Cable fiber-optic applications could take a quantum leap
forward if either or both of two product-development strategies brought to light last week
by two very different suppliers live up to expectations.
In one approach to fostering a more flexible use of fiber,
Scientific-Atlanta Inc. said it was investing $13 million in Bookham Technology Ltd. to
promote development of a new generation of optoelectronic devices that could drive down
the cost of extending fiber deeper into cable networks. Bookham has developed a means of
forming complex optical circuits on mass-produced silicon chips. This has the potential to
miniaturize and reduce the costs of products to be used in dense-wavelength-division
multiplexing, return path transmissions and fiber-optic nodes.
The other development, involving the new optical
transmission technology developed by startup SilkRoad Inc., is meant to greatly reduce the
costs and increase the performance of fiber supertrunks used to transmit analog and
digital signals among headends and primary distribution hubs.
SilkRoad, which recently conducted the first of several
scheduled field tests of its technology in non-CATV applications, said its initial line of
products slated for rollout in the first quarter of next year would include a system for
cable that will bidirectionally transmit multiple feeds of analog and digital TV as well
as telecommunications signals. This would occur in their native formats, on a single
wavelength at distances well beyond the unamplified transmission capabilities of today's
AM (amplitude-modulated) supertrunks.
The most immediate impact from the developments at these
firms could come from S-A, which is already testing some Bookham products in hopes of
quickly converting them for use in cable. "If these pan out, we could have products
available within six to eight months," said Paul Connolly, vice president of
marketing and network architectures at S-A.
"In instances where we have to design from the ground
up for our applications, we're probably looking at a 12-18 month timeframe," he
Bookham's patented technique involves the alignment of
passive optical components onto silicon chips while using standard silicon-chip production
techniques to produce many devices at low costs. The resulting miniaturization of the
complex circuitry involved in the interface between electronic devices and optical
receivers, waveguides and other components plays well with the needs of architectures such
as the "Lightwire" system AT&T Broadband & Internet Services is
preparing to test in San Diego, Connolly noted.
"What you end up with is the equivalent of ASICs
(application-specific integrated circuits) in electronics, where you can take standard
cells and cut devices in mass quantities to meet specific industry needs," Connolly
said. "We believe they're unique in the world in being able to do this on
Bookham, based in Oxford, England, has already taken on
Intel Corp. and Cisco Systems Inc. as equity partners for development of components in the
computing and routing sectors. S-A, as the third industry-specific equity partner, will
help guide the development of ASOPs for cable, Connolly said, declining to discuss other
terms of the agreement.
One of the applications S-A is exploring with Bookham would
be able to integrate a wavelength demultiplexer and photoreceiver onto the chip, thereby
creating a miniaturized, low-cost means of handing off a wavelength from a multiwavelength
stream at a micronode on the coaxial plant. This way a single fiber could serve multiple
micronodes with dedicated wavelengths, vastly lowering the cost of extending fiber deep
into the cable plant, as AT&T is doing in its Lightwire trial.
The technology could also be used to take in multiple
wavelengths and translate them to separate RF feeds, Connolly noted. "For example,
you could have one chip that demuxes and receives eight wavelength channels and converts
them to electronic signals," he said.
"The technology has implications for increasing the
density of wavelengths, for return path transmissions, where the transmitter and receiver
operate on the same chip and for many other applications," Connolly added. For
example, he said, one of the tests of currently available product from Bookham that could
lead to a quick product turnaround involves a low chirp, externally modulated
1550-nanometer laser, which, if it can be QAM (quadrature-amplitude modulation) modulated,
would produce the transmission distance benefits of external modulation at direct
S-A asaid it had also reached an agreement to acquire the
photonic business units of U.K.-based Thomas & Betts, which has been supplying
components used in S-A's four-port node return path system. The good density
characteristics of these components allow return signals to be multiplexed digitally from
different coax return feeds for return transmission over a single fiber.
At SilkRoad, the technical breakthrough has to do with the
ability to deliver very high volumes of traffic over a single wavelength, which could
prove to be a cheaper option for maximizing fiber capacity than DWDM, if the company's
claims prove out.
While SilkRoad eventually expects to offer gear that would
facilitate fiber applications in any part of any network, its initial CATV product is
focused on supertrunking, starting with an optical transceiver (transmitter/receiver)
product line to be made available commercially in the first quarter of next year that will
support delivery of four, eight or 16 separate bands of information on the same
wavelength. The company said each band can deliver up to 1.5 gigahertz of RF TV signals or
digital electrical spectrum of up to 650 MHz or a Sonet OC-3 (155 megabits per second) or
OC-12 (620 mbps) signal, with any combination of band configurations in either direction.
If the system is used to carry analog TV signals, it can
transmit up to 100 MHz without amplification or about twice the distance of standard AM
fiber systems, officials said. Transmission distances are longer in all-digital mode.
The immediate push in the cable TV product-development
effort comes from a regional Bell company that wants to be able to combine its analog TV
channels with delivery of OC-12 and OC-48 (2.5 gigabits per second) signals over a single
fiber, said SilkRoad president and CEO Kevin Doria. "They asked us to build and
design a network for them, and we'll be testing this in the next few months," he
Doria said the company will also introduce a commercial
product line for the enterprise market in the first quarter, with various versions
supporting anywhere from four to 16 mixed-technology payloads consisting of OC-3 or -12 or
digital video in each path or, alternatively, fewer paths of OC-48 payloads, ranging from
one to four.
The company conducted a trial of the technology July 23
with an unnamed oil and gas company at San Diego State University. It delivered a
computer-to-computer OC-12 payload over a distance of 80 kilometers directly from and to
OC-12 cards installed in Silicon Graphics computers, without use of intervening
multiplexing or switching equipment, Doria said.
"Our next test will be over a Qwest (Communications)
link between Los Angeles and San Diego," Doria added. This test, also operating at
OC-12 rates, will cover a distance of 200 Km. for delivery of a 300 gigabyte file, he
The commercial digital communications products will be
designed to propagate a distance of 500 kilometers without amplification, Doria noted.
While the capacity levels envisioned for these initial products don't approach the limits
of DWDM systems on the commercial market, the SilkRoad systems will achieve what DWDM
systems do at much lower cost. This is because they only require use of a single laser and
a high-frequency external modulation system, officials said.
Moreover, they added, the technology eventually will
support products that approach the physical capacity limits of fiber, at somewhere around
10 terabits per second, which is currently well beyond the reach of DWDM.
While many observers have been skeptical of SilkRoad's
ability to make its physical principles work in real-world field applications, Doria and
his colleagues asserted the test last week and others planned over the months ahead
clearly show that the technology is practical. The company has a licensing and development
agreement with NASA under which the two entities are creating high-frequency detectors now
operating at 300 GHz and soon to move to 500 GHz and later to 1.5 teraHertz, they noted.
High-frequency modulation and demodulation components are
one key to the way SilkRoad's technology works. Another essential element in what the
company calls "refractive synchronization communication" is the use of multiple
properties of photons to carry information.
Along with achieving high speeds and long distances, the
company also claims its signal aggregation and modulation technique supports the adding
and dropping of signals in meshed arrays with use of simple beam splitters. This
capability, using tightly coordinated timing between send and end points to control which
signals are dropped off, avoids the complexities of all-optical add/drop multiplexers and
cross-connects that are just coming to market in the DWDM domain, officials said.
SilkRoad's system is being tested by a number of
long-distance and competitive local-exchange carriers, said Dataquest analyst Ken Kelly,
who recently completed a report on the SilkRoad technology. One company, Level 3
Communications, was willing to have its name used in the report, with unnamed officials
telling Kelly that while they have some reservations, they expect to use the technology
once it becomes commercially available.
"In fact," Kelly reported, "an executive at
Level 3 believes all or most carriers will eventually deploy it."
A Level 3 spokesman refused to comment on the company's
position. "We don't discuss vendors, even the ones we're contracted with, let alone
any we're conducting tests with," he said.
As for the CATV application of SilkRoad's technology, Kelly
said, "I saw the system demonstrated with TV signals, and it lives up to its
SilkRoad officials said they had not set pricing for the
new CATV supertrunking system, but indicated it would match the costs of AM systems and be
significantly cheaper than digital options.