Lucent Technologies is eyeing the use of cable-based AM
fiber and data technology as a possible option for a new product line aimed at meeting
rising demand for fiber-to-the-home solutions in the telephone industry.
Lucent officials refused to discuss what one researcher
referred to as the "Quantum Leap" FTTH project, or to say when a product might
be introduced. But it was clear that after years of pursuing a multitude of ideas as part
of the normal course of work within the manufacturer's Bell Laboratories unit, the
systems group responsible for bringing product to market has become much more focused on
the opportunities for an all-optical broadband solution.
"The project is still in research, and we just
don't have the ability to talk about it at this time," said a Lucent spokesman.
"However, we will at some point in the future, although I can't give you a time
line."
A variety of ideas for supplying a low-cost means of
extending fiber all the way to the end-user's premises are under consideration,
sources said, but none seems to have a better cost-benefit profile than an idea recently
tested by Bell Labs researchers. This involves the use of AM fiber technology in
conjunction with optical amplification and dual wavelengths to provide bidirectional
service over individual fiber drop links.
"In our test, we're providing the same type of
service that an HFC [hybrid fiber-coaxial] network would provide, but we're doing
away with the corrosive and ingress-prone coaxial drop," said Tom Wood, a
distinguished member of the technical staff at Bell Labs.
While researchers, who only recently took on testing the
HFC-derived FTTH concept, have not had time to do in-depth cost analysis, the approach
appears to be "very attractive economically, since we're delivering analog and
broadcast digital video, as well as data services," Wood said. "If you compare
this approach with one that just uses digital transport, I think that the revenue you
could expect from being able to deliver mainstream TV services is higher."
This approach to FTTH has the advantage of using
off-the-shelf gear designed for HFC networks in conjunction with erbium-doped fiber
amplifiers, the costs of which have plummeted over the past two years, Wood said.
"The big thing that needs to be developed to support
this concept is the ONU," he said, referring to the optical network unit that
connects the fiber at the subscriber premises. Wood noted that Lucent would also have to
create EDFA housings that are rugged enough for field mounting in distribution plant.
The lab test performed by Wood's group split the
780-megahertz output of an optically amplified, 1550-nanometer, externally modulated
distributed-feedback laser over four 13-kilometer fiber runs to secondary EDFAs, where the
signals were amplified and split again in 1-by-16 array, resulting in service to the
equivalent of 64 ONUs. Wood said future tests would aim at achieving sufficient splitting
to assure that about 500 ONUs are served for each high-power transmitter, which is
analogous to the HFC fiber-to-the-node ratio of 500 homes per node.
A key part of the test was the use of cable modems in
conjunction with in-home coaxial wiring to handle data signals, thereby overcoming one of
the primary problems associated with other all-optical approaches. The cable modem -- by
handling media-access control and other functions within the dedicated 6-MHz downstream
and 1.6-MHz upstream data channels -- eliminates the need to design such functions into
the ONU, the researchers said.
Upstream signals are fed via uncooled, low-cost Fabry-Perot
lasers operating in TDMA (time-division multiple access) mode in the 1310-nm wavelength
window through the fiber drop and 1-by-16 combiner, each using separate time slots to
avoid interference. The upstream signals are fed from the point beyond the combiner back
to the headend via separate fibers, bypassing the EDFAs.
Whether by coincidence or otherwise, the Bell Labs test is
a very close replication of the concept proposed by Synchronous Communications and Earl
Langenberg, vice president of network implementation for Tele-Communications Inc., at the
Society of Cable Telecommunications Engineers' Emerging Technologies Conference in
late January.
The idea -- first reported in May, when Langenberg was
working as a consultant with Synchronous -- makes use of off-the-shelf components now
available from the manufacturer, including strand-mountable EDFAs, and it supports a path
to bandwidth expansion using multiple wavelengths in the 1550-nm window.
Langenberg estimated that the two-wavelength version of the
FTTH system would cost $636 per home passed, or $909 per customer, based on 70 percent
penetration, versus average costs of $372 and $531, respectively, for HFC. HFC costs can
be much higher, depending on levels of fiber penetration, amplifier capacity and powering
requirements, he noted.
FTTC costs vary greatly according to design, but they
should cost out at close to or in excess of the FTTH costs, he added.
Wood said FTTC systems such as what Lucent is supplying to
Bell Atlantic Corp. and SBC Communications Inc. might be "a little cheaper" on a
first-cost basis. But, he added, the real advantage in cost goes to FTTH over time, due to
the lower power-cost requirements and the fact that fiber eliminates the corrosion costs
incurred in the use of copper lines over time.
Wood said the use of dense-wavelength-division multiplexing
was "a key question" among researchers, but the two-wavelength version tested by
Bell Labs would probably be the right approach for initial deployments.
"This system allows you to economically deploy fiber
all the way to the home, and it gives you a network architecture that you can evolve using
DWDM in the future," he said.
Researchers made it clear that the HFC-derived concept was
not the only idea being considered for an FTTH product from Lucent. Another attractive
concept involves the use of Bell Labs-developed technology to vastly reduce the potential
costs of a DWDM FTTH system.
As described by Joe Ford, a researcher at Bell Labs, the
idea employs a MARS (mechanical antireflective switch) reflector, which is a reflective
surface deposited just above the substrate of a chip that can be made to move up or down
with a small shift in current, thereby passing through or reflecting back to the source a
lightwave sent to the ONU. Used in conjunction with a modulator, this component can be
made to send signals back to the central office over the same wavelength used for
downstream signaling, eliminating the need for a costly, wavelength-specific laser in the
ONU.
Because the MARS reflector can be switched on and off in
microsecond intervals using time-division technology to send signals upstream when the
downstream signal isn't on, the two-way communication system is fast enough to
support voice and other isochronous interactive connections, Ford said. Using a tunable
laser at the source, the system has been shown to support upstream signaling at 1 megabit
per second, and it "can go even faster, at least up to 4 mbps," Ford said.
The MARS component has been rigorously tested and shown to
be "extremely rugged," Ford said. "I feel that MARS has a strong chance of
being used in future systems," he added.
For now, though, it may be that the push that Lucent is
seeing from some carriers for FTTH solutions merits a less exotic approach, where DWDM
isn't the primary requirement for initial installations. SBC and BellSouth are among
the telcos that are said to be looking at the possibilities for newbuild situations such
as high-income housing projects, where full-service-network capabilities are a priority.
"There certainly are some operators who are very
serious about FTTH," Wood said. "These days, it's very important within
Lucent to move on ideas that might have merit for future products."
