Thursday, November 3, 2011

Telecom Path-Breaker?

Does the draft National Telecom Policy-2011 reflect true brilliance or smoke-and-mirrors? It will be a game-changer if a shared network is implemented effectively.
Shyam Ponappa / November 3, 2011

There’s much to criticise the government about for not initiating systematic reforms. Yet, the draft National Telecom Policy 2011 (NTP-2011), announced three weeks ago, is a stunner.1 It begins with a solid, integrated-systems preamble to IT, Communications and Electronics, followed by an excellent vision statement: “[to provide] secure, reliable, affordable and high quality... telecommunication services anytime, anywhere.” A sound beginning, although open-ended in terms of how the details could evolve.
There are potential problems with such high-level pronouncements, of course. A number of commentators castigate the motherhoods in the draft. With a lofty perspective and few details, much depends on how the open-ended possibilities develop, including the difficulties of execution in dealing with ground realities and obstacles.

An Assessment
NTP-2011 addresses six major areas: spectrum, licensing, broadband, convergence, roaming, and manufacturing. Focusing on the first two, there are sweeping proposals:
a) licences will not be linked to spectrum; and
b) spectrum sharing will be permitted.
Some view the separation of licences and spectrum as retrograde, because spectrum is essential for service delivery. Others suggest that transgressions that led to the scams are now being inducted as new policies, e.g., operators accessing networks they do not own, which is characterised as being against the public interest. Some heap opprobrium, alleging that like the previous policy, NTP-99, which they call retrograde (although it led to the phenomenal growth in mobile telephony), its main purpose is to allow companies to avoid paying licence/auction fees to the government.
  • The last expostulation is the most ludicrous, because revenue collections after NTP-99 far exceeded estimated fees foregone: Rs 20,000 crore estimated “loss” by March 2007, but Rs 40,000 crore actually collected, and Rs 80,000 crore collected by March 2010.2 Add tax collections on exponential growth with increased profits, and the result is even higher total government revenues.
  • Opposing operator access to networks arises from confused objectives; blocking access is like cutting off one’s nose to spite one’s face. The purpose of the sector is to provide services and access to users for legitimate activities. The public interest lies in facilitating access on appropriate terms.
  • To evaluate licensing and spectrum, begin with the premise of shared spectrum. Spectrum is essential for effective service provision, particularly in the rural and semi-urban areas with about 70 per cent of the population. An aspect not commonly known is that larger bands of spectrum enable more efficient throughput. For example, 1 MHz of a 12 MHz band carries 50 per cent more traffic than 1 MHz of a 6 MHz band. An estimate of the benefit to Indian operators of more bandwidth at international norms is a reduction of 20 per cent in operating costs.
Spectrum Occupancy
In practice, assigned spectrum is idle much of the time, except during the busy hours in India’s heavy-traffic metros, for extraneous reasons: too many operators, with too little spectrum, in too- narrow bands. This aspect becomes clear from spectrum utilisation or occupancy studies. For instance, the chart shows spectrum occupancy in Bangalore, Edinburgh and Stony Brook (New York) sometime in 2011.
Spectrum Occupancy in Bangalore, Edinburgh, and Stony Brook, NY.

dBm: decibel; MHz: Megahertz; GSM: Global System for Mobile Communications; CDMA: Code Division Multiple Access

The low readings (250 to 850 MHz in Bangalore, 600 to 950 MHz in Edinburgh, and 500 to 850 MHz in Stony Brook, NY) indicate available “white spaces” that can be better utilised.
  • High-traffic cities like Delhi and Mumbai have much higher utilisation than cities elsewhere in the world. It comes at increased costs to operators, because of advanced equipment and the closer spacing of towers, as well as having negative environmental effects. If a system with on-demand access to centralised, more efficient spectrum bandwidth were available, the capacity would be much higher, while operators would gain tremendous savings.
  • Another aspect has to do with the structuring and pricing of shared spectrum. One scenario for sharing is to enable operators to share assigned bands on mutually acceptable terms, leaving the onus of structuring and deployment on the respective operators, as for mobile telephony towers. As with the towers, there are likely to be coalitions of operators/independent entities who are able to work out arrangements among themselves, while not attaining the ultimate efficiency of unified coordination. For instance, participants who share towers in India share passive but not active infrastructure, and a critical element of active infrastructure is spectrum.
  • An alternative scenario would be mandated spectrum sharing. Spectrum on demand could be made available to any operator/counterparties for the duration of every communication “transaction”. This would need a database-driven Dynamic Spectrum Assignment facility, as deployed by Spectrum Bridge in the US. The more efficient throughput would mean higher traffic capacity for a given investment through better utilisation.
  • The distributed processing alternative through cognitive radio in every user device is (a) much less efficient, and (b) far more expensive. The market consolidation-through-acquisition approach, with more auctions, is the least efficient and most expensive.

Common-Access Networks

There would be further efficiencies if the entire network (and not just the spectrum) were accessed on-demand for payment per use. Another benefit from a public perspective would be much lower collective investment in resources, because of better utilisation. A third benefit would be the reduced environmental impact because of a lower carbon footprint and radiation from two or three common-access national networks (assuming competition is essential for effectiveness and efficiency).
In other words, database-driven, shared spectrum and networks have to be organised and managed as a coordinated unit if the potential benefits are to be realised. America is doing this with TV white spaces/the digital dividend, through the appointment of 10 database administrators (including Spectrum Bridge, Google and Microsoft). This should elicit our interest.
Once the government and stakeholders accept these concepts, the next major task is structuring the networks as consortiums to align the interests of operators and network providers, with state-of-the-art lead partners. In this process, incorporating and reorienting BSNL and MTNL as guardians of national interests with oversight by an adequately empowered regulator will be the remaining major tasks.

2 TRAI, 2005:
CAG: Report 2009-10.pdf