Home
> Inperso > Full Story
"The
USP of our products is the built-in software intelligence"
By
Mahesh Rathod and Shubha Murthy
Bangalore-based
Tejas Networks has partnered with US-based Sycamore Networks
to build a complementary set of products using a common
software platform and network management systems. Kumar
N. Sivarajan, co-founder and chief technology officer,
Tejas Networks spoke to Network Magazine and delineated
the optical networking trends
Can
you give an overview of Optical networking?
Optical
networking makes use of fibre for transmission of signals
in the form of light. Fiber optic networking is more than
a decade old, when we had this whole SONET and SDH hierarchy,
which today goes up to over 10GB. The more recent development
i.e. in the past six to seven years is DWDM (Dense Wave
Division Multiplexing). Today optical networking has become
synonymous with DWDM the use of multiple frequencies or
wavelengths for different signals from the infrared light
portion of the optical spectrum. This has been the trend
on the hardware optic side.
What is the current trend?
Currently,
the trend is to add intelligence on these optic boxes, which
are used by carriers such as VSNL, Reliance, and Tata Power.
All along bandwidth was viewed as a long-term investment,
in the form of leased lines that could last up to 10 years.
So there was no need for any intelligence on these boxes
as there is in an IP router, for instance.
IP routers are supposed to deliver packets very fast, but
depending on the circuit, it can take its own time. Now,
there is a rapidly emerging market for shorter and shorter
length circuits. If some one wants bandwidth urgently for
a few hours we can't take a few days to deliver the same.
Motivated by such demands, optic boxes are also trying to
accommodate a lot of intelligent software that is traditionally
associated with IP routers.
What is the optic networking scene in India?
In
India, we have had SDH networks for a while now, which have
been offering speeds from 155 MB through 622 MB up to 2.5
GB. With newer carriers coming in, we will see DWDM deployment
by the end of this year. Tata Power has already deployed
Sycamore's DWDM for its backbone, and the access part of
the network is supplied by Tejas.
What is blocking optical network deployment in India?
The
national long distance market was deregulated around the
same time last year. Since this is an infrastructure product,
it is bound to take time. Laying optic fiber networks is
similar to laying national highways. It takes a lot of time
to dig up, lay the fiber optic cable and then the equipment.
The work is of the same magnitude; the operations involved
are huge. But I think it is proceeding at a fairly rapid
pace.
Only ducts are being laid and not the cables?
Why?
Though
there is no cable crunch, it is also not available in abundance
as it is being laid all over the world. Around this time
last year there was a crunch in optical fiber, though now
it has eased out and is steadily available. It doesn't take
very long to lay the cable, as there are sophisticated technologies
for doing this.
There are also commercial implications. A trend might set
in where a company might own the ducts and may allow some
other company to lay the cables.
What products is Tejas offering the optical networking
market in India?
Currently,
we have introduced the family of multi service access products
that sit between consumers of bandwidth and the backbone
of the network.
Since
we address the carrier market, the consumers of bandwidth
are very big who have leased lines at 2 Mbps or higher rates.
Our customers include ISPs, who have high speed leased lines
to connect their data centers and so on; cellular operators
and carriers. For instance, Tata Power has deployed DWDM
backbone equipment. The equipment, which connects the consumers
to this backbone network, is what we are manufacturing in
India.
These products are also available from companies such as
Alcatel. But what differentiates our products is that they
are state-of-the-art. Our product is only three and half
inches high and on one card we can accommodate all the interfaces
i.e. sixty-three E1s, three E3s and two 155 MB interfaces
connected to the network. This is the densest product in
this category.
Besides, we have put a lot of software intelligence into
this product. This is our USP. Therefore we believe that
the product is going to proliferate. The network is going
to have lots of these boxes. For instance, a client will
want bandwidth for the next hour, so we can automatically
provision.
On a leased line such provisioning is going to take a lot
of time.
The intelligence that we have built in to the boxes performs
three things:
-
When a new box is added into the network, the box is
capable of learning the topology of the other boxes
and the existing ones of this new box. This is called
automatic
topology discovery. This intelligence is not there in
most other telecom equipment but can be found on IP
routers. The same concept has been incorporated
in these boxes. These boxes have a powerful processor,
FLASH RAM, a hard disk, and hence, in a sense, is a
mini computer.
-
The intelligence to calculate routes on the user side
is also incorporated.
-
After calculating the route, the last aspect is signaling.
Every box has to be configured to support a leased line
between two points. But such signaling of routes is
incorporated
into the box itself.
Therefore,
we have a very intelligent network management
system that relies on software running in the box and is
implemented using protocols and algorithms borrowed from
the IP world and adapted for use in the telecom space. This
makes the box scaleable and easy to use. These boxes are
priced depending on the configuration and range between
US$ 15,000 and US$ 35,000.
Are you lobbying with the government to promote
optical networking?
There
is no need for us to lobby with the government, as the carriers
realize that optic fiber is the only option they have, because
technologically speaking, there is not even a technology
in the lab that could be a potential replacement for optic
fiber. Thus there is no threat for optic fiber in the next
10 to 15 years as the risk factor is zero.
Since the carrier market segment is miniscule, don't
you foresee a saturation point?
There
isn't a very huge market in terms of the number of customers.
But each customer is a very large customer with networks
worth 10s of millions of dollars. Besides, every business
goes through a cycle. There is a slack period and a growth
period. Our business is going to be driven not by the number
of users but by how much bandwidth each uses.
Do you see optic fiber deployment on the LAN front?
Optical
networking within the enterprise hasn't taken off as it
has on the backbone. Optical networking is suitable only
for large data transfers. In the enterprise, Ethernet has
made a strong presence and now even gigabit Ethernet is
gradually moving in.
There are other optic technologies emerging for LAN deployment
such as passive optical networking (PON). It will take quite
a long time before optic fiber is deployed within the enterprise,
and not until there are applications that will warrant optic
fiber deployment.
Mahesh Rathod and Shubha Murthy can be reached at netmagindia@vsnl.com
>>
