'We can change the path of packets in the middle of a phone call'

Tom Evslin, chairman & CEO, was recently in India as a keynote speaker

for iLocus IP Telephony show. Evslin began his career by setting up a

communications software company, which was sold to Microsoft later. He formed

his own venture called ITXC on IP Telephony. In an exclusive interview with

Pravin Prashant of Voice&Data, Tom Evslin talks about IP telephony and the

company’s plans for India. Excerpts of the interview:

How would you rate your first visit to India?

It is an exciting time for telephony in India. We are just a year away from a

much more competitive atmosphere and everybody is looking out for that

opportunity. There is excitement about the events that are going to happen, how

they will unfold, and what the best way of making money is. As we are in a

country where tele-density is 3 percent, there is a huge opportunity for growth

in terms of data and voice. We see many opportunities for voice over IP and to

use Internet infrastructure for voice needs.

How should India move ahead in terms of telecom infrastructure?

I am recommending the use of IP infrastructure for use in voice

communications, too. Instead of building two separate infrastructures or

investing in traditional telephony switching, using the IP infrastructure for a

dual purpose makes more sense. There is already a much more competitive

environment in the ISP business than there is in voice business. When telephony

opens up, I think there can be a vigorous competitive atmosphere if IP

infrastructure is used. So, it does not require a huge amount of capital to

build a separate telephone network.

How cost effective is packet switched network in comparison to circuit

switched network?

My company has spent around US $100 million capital since we began in July

1997. With that capital we have a network in 100 countries which would have cost

billions of dollars if we were building a dedicated PSTN network. It simply

wouldn’t have been possible to achieve in a couple of years. At a time when

capital is so very difficult to get in the United States and Western Europe,

think how much more difficult it will be for the developing world to build a

network on top of the Internet infrastructure! Instead, the new technology would

enable them to spend capital only once for building up IP infrastructure.

About your India plans?

We haven’t formalized the India plan. The reason I am here is to learn

about all that is possible, what Indian companies plan to do, and what the

regulatory environment will be.

The model we’ve used to work in China, will be applicable for India, as

well. A couple of years ago, the Ministry of China granted some licenses for

Internet Telephony and one of the licensees was China Telecom, with whom we are

working today. They were able to bring IP Telephony in eighteen cities of China

within a month by just connecting to ITXC.net. Traffic between cities is being

handled by China Telecom and traffic leaving China goes through the ITXC

network. Our role is to provide technical expertise, build up the network and

provide connection to the rest of the world. Similarly in India, our role would

be to bring Internet telephony in and out of India and also to help Indian

companies in setting up IP networks. They can establish IP telephony networks

for meeting their internal needs and enabling them to exchange traffic with the

rest of the world.

The problem of bandwidth exists outside the country, but is far more acute

within India. How should India tackle this?

I think India is much more advanced than what China was as regards the

internal infrastructure and external connections. Actually, among developing

countries and checking the country’s size, the infrastructure is pretty good.

The availability of voice traffic and the revenue the voice traffic generates

will help to build the IP infrastructure as it will be paid for daily usage for

voice use as well. I think in some parts of India in the next few years, the

only IP connection will be a satellite connection. While that’s not ideal, it

is much better than having no connection at all. Many ISPs in India today have a

dedicated satellite connection. That’s how they are connected to the rest of

the world.

In future, we will see a mixture of landlines and broadband pipes but because

of limitations there will be lot of satellites (very small and very large

satellite locations) around India to make sure that all of India is connected.

How has your company performed in terms of product and services?

We have about 150 customers for our phone to phone services that run over

ITXC.net. It includes 14 of the 15 largest international carriers in the US and

six of the top 10 PTTs in the world. Interesting thing is, when these companies

use our services to provide service to their customers, the quality that we are

able to get on the public Internet is at par with the circuit switched network,

and customers are unable to differentiate between the call quality.

We started ‘Push to Talk’ service last year, a service that is getting a

rapid start in the travel and hospitality industry. Many hotels and hotel chains

internationally, are using ITXC Push to Talk service on their website and

including it in their e-mail, too.

Are you planning to promote Push to Talk service in India?

Push to Talk would be particularly interesting for India with respect to the

fast-growing call center business in India. The economies of Push to Talk button

on a web site is seen primarily in the US and UK but India being a call center

hub, shows promise too. We would be talking to Indian ISPs and they will offer

services directly to call center companies. Especially seeing how strong the

call center business is here in India, and that Push to Talk is a call center

product.

Any plans of opening an India office?

We do not have any specific plan now. We do have an Asia-Pacific office at

Singapore and our India operations are managed from there now. But it is a

possiblity next year in 2002 with more opportunities opening up.

What are the immediate obstacles that you face as a carrier’s carrier?

The first challenge is to provide interconnection between packet switch

network and circuit switch network because most people are in the circuit switch

network. Second stage, get good quality over Internet. Initally, Internet

Telephony got a bad reputation for calls, which it deserved. Internet being a

public highway system, packets going through it got stuck due to congestion and

delay.

So, until we developed the technology it was not possible to get carrier

grade quality on net. But that’s largely been overcome with the new

technology.

The third challenge is to provide end to end IP Telephony first to the

premise of corporate customers and then to the premise of residential users.

What are the problems on the IP backbone and how have you tried to

minimize them?

We use our best value routing technology. What we do is, send test packets of

all kinds though the routes we use to send voice traffic and then measure those

routes for delay and packet loss. If we find that there is congestion starting

to build up resulting in long delay and high packet loss in a particular route,

we steer the packet to a different path through the Internet to their

destination. We do it quickly enough to prevent the quality of call from being

degraded. The aim is to not let congestion get to the point where it noticeably

interferes with the quality of the call. Since it is a property of IP that all the packets of conversation do not have to take the same path, we

can change the paths of the packets in the middle of the phone call. And without

interrupting the phone call in anyway.

What about a separate country code for IP Telephony?

It will not happen. If most of the Internet Telephony were happening from PC,

it might have been the case. However, most of the Internet Telephony happens

through phone. It doesn’t make sense for the phones to have different numbers.