|
|
|
|
In 2000,
the telecommunications boom was at its peak. Companies rushed in
headlong to make good of the opportunity. By one count, some 80 new
telecom companies were set up across the world. Among them was the
Bangalore-based Tejas Networks.
The next year, the boom imploded spectacularly. Many firms, old and
new, went under. Analysts estimated that the crash wiped out some
$1,000 billion of wealth globally.
Tejas survived the crash and, appreciably, went on to thrive in a
small niche while many of its peers were still reeling from the
aftershocks. And it did so by
building an innovative product in India.
True, the company has the advantage of being guided by renowned
telecom entrepreneur Gururaj Deshpande, who is the largest
shareholder and chairman of Tejas. True, even after five years of
operations, the company has just about breached the Rs 50-crore
mark in sales. But none of that takes away from its credit of
having built an innovative product from ground up, that too in
record time.
The product
in question is a piece of transmission equipment - a box that
connects users like large enterprises to a telecom service
provider's network using fibre optic cables. Tejas has already put
its boxes at the core of some of the largest telecom networks in
India - those of the Railways, BSNL, and the Tata group. Last year,
it broke into the global market by becoming an original equipment
manufacturer (OEM) to a large global vendor (which enjoys a market
capitalisation of $10 billion).
The whole effort has grabbed the
attention of one of the best-known telecom investors in Silicon
Valley, Battery Ventures. In January, it led a group of investors
to plough $15 million into Tejas, in what turned out to be the
second largest venture capital investment in India during that
financial year.
It's hard to imagine such a feat from a company where two of the
three founders, CEO Sanjay Nayak and head of engineering Arnab Roy,
didn't even come from the telecom industry. And they hadn't even
met the third founder, chief technology officer Kumar N. Sivarajan,
before forming the company. The credit for putting the team together
goes to Deshpande.
Setting Off
Tejas's journey started one night in March 2000, when Sivarajan,
who was then assistant professor at the Indian Institute of Science
in Bangalore, received a call from Deshpande's representatives.
Would he meet Deshpande?
Sivarajan, a PhD from CalTech whose book, Optical Networks: A Practical Perspective,
is a prescribed text at several universities, didn't want to leave
his academic career for starting a company in the US. But since he
was travelling to the US soon afterwards for a conference, he
agreed to meet Deshpande.
The meeting changed Sivarajan's mind, especially when he heard
about the opportunity in fibre optics, among the hottest
technologies back in 2000, and the fact that Deshpande wanted the
start-up to be based in India. Ten days later, he flew back with a
letter appointing him as the CTO and employee No. 1 of the
yet-unnamed start-up.
|
|
Tejas chairman Gururaj
Deshpande was clear that his company had to win the Indian
market first
|
|
|
The search for the CEO took
Deshpande to the then CEO of Synopsys India, Sanjay Nayak. By then,
Nayak had set up several subsidiaries for high-tech multinationals
like Cadence, ViewLogic and Synopsys. Time was on Deshpande's side
- Nayak and colleague Roy (they had worked together since 1989)
were already thinking of starting out on their own. Deshpande
agreed to take both aboard.
Tejas was incorporated on 25 April 2000 with Deshpande as chairman.
The initial $5 million came from Deshpande and his other start-up,
Sycamore Networks. The founders spent the first few weeks
brainstorming in Boston, where Sycamore was based. The question
was: what should Tejas build?
Deshpande was sure that Tejas had to start with the domestic
market. Chinese companies like Huawei, ZTE and UT Star - who
together command 7 per cent of the global telecom equipment market
- had shown that global scales could be attained even while
catering to the domestic market, especially large ones like China's
or India's.
But what would be the product? While still searching for the
answer, the Tejas team decided to sell Sycamore gear in India. The
Indian telecom boom was just starting, and Sycamore, which was
selling optical networking gear capable of transmitting terabits of
data in a few seconds, wanted a piece of that action.
While selling Sycamore ware, the Tejas team got a view of what
network providers were looking for. A few equipment vendors tried
to cash in on the fibre optics opportunity by adapting the earlier
core networking gear (used in copper-wire networks) for use in the
new optical networks. That proved to be too expensive for
customers. Issues arose even with the Sycamore gear. The slowest
traffic that the Sycamore product could accept was 155 megabits per
second (Mbps). Indian telecom providers, who had just started
building their networks, wanted equipment that could go down to
capacities of 2 Mbps and 34 Mbps, which wasn't readily available.
Tejas had found its sweet spot. It decided to build access
equipment that would provide speeds starting at 2 Mbps, going up to
155 Mbps. The strategy made sense because, by complementing the
Sycamore gear, Tejas could sell wherever its parent was selling.
The company started developing its first product, the TJ100, in
October 2000.
|
|
Chief technology officer
Kumar Sivarajan, a former academic who has co-authored a
widely-prescribed textbook on optical networking, was Tejas’s
employee no. 1
|
|
|
Within a few months, Tejas found a
customer in Tata Power, which was looking at a broadband foray in
Mumbai. Tejas had to come up with its products by June 2001.
Building a product in such a short time was a huge challenge. The
Tejas team had several software heavyweights conversant in the
complexities of electronic design automation, but it lacked
experience in designing a system that included hardware and
software.
Getting the software architecture right was critical. Since the
board (hardware) design was also being done from scratch, there
would not be time enough for running and debugging the software
after the board was ready. This is where the strong background in
design automation helped. The team built an emulator of the entire
box on a workstation, and tested the application software - about
80 per cent of it - even before it was put on the hardware.
Tejas brought out the TJ100 within the deadline. It sold the
product to a few more clients, including China Netcom. But before
it could scale up, the market changed again.
By late 2001, telecom spending had started dipping globally.
Sycamore pulled the plug on its optical transmission products on
which Tejas was riding. Without that connection, selling globally
was tough. The Indian market was opening up, but it expected a lot
of performance at very low costs. Without Sycamore, Tejas needed a
stronger product proposition.
Turning Point
A transmission product has two ends: one facing the user and the
other facing the network. The old networks were designed to carry
just voice. Data traffic (most commonly, our emails and Internet
downloads) grew only in the past decade. Therefore, the
transmission gear of that time accepted only a voice-like interface
at both the ends. Enterprises, the predominant users of data
services, could not connect a data (or Ethernet) cable directly to
the transmission equipment. Instead, they had to use a router to
receive the data traffic, convert it into a voice-like interface,
and then put it on the transmission equipment.
Also, the capacity of the transmission equipment was limited. All
that the network provider gave were fixed bandwidths of 2 Mbps and
34 Mbps. If an enterprise wanted 8 Mbps, it had to take 4 pipes of
2 Mbps each, for which the router maker had to create additional
interfaces. Or it had to take a 34 Mbps cable, and use only 8 Mbps.
This was a major bottleneck in expanding bandwidth.
So the proposition for Tejas was to build a transmission box that
could directly accept data traffic and also provide flexible
bandwidth.
The question was what technology
Tejas would pick - voice and data services use different
technologies and follow different standards. While voice traffic
works on time division multiplexing (which runs on the synchronous
digital hierarchy standard, or SDH, the North American equivalent
of which is Sonet), data runs on IP (Internet protocol). Since
Tejas was building something that would mean a convergence of the
two, it was important to use the right standard.
Fortunately for Tejas, in 2001, the International Telecom Union
(ITU), the organisation that sets the standards for telecom
service, announced a standard for carrying data over SDH/Sonet.
Tejas chose the ITU standard. But once again, it had to move fast
because once a standard is announced, every equipment maker would
start launching products on the standard.
There were other issues, too. Since the standard had just been announced,
there was no commercial silicon available for designing a box
running Ethernet over SDH/Sonet. In a smart move, Tejas decided to
develop its own silicon. Not only did it beat others to the market,
in-house development helped the company keep the costs low.
January 2003 saw the launch of the
TJ100 MC1 (featured above), a single box that could accept an
Ethernet connection and provide flexible bandwidth up to 100 Mbps.
Subsequent versions of it can now take bandwidth up to 2 Gbps. By
putting Ethernet directly on the transmission equipment, MC1 did
away with the need to put a router in that place.
Within a year of the new launch,
Tejas trebled its sales. By the end of March 2004, the company had
done about Rs 30 crore in turnover and had sold the MC1 series of
products to 25 companies. Though the latest financials haven't been
audited yet, the company expects the turnover to be close to double
the previous year's.
Going Global
The optical networking market has gone through a roller-coaster in
the last decade. At the moment, the demand for optical access
equipment is growing as telecom carriers are redesigning networks
to deliver IP-based voice, data and video services. According to
Heavy Reading, a telecom market research agency, 2005 promises to
be a turning point for optical networking.
While this represents an opportunity, it also poses a challenge.
Tejas still gets 80 per cent of its revenues from the domestic
market, which is all of $250 million in size. Compared to this, the
global market for Tejas's products is worth more than $5 billion.
Apart from large companies like Lucent, Nortel and Alcatel, this
global market is served by players like Huawei, ZTE and UT Star.
And Deshpande estimates that the Chinese companies are going to
corner a quarter of the entire telecom equipment market in a few
years.
He says: "Tejas needs to grow 100-fold to get the critical
mass to compete globally." That's why the OEM route is a good
way for Tejas to grow. Though it earns lower margins, it ensures
that the company doesn't have to spend large chunks of its earnings
in marketing abroad. It would also help Tejas achieve its target of
getting 40 per cent of revenues from global sales in two years.
The other challenge is to beat technological obsolescence.
"Most products [in this market] become obsolete in about three
years," says Deshpande. That puts tremendous pressure on the
team for getting more innovative products to the market. (The
effort continues apace. Tejas claims that this year it will be
among the first companies in the world to launch Switched SDH, an
equipment that allows even greater flexibility in allocating
bandwidth.)
Deshpande sees an opportunity where others may see a pressure
point. He says: "Companies have to constantly invest in
R&D. During the bust years, many telecom giants haven't been
able to invest in R&D. By tying up with a company like Tejas,
they can offer the latest products." In effect, Tejas looks
set to stay on the OEM route for a while.
Agrees Carl Stjernfeldt, partner at Battery Ventures: "I view
this as a great opportunity for Tejas to grow and become a world
class player." If Tejas keeps growing, it will be an
achievement for Indian manufacturing as a whole.
|
