Now that Y2K projects are over, the new big growth area for the Indian IT industry is telecom software
There are very few entrepreneurs in India as unhappy as the Nandas of the Escorts group. A share they divested at about Rs.190 in Hughes Software Services two years ago is now setting the markets, ablaze at nearly 25 times that amount, hardly two months after its IP)O at Rs.630! The Nandas, along with a handful of others like Pradman Kaul of Hughes Network Systems and K.V. Ramani, backed a project to develop telecom software in India way back in 1991. And that proved to be a far-sighted move. Eight years of evolution from a privately-held company creating software for various Hughes enterprises in India to a listed company with a formidable global client list of 45 is a genuine success story. Suddenly Indian investors have woken up to a new opportunity–telecom software.
The Indian IT industry, however, realised its potential way back. “In fact, our first product in that direction was in 1985,” says Ajit Swant, vice-president of Tata Consultancy Services. “It was just local area networking product called Falcon (fast-access local connectivity). But then, networking itself was relatively unknown. Today we are engaged in real time maintenance of switching software for several clients like Nortel, Lucent, and Nokia in dedicated centres. In fact our centre at Borivli (a northwestern suburb of Mumbai) has Nortel switches with 300,000 lines to simulate a real time problem and try out a solution. When our client informs us of a glitch we remotely diagnose it, work out a solution, try it out, and then send it over a dedicated satellite communication link to his system,” he adds.
Wipro and Infosys too have dedicated centres for telecom equipment vendors. The experience gained in designing the hardware and software for C-DOT switches has is of good use for the IT industry as a whole. In fact, the very mention of “switching” excites Bishnu Pradhan, former executive director of C-DOT and currently vice-president at ICO-Global. “Switching is a good example of an area where we have significant scope for software development. When joined C-DOT in 1990 after working in the IT development area, I saw tremendous potential in doing developing software for switches. No one in the country had as much experience in this area as C-DOT. I conveyed this to a number of MNCS, including Siemens, which subsequently set up a captive centre in Bangalore.
“A large number of other MNCS in the business followed suit. The reasons were simple–the software development cost for a new switch was as much as 70 per cent of the total cost. Manpower availability was also limited, even if they were willing to pay the cost. So coming to India, where a large number of former C-DOT employees were available, was but natural. The second area in telecom is network management systems for large national and international networks. These systems conventionally use standard high-end platforms and system software, and application-specific software needs to be developed. The third area that comes to my mind is in wireless technology. Whether it be in new CDMA-based systems or RF designs for networks being set up all over the world, these are all heavily software-oriented, requiring good radio engineering and network engineering knowledge,” adds Pradhan.
True to his prediction, almost all major equipment vendors of the world – Motorola, Siemens, Alcatel, Nokia, Lucent, and Nortel Networks- are here. In fact Motorola India Electronics, which has two centres in Bangalore and Hyderabad with over 550 engineers, has become a major global resource for Motorola worldwide. Mohan Kumar of Motorola says: “If the PC is sold in millions, telecom information appliances can be sold in billions. That will depend on the production of low-cost ‘system-on-a-single-chip’ solutions. Indian companies can generate a lot of IPR in this area.”
Today ex-C-DOT engineers dot the entire telecom software landscape and they are much sought after as there is a genuine shortage of people who understand both computer programming and telecommunications. “Though we recruit telecom and electronics engineers along with computer science graduates, we still have to give the man intensive three months of training in today’s convergent technologies,” says Ajit Sawant. Arun Kumar, president and managing director of HSS, in fact is involved in setting up a separate institute attached to Delhi University which will help narrow the gap.
However, switching is not the only component of telecom software. What is fuelling the entry of many players into telecom software, with an expected growth rate of 100 per cent year on year in India, is the magic work “convergence”, since it is not difficult for a computer software engineer to pick up Internet technologies. In fact, voice-over IP is the hottest item on the agenda of all Indian companies (see box).
“The telecom industry has remained a big area of innovation, if there is scope for innovation and IPR anywhere, it is in this industry,” says Kiran Deshpande, CEO of Mahindra-BT, 90 per cent of whose revenues come from telecom. Joseph Samuel of Nortel Networks, a Canadian telecom major, confirms that there is great opportunity for Indian companies to do high-end work and owning IPR, but cautions that for the time being it will have to be in partnership with global telecom majors.
K.V. Ramani of Futuresoft agrees. His company already owns considerable amount of IPR in terms of software stacks. “As the industry moves from the outsource model to joint development, it will automatically move up the value chain,” he says.
It is said that, in computer software, Indians missed the Cobol bus but caught up with the rest of the world when Unix arrived. Similarly, though Indian companies missed out the earlier developments in telecom, the development of packet switching (the technology underlying the Internit) and wireless technologies have now provided the opportunity to jump on to the band-wagon.
The entry of private operators into basic and cellular telephony and Internet access in India has created opportunities for the Indian IT industry to even provide end-to-end solutions and thereby gain extremely valuable” domain knowledge”–detailed knowledge of all aspects of the telecom business from billing to network management. As is well known, “Indians have no problems with ‘platform knowledge’ whether it is C++ or Java, but it is domain knowledge that is most valuable. “Our experience with providing the complete solution for Tata Teleservices in Andhra Pradesh has helped us a lot in this regard,” says Ajit Sawant.
“The Indian software industry does not face a problem of competence. Major Indian software companies are in all areas of wireless, wireline, and switching. Revenues will, however, depend on what products, services, or solutions they sell in the international domain. Typically, earnings will be low for software engineering services and higher as we move up the value chain from components for OEMS and end-to-end solutions for network operators and service providers.” says Arun Kumar. “In telecom software the value chain starts at body shopping, with the lowest margins of about 5 percent, and moves up through onsite and offshore work to products and solutions (60 per cent). At HSS we have been able to reach the last three–IT project services, packaged software, and solutions,” he adds.
Software industry pundits predict that revenue from this sector will reach $3 billion by AD 2003. In fact some analysts have pegged volumes at $15 billion by AD 2008. They expect telecom software to contribute a healthy 25 per cent to total software exports. Even the most conservative growth estimate is a mind-blowing 100 per cent.
“We have to invest heavily in training. HR is a major issue in this field as it takes time to develop skill sets required for high-end work. If the engineers of not get challenging assignments, if the company’s main attention is either towards low-end work or quick money through dedicated offshore development centres, then we will see a lot of migration,” says an industry source. Thus, for further growth in this area, the key is building knowledge collectives and retaining them. fortunately we are also seeing the trend of major telecom equipment vendors now trying to become telecom solution vendors.
Their strategy will involve partnerships with software developers right from hardware design stage and even strategic cross-holdings and acquisitions. Fortunately, frontranking Indian companies in IT like TCS, Infosys, Wipro, Satyam, and Mahindra-BT, along with the specialists at HSS and Futures soft have already proved themselves as cost-effective developers with short concept-to-market cycle times. Thus telecom majors are ready to become codevelopers with Indian companies today, a major shift from legacy systems and software maintenance.
Leaf through any of the various international research journals produced by IEEE in communication technology and you cannot fail to notice the disproportionately large number of Indian contributors. Converge that solid academic expertise with the entrepreneurship of the likes of Pradman Kaul, Arun Netravalli, or Gururaj Deshpande, add a bit of networking, and voila, Indian companies can still emerge as frontrunners in a major communication technology revolution in the 21st century.
What’s telecom software?
As digital telephony came into being in the 1970s and microchips started getting integrated into switches, it became possible to build several functions which earlier needed human intervention directly into a switch. The process accelerated in the early 1990s. Earlier, an operator in the middle or a secretary at the end could redirect a call or inform the receiver that there was a call holding or find out the identity of the caller and so on. Now it can be done electronically. Moreover, as the number of telephone users and, consequently, the traffic increases, network problems like congestion, alternate routing in the case of a particular segment of the network breaking down, and so on, became critical to the quality of service.
Today telecom networks can do things one couldn’t imagine earlier, and that too automatically, millions of times a minute. Hence they are called intelligent networks. The equipment is imbued with ‘intelligence’ in the form of the software in the switch. As competition increases and new technologies are developed, newer features can be provided to the end user by upgrading the software. It is not uncommon for a digital switch handling anywhere between 10,000 and 100,000 telephone lines to run on software that has 15-30 million lines of code. The size of the programme does not depend on the size of the switch but on the number and complexity of features one wants to build into the application.
Earlier most of the code used to be written in proprietary “assembly languages” consisting of instructions written in a code that could easily and quickly be understood by the machine. Programming languages like C, Java, Basic, etc, which are very close to normal English, take several seconds to process, whereas a “real time” programme should run in a thousandth of that time. If properly optimised (cutting away verbosity in programming), assembly language programmes, which look like a string of 0s and 1s, can be executed in microseconds and milliseconds. Such software was embedded into the system as “read-only”. Today such proprietary software is giving way to open systems which are easier to develop, upgrade, and maintain.
Maintaining system software involves finding bugs and fixing them or writing new patches to add new features and so on. However, when new patches are added on, the whole process needs to be optimised again so that the functions can still be completed on the fly.
With growth in the sophistication of the microchips used in the switches and the development of new programming languages, software could be made more “object-oriented”. That is, instead of a programme looking like spaghetti with entangled code, which becomes more and more difficult to modify or be even understood by a different programmer, it could be written in a modular format, where each module has a well-defined function. Such programmes, written with open architecture in Unix and C, are popular but they have to be interfaced with the old proprietary software. Such interfacing is another major part of telecom software development.
With the advent of the Internet and the possibility of using the network for transmission and switching of voice, data, and video, a whole new exciting field has opened up for software developers. And now the urge is to add on the feature of mobility to such convergent technologies so that wideband mobile communication becomes possible.
In short, there is a veritable explosion of new technologies as service providers compete to come out with newer features and ever-increasing bandwidth. That is what is driving the phenomenal growth of the business of telecom software. From the current level of $ 15 billion it is billed to shoot up to about $ 80 billion in 10 years.