A few years ago, John Chambers, CEO, Cisco Systems—the company that is regarded as the VoIP pioneer—made that famous statement, "Voice would be free," at a service provider gathering in the US. He was obviously taking a shot at the century-old public phone network and the familiar circuit-switching TDM technologies that are a hallmark of telephony. IP, he declared, was a revolution, that would change telephony by carrying voice calls over the Internet in the form of packets just like it carried any other application data.

Much water has flown under the bridge since then. Fresh in our minds are the sad stories of bursting dot-com bubbles, melting ‘new world’ CLECs, telco capex embargoes, and billions of dollars lost in market capitalization (including that of Cisco).

Let’s take a close look at VoIP. Many of the shortcomings of IP have been rectified. IP can intelligently identify which packet is to be routed on a priority basis. MPLS, DiffServ, QoS, etc, have given the much-needed respectability to IP. Today, it is not just Cisco that is into VoIP; VocalTec and Sonus are also selling VoIP solutions to carriers across the globe. What is more, traditional telecom equipment vendors like Nortel, Lucent, Alcatel and Motorola, too are addressing the VoIP needs of their existing and new service provider clients.

So…is VoIP about to revolutionize the way we place and receive calls? The answer is a clear no. True, VoIP is fast garnering up popularity among enterprises for their internal voice communications. But public telephony… it has been a depressing experience. The so-called Greenfield operators couldn’t help falling from their tall castles that turned out to be built in air. And as far as implementations by incumbents and larger telcos are concerned, they can best be termed as experiments… research findings substantiate that.

Global Traffic Pattern
According to a TeleGeography research, VoIP traffic accounted for only 6 percent of the world’s international long-distance (ILD) voice telephony traffic in 2001. The US–Mexico route had topped TeleGeography’s VoIP route rankings, with one-third of US’ international VoIP traffic in 1999 and 2000. TeleGeography had also indicated that VoIP call volumes into China, Brazil, and India would increase dramatically over the next few years as their IP infrastructures matured and VoIP termination arrangements got expanded.

According to another very recent report by iLocus, over a billion VoIP calls are handled per month by a variety of wholesale and retail providers globally. The report, based on a survey of 100 IP telephony operators worldwide, estimates that 6.8 billion calls traveled over IP during 2001. Significantly, China contributed around 10 percent of the monthly VoIP traffic in January 2002.

VoIP vendors argue that this is only the beginning for VoIP and its share would considerably go up to have the big telecom operators take note.

The Indian Scenario
VoIP uptake in India has been quite dismal so far. The current situation only lets one realize that expectations from India were far-fetched. VoIP here is restricted mostly to international calls. And almost all ILD traffic is from PC-to-phone through the Internet telephony calling cards channel. Data Access, the only pure-play VoIP ILD player, is yet to take off on a significant scale.

The remaining two operational ILD players—incumbent VSNL and challenger Bharti Tele-ventures—still operate on traditional telecom switches. Does VoIP figure in the immediate expansion plans of these players? Only minimally. VSNL issued a tender for VoIP equipment recently, but industry watchers say that the size of the tender is negligible in comparison to its regular TDM purchases. Bharti Tele-ventures too, is juggling with a few solution proposals to find its suit, just for about 10 percent of its ILD traffic requirement in the short- to medium-term.

The other company that is expected to be a significant player in the ILD game is Reliance. It is said to be evaluating about half-a-dozen VoIP vendors. But at the same time, there are talks of large TDM equipment orders as well.

If biggies like Reliance, Bharti and Tatas are not going whole-hog for IP, then it means only one thing—VoIP is still far away.

The Miscalculations, or What Went Wrong
The IP era has always been associated with speed and fast-forwards. It was all about taking a risk and benefiting from it with huge VC finances, IPOs and market caps, etc. All those who did not appreciate this point of view were termed as ‘old world’. IP vendors had big hopes from the Greenfield operators. These new operators, armed with IP, were expected to do to incumbents what MCI did to Bell operators.

Now that the Greenfield world is licking its wounds, IP’s generals are nurturing dreams of conquering lands like China and India, where the governments are de-regulating their telecommunications services and inviting new players to the fields.

Talking of India, the VoIP dream so far remains that—a dream. And IP’s generals are trying to understand why new Indian operators have not leapfrogged? Why are they still investing hundreds of millions of dollars in the purchase of ‘old world’ equipment? What went wrong? What were the miscalculations?

n Miscalculation 1: Cheap VoIP services will cannibalize the ILD telephony market!

It was presumed that in cost-conscious India, people would start calling up through VoIP, to slash their long-distance telephone expenses. Yes? No.

Cheap is a misnomer for toll-quality VoIP because only PC-to-PC and PC-to-phone calls over the un-managed Internet jungle come cheap. Toll-quality VoIP over a managed IP backbone will be costlier. Let’s look at it this way: why is Internet telephony cheap? It’s the compression, stupid! IP telephony companies can compress from 6 up to 16 E1s (a telco’s E1 is typically capable of carrying approximately 30 telephone lines) into one stream on the trunk side. That translates into better utilization of bandwidth, thus saving on the access bandwidth.

But does saving matter much when the bandwidth prices have crashed to a record low? Moreover, DWDM dynamics are trying to imply, "Bandwidth will be free". Also, significant price drops in TDM equipment have put traditional voice carriers in a position to further reduce prices over and above the mighty cuts that we have already seen. Jagbir Singh, V-P (technology), Bharti Telenet, claims that his company can also compress 4 to 6 E1s in a trunk stream. "Even with that comparatively poor compression rate, we can be much cheaper than VoIP". He promises to meet every price combative move of pure-play VoIP operators like Data Access with competitive price cuts.

Even if pure-play VoIP operators set an unmatchable ILD tariff, much of their success in carving a survival strategy will depend on how Indian customers respond to their services. The government might have already clipped VoIP’s wings by putting the condition that no PSTN customer equipment can be interfaced with VoIP traffic. This means that even if an enterprise wants to subscribe to a cheaper ILD service, he will have to purchase VoIP terminals that are almost ten times costlier than a normal telephone terminal. The tariff of VoIP in that case has to be significantly lower for the enterprise to consider buying a VoIP connection. In most cases, it will be limited to a few lines for the top executives. Surely, the investment doesn’t appear justifiable.

That leaves segments like the SME, SOHO and home. But the cream in these segments might have already gone for PCs.

And once you have a PC and an Internet connection, you can purchase a Valuefone card from HCL Infinet or a Phonewala card from Net4India and make those occasional ILD calls to a business associate/relative in New York or London. Why bother to purchase a Rs 15,000 VoIP phone now, when you are not sure how long VoIP tariff will remain the cheapest in the ILD call market?

n Miscalculation 2: Telcos, especially new ones, will choose VoIP for its cost benefits

VoIP vendors claim that operators can save considerably on both opex and capex by deploying a purely converged network for voice, data and multimedia services. It is claimed that it is in opex that VoIP has the biggest advantage over TDM networks.

In the circuit-switch paradigm, a point-to-point connection is established and the entire circuit is busy in spite of the fact that there are several periods of silence during the conversation. IP on the other hand, can send packets in no particular sequence, thereby enabling several packets from several calls to be jumbled up and sent over the international fiber and satellite trunks (in fact different packets of the same call could take different routes to reach the destination). Thus with IP, voice traffic can be transported with much more bandwidth efficiency. Over and above that, compression technologies give a further edge to IP in saving bandwidth cost. Also, by deploying VoIP switches telcos may reduce manpower costs that could be reduced by centrally managing the network.

So why are new telcos not embracing VoIP? And why are incumbents not strategizing to migrate from legacy networks to IP? Simple—it’ll make poor business sense as of now. Out of the total international telecom traffic terminating in India, less than 10 percent is IP traffic. The same is the case with the outgoing traffic. Given this small amount of IP traffic, where is the need for large-scale deployment of VoIP equipment?

Sections in the vendor as well as service provider communities believe that practically, VoIP does not distinguish itself from TDM cost-wise. While VoIP may have opex competitiveness, TDM has a capex edge over VoIP. Though one requires lesser number of VoIP ports to complete a call, these ports cost as much as five times the TDM ports.

Many are of the opinion that the battle between TDM and IP will be finally decided in the long run by the killer value-added services that can be deployed over and above plain telephony. Here too, TDM switches with their Class 5 architecture have a distinct edge over VoIP switches, which are still at Class 4 level. Class 5 switches, with their Intelligent Network (IN) architecture can allow service providers to deploy more than 250 applications.

Not all these applications are popular though. VoIP vendors argue that seldom do we use more than 5-6 of these applications. They also insist that IP, with its converged voice and data applications and flexible service provisioning facilities, is getting popular day by day. Vijay Yadav, country manager, CommWorks, says, "Operators are realizing the value of picking up the popular services parts from a softswitch and integrating them with the traditional TDM switch platforms".

But picking up IP components is not that easy. The problem of integration with the complex billing and operational support systems of existing TDM networks is very complex. Such issues will need to be resolved first.

n Miscalculation 3: Days of TDM switches are numbered

In reality, there’s no way that telcos are going to rip off their existing equipment in which they have put in billions of rupees over the past several years. Yes, they will be requiring a transition strategy from whichever vendors provide their switching equipment in future. And that in any case almost all TDM vendors are addressing through their research or acquisition investments in IP. Vendors like Alcatel see the markets for both TDM equipment and IP equipment separately in the short- and medium-term. Gerard Chevalier, director (voice networks division), Alcatel India, says, "Today, new generation network (NGN) is a fashion. But, the life of a TDM switch is far from over. We are in the middle of a river. The other side is clearly IP with its big backbones. However, to reach the other side, you have to swim quite a bit."

Alcatel, like Lucent and Ericsson, the other largest telecom equipment manufacturer according to ITU, is in the process of getting ready for the all-IP days ahead. In other words, these vendors are suggesting that operators need not hurry and invest in completely new equipment. Instead, they can keep investing in time-tested TDM switches, which are now upgradable to the soft switches.

If that’s a promise, operators like Bharti Tele-ventures think there’s reason to believe it. New operators like them doubt that VoIP equipment will enable them to match the quality of voice services offered by incumbent operators who use TDM switches that deliver 99.999 percent reliability and a host of intelligent features. IP equipment is still not as reliable as the TDM switches. And when you try to make it more reliable, the costs go over the roof. In India, you cannot terminate on a BSNL equipment without SS7 signaling support (BSNL is the default domestic carrier of telephony traffic to the international gateways of VSNL in most of the cities in India, as most often there are no long-distance carriage alternatives at all).

Given all this, it’s hardly surprising that operators in India are considering VoIP only as an add-on rather than the de rigueur.

Almost all operators, including the incumbents BSNL, MTNL, and VSNL, and the new operators Bharti Tele-ventures, Reliance, Tata Teleservices, and Hughes Tele.com, are still buying TDM equipment in bulk. VoIP purchases are currently aimed at targeting niche segments rather than mass deployments.

IP’s Future, but…
All said, one thing is beyond doubt—IP is the future. Even TDM equipment vendors agree to this. Look at the roadmap of any telecom equipment vendor, and IP is there. Every major telephone switch vendor has a switch with IP capability. PBX manufacturers like Avaya and Nortel are clear that every new product that comes out of their factories is IP capable.

Comparing port-to-port, today VoIP is definitely costlier for telcos on capex-basis. But this can change when volumes start driving down the costs, as witnessed in the case of RAS ports for data communication. And VoIP certainly has opex advantages that come through service flexibility and efficient usage of shared bandwidth.

Also, it can’t be denied that VoIP has drastically brought down the entry barrier for telecom operators. The biggest testimony to this is that an ISP like Data Access, with an investment of a few million dollars, is operating an ILD service and competing with companies like VSNL and Bharti Tele-ventures.

But changes of colossus scale can’t happen in a couple of years. Much will need to be done before VoIP can become the norm for public telephony.

n Technology Improvements: First of all, VoIP vendors have to change the perception about VoIP’s inferiority to TDM switches. In spite of recent developments like MPLS, DiffServ and IPv6, there is still some doubt among telcos when it comes to VoIP. Vendors will have to work for improvements in areas like reliability and features. For example, there is a need for VoIP switches that are reliable for Class-5 applications, so that the technology can be used at the access level and not just in the longhauls and backbones.

n Standardization: The other major focus should be on standardization. VoIP vendors must work hand-in-hand with standards bodies of telecom technologies, such as ITU, ANSI, TIA and EIA, apart from the standards bodies of individual countries, like the TEC in India. The turf war between the ‘new world’ and ‘old world’ vendors must end. Yes, they can fight out there in the market. But back at the standards-making forums, both parties must work amicably towards standards that will benefit their common end-customers—the operators. As of now, operators complain that internetworking among various VoIP equipment itself is an issue, leave alone internetworking with TDM switches.

n Better Interaction with Telcos: It’s the basic rule of marketing—understanding the customer, his needs and preferences, and evolving products and services oriented towards him. To say that telcos have a mindset problem is incorrect. They are faced with huge challenges in terms of network complexities, operational issues, and most importantly, financing. They are considering VoIP but they need more time to transit to the ‘new world’. They also want better utilization of the installations done so far. So VoIP vendors will have to come out with a transition path that is more viable for operators. It can be a good idea for VoIP vendors to work closely with telcos to build VoIP success stories, which will convince telcos to go for more large-scale VoIP deployments. Playing the pipe hasn’t worked, but a hand-in-hand approach should.

Nareshchandra Laishram

VoIP - TDM FACE - OFF