Corrections or additions?
These articles by Barbara Fox and Kathleen McGinn Spring were
prepared for the May 9, 2001 edition of U.S. 1 Newspaper. All rights
NJ: Hot Telecom Territory
From the time that Ma Bell presided with matriarchal
splendor in the telecommunications industry, New Jersey has been a
leader in telecom technology. The telephone system worldwide is the
most complicated machine in the world and most of those systems have
been engineered right here. Even after deregulation — perhaps
especially after deregulation — scientists here are making
exciting discoveries in the new research areas, and not all of these
work is being done at AT&T or Lucent Technologies or even Sarnoff.
Many small, exciting companies have sprung up.
"As a result of AT&T and Lucent and Sarnoff, this area is a hotbed
for optical device companies," says Joe Montemarano, industrial
liaison for the Princeton University Center for Photonics and
Materials (POEM). "The people who actually make fiber
tend to be a few very large companies like Corning and Lucent, but
for every big company there are half dozen little companies all
on a slightly different piece of the action."
The profit curve can only go up, because fiber optics is such a very
young industry, says Mark Itzler, chief technology officer of JDS
Uniphase’s Epitaxx division. He takes this view even though his own
company had to trim its workforce in West Trenton. "Five years
ago," says Itzler, "there was negligible revenue in this area,
comparable to what it is now, and what it will be."
Whether they are making chips, or switches, or converters, or
or modulators, or measuring devices, or repair kits, these mostly
new, mostly small companies are coming up with next-generation fiber
optic components. In addition to spin-out support from the Lucents
and the Sarnoffs, they get lots of R&D support from Princeton
Unlike Rutgers, which focuses on basic research in fiber optics,
researchers work on devices for optical signal switching (such as
modulators) and for signal detection (such as receivers).
"New Jersey is strong in telecommunications because of the AT&T
legacy. Now POEM, Sarnoff, and the abundance of other companies doing
the same kind of thing provide our technology base," says Roy
Howard, applications engineer at Discovery Semiconductors on
Road. "We all work on the same problem. Everybody looks at from
a different perspective and solves a piece of it."
"We’re talking about enabling broadband," says Tom Lento of
Sarnoff. "All these companies are working in areas that will
the network to carry far more data than it ever has been able to do
First, let’s define terms, starting with fiber, the term often
used by real estate brokers to lure companies to certain office parks
("Great fiber! Redundant loops!") by promising them fast Internet
connections. Data and voice message travel faster on fiber, because
it travels at the speed of light, as a light signal, in packets of
1s (where the light is bright) and 0s (where the light is dim). On
copper wire, the messages are sent with an electric current.
Scientists use the term fiber optics to describe the entire
field of carrying light with fiber and all the devices needed to
The term photonics is also about the technology of manipulating
light, but with a different connotation. If "electronics"
works with electrons to produce electrical impulses,
works with light to produce light waves.
Telephone companies are gradually replacing their copper wire with
fiber optic lines, but even messages sent on fiber must start as an
electrical impulse (in your handset or computer) and end as an
impulse (in the receiver’s handset).
Converting electrons to photons and back again requires
and the conversion is done with lasers. Greg Olsen founded Epitaxx
(now JDS Uniphase) as Epitaxx Optoelectronics, for instance. Vladimir
Ban’s PD-LD and Olsen’s new company, Sensors Unlimited, for instance,
are doing the laser conversions.
What about wireless? Won’t that replace everything? No, because at
some point the wireless signal must be converted to some kind of wired
All this switching from wireless, to fiber, to copper wires, and back
again requires an endless assortment of switching and modulating and
monitoring devices. Princeton’s small telecommunications firms are
busily working on them.
River Road, speeds up the switching connections that take place within
the telephone companies’ offices.
Route 130, has laser and chip products that pack extra power to
bandwidth for both long and medium distances.
optoelectronics with standard technology — CMOS semiconductor
chips made of silicon to send signals at optical speeds.
to its own 33,000-foot building on Black Horse Road in North
wants to put many active and passive devices on the same chip, making
the optical equivalent of an integrated circuit.
operation in West Trenton, has fiber optic receivers (to detect the
data stream from any optic channel and convert it from light to
and optical monitors (to regulate the flow of light on the network).
other end of the network. Its current cash cow is packaging lasers
to convert electrical signals into light or optic streams, but it
is doing leading edge research for future work in networks with
glass (see page 15).
packs photo detectors into infrared cameras and has a new kind of
modulator to monitor the wave lengths of light on the optic networks
known as "dense wavelength division multiplexing systems
those with different length signals on the same fiber.
points out, that until a couple of months ago nobody had been through
a downturn. He is sanguine about the current company layoffs, a wave
that took 200 of the 1,000 workers at his West Trenton manufacturing
facilities, and compares it to a soft depression in what has been
until now an upward slope. "This is a sign that the industry is
maturing, and we have a great deal of optimism going into the
Unveiled in March as a venture from Lucent Technologies,
Aralight has a 75-patent portfolio from Bell Labs and a goodly number
of Lucent’s alumni. The company has switching devices that work at
50 bits per second but aims to create devices for "terabit"
systems that can transmit a trillion bits per second while using just
one-tenth of the power that today’s much slower optoelectronic systems
Right now, it targets switches that are located in telephone company
central offices. "We have demonstrated aggregate band widths of
50 bits per second," says Gary Ger, vice president for business
development at Aralight. "This technology is now being used for
switching, moving information from one box to another, but it can
and will be used for transmission applications."
Unlike the much smaller Sarnoff, Lucent makes it a standard practice
to populate new ventures with the same scientists that were working
on that technology in the lab, says Steve Socolof, vice president
of Lucent Technologies New Ventures Group. He was a consultant at
Booz Allen before joining AT&T and the Lucent ventures group, and
he has degrees in math, computer science, and economics from Stanford,
Class of 1982, and an MBA from Dartmouth’s Tuck. His group was the
leader in a $10 million first round of funding, which was joined by
Ridgewood Capital, Signal Lake Ventures, and Solar Venture Partners.
Socolof explains the Lucent strategy. Five years ago, when Lucent
became an independent company, it began to create new ways to bring
technologies to market. "The success of a business based on
technology has something to do with the commitment of people seeing
it through to product delivery," says Socolof. "Instead of
trying to retain them we actually encourage them to go with the
We want to work with an entrepreneurial team. Our goal is not
to flow compensation back into the labs."
"Conversely, if we see an opportunity to create a venture and
we find a set of people who like being in the labs, that is
a deterrent for us to move forward." Yes, this policy does incur
the cost of losing good researchers. "But that kind of culture
and movement and replenishment of people is a good thing, culturally,
for the lab. And out of 20,000 people we are not talking about a lot
Three of Aralight’s co-founders are indeed Lucent alumni: Ashok
Keith Goossen, and Ger. Krishnamoorthy, the chief technology officer,
is an electrical engineer from the California Institute of Technology,
and has graduate degrees from the University of Southern California
and the University of California, San Diego. Goossen, the chief
is an electrical engineer from the University of California at Santa
Barbara with a PhD from Princeton. Ger is an electrical engineer from
Drexel who has an MBA from Northwestern’s Kellogg.
Michael Camp, former CEO of Aralight, has moved on but retains a board
Socolof also tells how Lucent has tried to give its ventures a good
start with a good patent position. Aralight, for instance, has patents
in such areas as microbump flip-chip bonded VCSEL design, array
device design, high-speed transceiver arrays, III-V semiconductor
processing, 850nm photodiode design, and VLSI design of low-power
receiver/driver and switching circuits. "People trust what comes
out of Bell Labs is good stuff," he says.
"Besides technical credibility," he says, "it is nice to be
friends with Lucent. It has been demonstrated time and time again
that some sense of the Lucent name might open doors for partners or
But investors need to know that the parent will continue to take a
benevolent attitude and not set up future competition with its
"Lucent is a big company with strategic intent of its own, but
we have spent years educating the market to ensure that, for Lucent
spinoffs, the Lucent influence is aligned with any outside
Lucent typically does not transfer patents but provides
a broad background license with exclusivity in a particular field
of use, to affirm that Lucent wants this business to grow and prosper.
"People will never invest in a young company unless they believe
it has every thing it needs," says Socolof.
As a huge company that comes up with four patents a day, Lucent has
two kinds of discoveries that it might want to license to young
don’t fit in with what Lucent is doing.
of Lucent but may not have exactly the right home in Lucent.
and very short range applications, a particular market segment that
none of Lucent’s business was focused on. "We thought this would
do better outside, where the company can talk with the various
who are on the leading edge," says Socolof.
The successful companies will be protocol independent. Optical, DSL,
ATM, frame relay, IP — all need different languages, Socolof
out. "Each is appropriate for certain parts of the communication
stream," says Socolof, "and with networks of networks, the
networks have to be able to talk with each other. Lucent is a good
vendor because we are able to work with all of them."
"Our challenge is to make Aralight a high volume manufacturing
company," says Ger. "We have tremendous opportunities in
and massive parallel processing, but as a start-up, you have to focus
on a niche and execute properly. Down the line, as we generate
we will enter other gigantic markets."
3, Monroe Township 08831. Catherine Kopko, office manager.
fax, 609-409-3335. Home page: www.aralight.com
Princeton Lightwave Inc. is a start-up with an unusually
valuable portfolio — the exclusive rights to the Sarnoff
key optoelectronic intellectual property for telecommunications. The
company also walked out of Sarnoff’s door on Fisher Place with an
unusual number of personnel. Sarnoff’s usual method for spinning off
companies is to give its researchers an equity stake in the new firm
but keep them at home working on more new projects.
"PLI had products from day one," explains Jim Keszenheimer,
Sarnoff’s director of business development for optoelectronics.
was felt that the best way to maximize the company’s value was to
maintain continuity in personnel."
"I picked the group of people who would build the best foundation
for the new organization without hurting Sarnoff," says John
vice president of engineering. A graduate of Rutgers, Class of 1977,
he has a master’s in metallurgical engineering from Polytechnic
of New York. At Sarnoff he had been in charge of the laboratory that
commercialized not only telecommunications applications but also those
for medical diagnostics, power generation, and electronic sensing.
Other Sarnoff alumni here are Ray Menna, a crystal expert who is
of materials growth, and Dmitri Garbuzov, vice president of research;
he had done research at Princeton and Northwestern universities and
in the Soviet Union and is a winner of the Humboldt Award.
Two women left Sarnoff for Princeton Lightwave. Pamela York, vice
president of corporate development, was a pioneer in several kinds
of lasers and also led the effort to create a silicon and glass
chip for drug discovery and genomics. She helped build one of the
successful the Sarnoff spinoffs, Orchid BioSciences.
The director of new product introduction is Nancy Morris, who worked
at Laser Diode Inc. before coming to Sarnoff, where she led the
transfer and product commercialization for one of Sarnoff’s venture
companies, Secure Products.
Greg Blonder, of Morgenthaler Ventures, moved in to be the acting
CEO after the departure of the chosen CEO, who had been plucked from
Nortel. Blonder has been chief technical advisor to AT&T and director
of materials and technology integration research at Bell Laboratories,
and he led Morganthaler’s investment in another Princeton firm,
Also in last year’s $28 million first round of investment are Venrock
Associates and U.S. Venture Partners.
Princeton Lightwave uses its unique chip level design and integration
technology to come up with high-performance optical components for
advanced network applications. By integrating design with
Princeton Lightwave hopes to create the materials and structures for
the next generation of optical subsystems — such products as power
pump modules, advanced source lasers and Broadband GainChips (BGCs)
— with production slated for later this year.
Last month PLI announced a breakthrough, achieving the 1-watt power
level from a single narrow-stripe pump laser chip, taking the market
leader position in that particular kind of high speed laser pump.
Using an indium phosphide platform, these pumps can push a pulse for
longer distances, leading to wider bandwidth at lower cost.
Additional power does not increase the speed; it increases the
between amplifiers. "As light travels down the optical fiber the
signals weaken and need to be regenerated," Connelly explains.
"Each amplifier gives the signal strength. Lasers are used to
optically pump the fiber, which creates a signal for the
PLI has amplifiers for both traditional transmission methods and the
more advanced Raman amplifiers, which allow for even broader, less
Connelly feels the Sarnoff alumni gleaned plenty of start-up knowledge
from working with early stage companies: "When you are dealing
with senior management, you get to see all the warts, and it doesn’t
take you long to see which ones are running properly."
Princeton Lightwave also has access to the continual feed of new
technology from the team at Sarnoff that works on next generation
products. "We are continuing to do development work on PLI’s
says Sarnoff’s Keszenheimer.
— Barbara Fox
Cranbury 08512. Greg Blonder, acting CEO. 609-925-8100. Home page:
Velio Communications Inc. opened a design center in
the Princeton area for one reason, and one reason only: "Purely
to tap the manpower," says G. Ramamurty, chief system architect.
The semiconductor company, with headquarters in San Jose, California,
designs minute components for next generation optical telecom switches
Unlike all the other companies in this lineup, Velio does
with standard materials. It uses the more traditional technology —
standard CMOS semiconductor chips made of silicon — to send
at optical speeds. "It is difficult to get CMOS running at optical
speeds, so people are going to other materials like indium phosphide,
but that is expensive and is not a mature technology. We can transmit
and switch at optical speeds using standard CMOS technology. There
is no risk," Ramamurty says. Velio’s switch fabric is being used
now in the backbone of optical networks.
The 120-person company is not alone in hunting for engineering talent
in New Jersey. "All the big California companies have started
design centers here," says Ramamurty, reeling off Motorola, Intel,
and Broadcom as examples. "Everyone has a presence here."
Velio moved into 10,000 square feet in 210 Carnegie Center in the
fall. These offices put the company near the carrier class equipment
manufacturers that are its customers. And while Ramamurty would not
name clients, its targets are companies like Nortel and Lucent. A
sales presence is not in the picture for Velio, however. All of the
company’s 12 Princeton employees are engineers, six of them Ph.D.s.
Ramamurty says the company expects to employ 40 to 60 engineers in
its new offices.
Velio is a start-up that has taken in $50 million of venture funding
from Sequoia Capital, IVP, Global Catalyst Partners, Capital Research
Group, and HarbourVest Partners. In addition to its new Princeton
offices, the company has design centers in Lowell, Massachusetts,
and Chapel Hill, North Carolina.
The telecommunications industry is now in a slump, but Ramamurty says
his company is not feeling the effects. Next generation communications
technology like that on which it is working is a necessity, he says,
and Wall Street is looking at products telecommunications companies
have in their pipelines to meet the need. "Traffic is
he says, using a highway analogy. "One way to cope is to make
roads broader, that is the way optical systems work." But
capacity is not enough, as any shore-bound traveler encountering a
toll booth on the Parkway on a summer Friday knows. "You need
efficient interchanges," he says.
Communication circuits get flooded with data, Ramamurty says, "and
you have to switch intelligently." The chip elements Velio
called "blocks," move massive amounts of data between circuit
cards on a backplane, the surface on which the cards are mounted.
"In each generation, we pack more and more," on the chips,
he says. "What occupied an entire cabinet, we now put on a single
chip." As the chips acquire more processing power, however, they
run faster and faster. This becomes a problem, because "they’re
not like autos that let all the emissions out." Velio addresses
the power issue by designing very low power consumption into its
The company is a "fabless" shop, meaning that it designs,
but does not manufacture, or fabricate. "If you take companies
like the old Lucent, they used to design and manufacture,"
says. "Today that (manufacturing) in itself is a billion dollar
industry. It’s too expensive for everyone one to do their own."
Velio outsources manufacturing to companies that "have no idea
what we’re building." The company, which started to sell its first
family of products earlier this year, could file for an IPO in 2002,
Ramamurty left NEC, where he had worked for nine years, to open
Princeton office. Asked why he cast his lot with the start-up, he
says, "for the same reason everyone else does — it’s exciting,
and the upside is high." After reflecting a moment, he adds,
a chance to build something you always wanted to."
The son of a chemist and a homemaker, Ramamurty was born in India
and received his master’s degree in electronic communications
from the Indian Institute of Science in 1974. He holds a Ph.D. in
computer communications from the University of Aston in the United
Kingdom. In addition to NEC, he has worked for Indian Telephone
and for Bell Laboratories in Holmdel. His wife, Hema, a librarian,
worked at the West Windsor library until he started working for Velio,
and putting in longer hours. She now stays home with their two
Despite the fact that he is heading up the growing, new office of
a technology start-up, Ramamurty says he works 50 to 60 hours a week,
a schedule that would be considered part time at some pre-IPO
"The important thing," he says, "is to work smart."
— Kathleen McGinn Spring
Floor, Princeton 08540. G. Ramamurty, chief design architect.
fax, 609-419-9777. Home page: www.velio.com
Corrections or additions?
This page is published by PrincetonInfo.com
— the web site for U.S. 1 Newspaper in Princeton, New Jersey.