s a no-brainer. Consumers can get a better tv deal from cable or satellite. RBOCs should run EFM to 802.16/20 towers first. Do FTTH later.
The number of local telephone lines has been dropping steadily since 2001, according to In-Stat/MDR. In the first half of 2001, local phone line subscribers dropped by 2 percent. In the first half of 2002, the Bells lost 4 percent of their local lines. The main reason: people are replacing their landline with cell phones.
Twisted pair can’t compete with satellite - ever. It’s too expensive to push a dedicated video stream over twisted-pair. It’s distance-limited. It requires a truck roll. Competing in a saturated urban market is a receipe for disaster.
VDSL, like wireless, is just a tool. It’s not salvation. RBOCs, without careful planning, could implode like AT&T.
Poof.
.gif)
