Telecom Blog

Twenty years ago, “spectrum” implied the colours of the rainbow. Now, we understand that spectrum also relates to mobile phones. We encounter spectrum daily, in TV remote controls, microwave ovens, even sunlight. So, what exactly is spectrum, and how do government and commercial decisions on this scientific phenomenon affect public facilities and costs?

“Spectrum” is short for “electromagnetic spectrum”, the range of radiated energies that envelop the Earth. This electromagnetic radiation (EMR) is primarily from the sun, and secondarily from the stars/cosmos, radioactive elements in soil, rock and gases... .

One section of EMR is visible light; another is radio frequency (RF) spectrum. There are many other “wavelengths” in EMR with different characteristics and effects, such as infrared and ultraviolet rays. All countries have the same RF spectrum in equivalent areas.

How is spectrum used?

The length of a wave, its associated frequency (“wavelengths” or “cycles” per second) and energy determine its usage (see Figure 1).

1. Radio waves are relatively long, with wavelengths from 1,000 metres (1 km) to 10 cms, and frequencies from 3 kilohertz (3,000 cycles per second) to 3 gigahertz (GHz) or 3 billion cycles per second for the shortest, sometimes also called microwaves. (There are longer waves, e.g., electric power, of several km.)
2. Microwaves in the centimetre and millimetre range can have frequencies up to 300 GHz. There is an overlap in terminology depending on use; microwaves for cooking use several hundred watts of electricity at RF wavelengths of about 32 cms (915 MHz) and 12 cms (2.45 GHz). Microwaves from low-powered devices of a few watts at these frequencies are used for communications, and emit insignificant heat.
3. Infrared waves are smaller, and are felt as heat, e.g., from lamps and infrared grills used for cooking. Higher infrared bands used for communications in remote control devices and for imaging/night vision have no heating effect.
4. Wavelengths between 700 and 400 nanometres (about 430 to 750 terahertz or THz) form the visible spectrum from red to violet, combining to form white light. For example, we perceive wavelengths of about 635-700 nm (430-480 THz) as the colour red.
5. Shorter wavelengths form ultraviolet rays, of which those around 380-280 nm cause sunburn. Sunlight at sea level comprises about 53 per cent infrared, 44 per cent visible light, and 3 per cent ultraviolet rays.
6. Yet smaller waves are classified as X-rays, and the smallest as gamma rays, both used in medical and industrial imaging.
   

The sweet spot in the RF spectrum for telephony and the Internet

For telephony and broadband, lower frequencies (700-900 MHz) are most cost-effective, as they traverse long distances without attenuation, penetrating walls and foliage. Radio waves in the atmosphere are affected by water vapour and ionisation, as well as events such as solar flares with bursts of X-rays. Depending on temperature, moisture, etc., radio waves may be absorbed, refracted, or reflected in the atmosphere, and by hills or other obstacles. Low frequency waves penetrate buildings and trees, and curve over slopes. Higher frequencies are more absorbed or reflected by the atmosphere; they are also more attenuated by distance and rain. Networks at lower frequencies require fewer towers than at higher frequencies.

What are 2G and 3G?

These signify different stages of technological development, starting with 1st Generation (1G) analog wireless in the 1980s, e.g., in car phones. 2G (2nd Generation) began in the 1990s with the digital wireless GSM standard for mobiles, extending to other standards, e.g., CDMA. 3G (3rd Generation) has faster data speed and greater network capacity.

What is 2G/3G spectrum?

There is no difference in the spectrum; only the convention of government regulations and harmonisation between countries by the International Telecommunications Union (ITU) earmark wavelengths for different applications. Both 2G and 3G can and do work at 800-900 and 1800-1900 MHz.

Combined with the advantages of prices dropping as volumes rise, one estimate puts 3G coverage with 900 MHz at 50-70 per cent lower cost than at the designated 2.1 GHz. 3G networks using 900 MHz (“2G spectrum”) exist in Finland, Iceland, Australia, New Zealand, Thailand, Venezuela, Denmark and Sweden, and countries like France encourage 2G networks to upgrade to 3G services.

Spectrum allocated for 2G and 3G by various countries is at Figure 2; the current and proposed allocation in India is shown below.

This shows India’s dearth of spectrum for public use because of government and defence allocations. We need innovative methods to maximise capacity given our needs, limited landline networks, and the relative costs. (For details on the chart, please see umtsworld.com.

For example, China has allocated 250 MHz in the 800/1800 MHz bands. By not charging auction fees and spectrum charges, ubiquitous networks were built at lower cost with high capacity. These result in lower costs for users and higher productivity. With its focused approach, China also developed its own standard (TD-SCDMA).

India’s spectrum allocation is burdened with short-term revenue collection for the government, and a shortage mentality. There is apparently insufficient clarity on spectrum usage for ubiquitous broadband/telephony as in other countries, let alone more ambitious targets, such as developing an Indian standard.

Our policies could address the requirement for enhanced coverage/capacity at low cost to make services available everywhere at reasonable prices. Innovative approaches to spectrum management could help get these, through:

1. Technology-neutrality: the UK and Norway have not restricted the use of recently auctioned spectrum to any technology.
2. A focused strategy for service delivery at low cost, as in China.
   

This needs a combination of methods, e.g., along with technology-neutrality, (a) data-base driven, shared spectrum usage, under trial in the US, (b) “Cognitive Radio”, whereby smart devices sense available channels for dynamic, non-conflicting use in unlicensed spectrum bands, (c) incentives for rural broadband delivery, e.g., by subvention of fees and government charges, with (d) subsidies.

Follow the original article on Business Standard

Like the industrial revolution, India missed the infrastructure systems building stage. As a consequence, even in 2001, the telecom network covered a mere 4 per cent of our population. Now, it covers about 48 per cent, but with only 20 per cent rural coverage. Our need being extensive coverage, the following what-if scenarios explore how alternative approaches might pan out.

The market-driven scenario

One approach is that all that’s required for an effective communications infrastructure is to go ahead with the spectrum auctions — that long-delayed, but always expected “3G” auction, to begin with. Imagine that it happens. What then?

Current policies will result in three winners of 10 MHz each. If they are from among present operators, they could be any three of Airtel, Vodafone, Reliance, Idea, Tata… or one or more new players: Google, Intel… until one of these wins the fourth “3G” slot when that band is made available, and so on. These operators will probably roll out networks and services where heavy traffic is expected, as with 2G so far: more extensively in urban areas. Provided other policies evolve rationally, e.g., that acquisitions are allowed and spectrum holdings can be consolidated, in the long run India may have around five or six large countrywide operators. There may be regional/segment operators with lesser franchises, or addressing specific segments. Each company will incur capital costs for spectrum and network investment, which then must be recovered from users. Network growth is likely to be on similar lines as before.

Take the evolution of India’s telecommunications policies in the 90s, and the desultory state of the sector until the reforms of the National Telecom Policy ’99 (NTP ’99), followed by reductions in revenue share to more reasonable levels in 2002. Even so, the facts show that:

• network growth is skewed heavily towards urban users; and,
• broadband coverage is abysmal.

Urban bias in network growth
By November 2009, urban coverage was at 107 per cent of the population, while rural coverage was at 20 per cent. In addition, rural wireless lines grew to 91 per cent, while the wire-line share dropped to 9 per cent; hence the increased importance of spectrum. Networks need more rural reach.

Low broadband coverage

Broadband subscriptions in August 2009 were at just seven million, two million short of the estimate for 2007. According to Comscore, at the end of September 2009, India had under 36 million Internet “unique visitors” (excluding access from Internet cafes, mobile phones and PDAs). This is roughly equivalent to the installed base of PCs, compared with about 560 million phone lines, of which under 40 million are wire-line. Something must be done to increase broadband coverage at lower prices.

The shared-network scenario

Now, imagine what shared-network facilities could do to lower costs, with no duplication of capital investment. Consider the added benefits of shared spectrum as part of this shared network — which, given the fragmented, inefficient present allocation, is the primary need for effective last-mile coverage. Then, add the benefits of substituting revenue sharing for up-front spectrum auction payments. With incentives for performance, the savings in time and money in network build-up and throughput will be immense, while the green footprint from less network hardware will be a double bonus. Government revenues will be far in excess of the foregone auction bids, together with more tax from higher profits, provided the revenue-share percentage is reasonable, as witnessed after NTP ’99 plus reduced revenue-share.

Need for reforms: Networks, spectrum and broadband

Significantly, much of the wire-line rural network is reportedly unsuitable for broadband, because of the length of “last-mile” connections, their quality and the problems of maintenance in difficult terrain. Besides, the cost — more than five times wireless, according to one operators’ association — and difficulty of laying cables in rural terrain, compounded by the impediments of clearances from multiple local authorities, render this impractical. The need is for more coverage with the same investment, even if it is private sector investment.
Therefore, network-building with spectrum reform and broadband need more supportive policies. In particular, incentives and disincentives/penalties are needed for intensive rural coverage as well.

Imagine the IT companies capturing the Y2K opportunity and outsourcing without special communications facilities and tax breaks. Those regulatory measures enabled the development of an essentially outward-oriented IT services sector. Likewise, NTP ’99 with lower revenue-share has led to high growth in telecommunications. This appears to be the best way to establish broadband as an essential infrastructure, especially in rural and semi-urban areas.

Required measures

The initiatives required cover three areas:

• Policies that make it profitable to build networks and provide broadband services all across the country, not just in heavily-trafficked areas. This will enable communications access to all, providing a platform for service delivery for government and the private sector with tremendous user benefits. These services could encompass education, health and sanitation, extension services related to economic activities, including logistics, telecommuting, entertainment and information.
• Formulating incentives and implementing them so that the primary objectives are achieved. The public-interest broadband objectives are likely to be on the lines of access anywhere — realistically, in most populated places — at reasonable prices. Key results have to be defined and tracked to ensure achievement. There’s a mountain of work in defining reasonable cost so that many more people can access broadband, while the business is commercially attractive. However, that is a separate issue. It needs to result in a large number having subsidised access, just as they must have access to food, education, and other necessities.
• Equally important, formulating disincentives that are then applied impartially, so that transgressions that detract from the objectives are penalised.

These issues must be addressed simultaneously from the perspectives of technology, economics, defence and security, and commercial interests, including existing operators’ legacy interests. For this, the government has to work with all stakeholders and specialists to develop solutions with experienced, objective facilitation. Business, government, and consumers can benefit.

The article appeared in Business Standard.

What is a good way to plan and build enduring systems, e.g., for sanitation and water in our cities and countryside, roads (rail/waterways/air for logistics), or a broadband network for communications? Some thoughts on Standard Operating Procedures (SOPs) — beginnings, processes and ends — and on some “invisible” aspects that facilitate good outcomes.

The Big Picture

Start with the big picture: the engineering background of China’s leaders has no doubt contributed to their conceptualisation and achievement so far. President Hu Jintao and Premier Wen Jiabao were engineers, as were former President Ziang Zemin, and Premiers Zhu Rongji and Li Peng. While pondering if that’s what it will take to improve India’s record on conceptualisation and execution, we find that former President Deng Xiaoping, who got China going on its current track, didn’t have it. No engineering degree, although he went to France when he was 16 for a work-study programme. Despite a difficult experience with entry level jobs in shoe manufacturing, metals, automobiles, and restaurants, he was very effective in applying himself to building China. So, there’s hope if our leaders apply themselves to long-term solutions, rather than to self-aggrandisement. This might be of their own volition, or because the public and/or circumstances force them to do so. For instance, if RTI activists concentrate on one major objective at a time, while paying attention to facts, thinking, talking and acting logically in close cooperation and coordination, i.e., with sound direction, we might get results.

Fundamental SOPs

Some fundamentals are clear enough, although we rarely seem to follow them, like an integrated systems perspective with disciplined project management (listed below). There are, however, many assumptions and enveloping circumstances that affect the drivers directly, as well as their boundary conditions and interactions. These can be easily lost sight of in pursuing a line of thought or action, or even particular disciplines. This is valid for all issues, for instance, increasing the hit rate for road projects put to bid by the National Highways Authority of India (NHAI), or the successful completion of power projects, or efforts to structure and manage spectrum for broadband. Therefore, for user-centric area planning/spatial planning, the overriding emphasis is necessarily on an interdisciplinary (i.e., multidisciplinary) approach. This is because societies and their needs are multi-dimensional, and solutions must work in a complex set of circumstances. Silo thinking and action won’t work.

This is true whether for neighbourhoods or for country-wide networks such as road systems, railways, or broadband. It is also true for the content, i.e., for broad areas like education from kindergarten to postgraduate levels including vocational training and Continuing-Education for all people, or for a single vertical space, such as health care or hospitality.

The fundamental elements (SOPs) include:

• End-to-end systems, i.e., comprehensive, integrated pieces that fit.
• Convergent objectives.
• Systematic, disciplined project management, starting with the desired end results, and a backward induction for intermediate goals at each step with the required resources and time, all the way back to the start.
• An interdisciplinary/multidisciplinary approach, because sound inputs are required from multiple perspectives, such as overall strategy, structure, systems, technology, human resources, finance, and markets, tailored to our culture and practices, even as we improve them.
• Coordination & Direction: above all, there needs to be convergence of efforts to achieve a desired goal or direction. Without coordination and direction, efforts are unlikely to converge, and therefore unlikely to achieve desired outcomes.

Other Essential Aspects

1. Self-Governing Systems vs Government Intervention
   Much has been made recently of Prof Elinor Ostrom’s ideas on polycentric governance and self-regulation. However, there is insufficient appreciation of and attention to her stress on (a) trust as the most critical attribute, and (b) checks and balances (incentives/penalties) that are “accepted”, as she understates it. Cooperative action is certainly a winner, provided there is an effort by players to build trust, and sound rules are devised and applied impartially. Can you imagine a country-wide highway system or broadband network in the public interest, designed and developed by independent commercial interests? Possible, but unlikely. That’s why governments need to act in the public interest.
2. Allowing for the Non-Rational & Emotional
   Years ago, Carl Sagan popularised the ideas of Paul MacLean, who headed the Laboratory of Brain Evolution and Behavior in America’s National Institute of Mental Health. The concept was of a three-part structure of the brain: the deep down R-complex (for reptilian-complex) where aggression resides, the limbic system which is the seat of emotions, and the neocortex, which is rational and cognitive. While neurology has moved on in the details, e.g., the hippocampus is now apparently thought to be less important in emotions than in MacLean’s view, and the brain may be less simply compartmentalised, the idea of rational man is no longer assumed as a truism.
3. The Normal Curve & Dysfunctional Elements
   For those not familiar with statistics, there is a universal phenomenon of distribution along the “normal” curve: any group of objects (or people) measured for any attribute — height, weight, goodness — is likely to be distributed along a probability curve, as in the graph above, with some outliers spread over the lower and higher ends or “tails”, and the rest bunched around the middle/mean/average.
   

The takeaway: plan for the dysfunctional elements in the left tail, and build protection mechanisms in systems. Consideration with item 2 above indicates the kind of protection robust systems might need.

Good SOPs are a starting point, but there’s more under the surface that will affect results, regardless of external factors.

As the Telecom Regulatory Authority of India (Trai) deliberates on spectrum and licensing after the hearings ending December 2, some important points are worth highlighting. Spectrum is public property and, therefore, need not add a layer of cost (through auctions and such other artifices), provided it is available to network builders, and these networks are available to operators for their customers on payment. The question is whether the government should give spectrum
free, or for an up-front price, i.e., a hefty spectrum fee, or through a progressive revenue-sharing arrangement as for telecommunications. This can be to network builders for their use, or to operators, to pool through either their own arrangements or through network builders-cum-operators.

One way to think about communications networks is to consider an analogy with road networks. The road network is accessed by paying road taxes and special tolls as required, e.g., when using a toll bridge or highway. The rest of the time, once a transport operator pays road taxes, the fleet’s vehicles have access to the entire public road network.

In the same way, it should be feasible for operators to access communications networks. These networks may be the operator’s own, or the public network, i.e., the Public Switched Telephone Network, paying as they go. In other words, whether operators use their own or others’ networks should be immaterial as long as they pay the tariffs, which result from a mesh of interconnection agreements. In this manner, network builders/service providers can use the spectrum as part of their “plant” for wireless transmission, just as they use optical fibre and copper wire for wire-line transmission.

Networks are already being built and operated by network builders-cum-operators. According to The Economist on developments in network operations, initially in New Zealand and then extended on a much larger and broader scale in India, “The vendors... gain economies of scale because they build, run and support networks for several Indian operators. Ericsson’s Mr Svanberg says his firm can run a network with 25% fewer staff than an operator would need. Bharti’s operating expenses are around 15% lower than they would be if it were to build and run its network itself, and its IT costs are around 30% lower, according to Capgemini.”*

Meanwhile, a momentous experiment in spectrum sharing is taking place in America. A company called Spectrum Bridge has developed a database-driven model for dynamic spectrum allocation in unused spectrum bands, the “white space” in the TV bands. This is in the 200 to 600 MHz band, with considerable advantages in propagation over distances, through foliage and walls, without attenuation as experienced at higher frequencies.

This system is being tried in Claudville, a rural community on the border of Virginia and North Carolina. As is likely to be true in rural India, using open spectrum that is unlicensed is impractical because of the distances, terrain and foliage. Fibre and copper lines are not only impractical, but also prohibitively expensive, a fact that people who suggest the use of existing wiring for broadband don’t seem to realise.

In this context, given the discussions on the possibility of spectrum trading as a solution going forward in the Trai hearings, it is instructive to note that despite the US Federal Communication Commission’s secondary market initiatives taken in 2003, not much spectrum trading had actually taken place until Spectrum Bridge’s introduction of their tracking and trading model, SpecEx (see www.specex.com). Subscribers view available spectrum at a chosen location and frequency band with pricing details when they want to buy, or list available spectrum to sell by location and frequency band. Therefore, any recommendations by Trai or decisions by the Empowered Group of Ministers (EGoM) or the government should take this into account in considering the path of market traded spectrum based on exclusively assigned bands. It would be unrealistic to expect such trading to take place simply because it is allowed, without other
facilitating developments as have been achieved by Spectrum Bridge in America.

A second problem is that trading in spectrum can result in effects equivalent to land-grabbing in real estate. This serves less for effective communications than as an asset play for profit.

Like SpecEx for priced spectrum, www.ShowMyWhiteSpace.com is a free website that the company supports to show free TV white space (the “digital dividend” that is talked about) that can be used on the basis of open access to unlicensed or open spectrum.

In the trial at Claudville, Spectrum Bridge deployed the network with Dell and Microsoft contributing computer equipment and software to the local school. Teachers can now incorporate distance learning resources into the school’s curriculum.

Our policy-makers need to move beyond debates over slicing and dicing the spectrum to determine the smallest efficient band — 2.5 MHz for CDMA and 4.4 for GSM? Is 6.2 MHz all that any operator needs?... and so on. A direct solution is to not assign spectrum for exclusive use, and instead enable its use as a common resource that must be accessed by everyone
who needs to communicate on the network, exactly as public roads are accessed by paying road taxes and tolls. If spectrum must be assigned nominally to operators for administrative reasons, they should be obligated to pool this spectrum for common access.

Once we are able to aggregate spectrum in the frequency range which allows propagation over distances and through natural and man-made obstacles — buildings, foliage, etc. — we will have the open “highways” for broadband for its widespread usage across the country. This can only be achieved at relatively low cost through a progressive revenue-sharing arrangement, which is what happened eventually for voice communications with the National Telecom Policy 1999.

These are complex technical and commercial issues, and require the concerted effort of stakeholders and experts to devise the most effective solution in the public interest. The Trai hearings are the first step in this process.

shyamponappa@gmail.com <mailto:shyamponappa@gmail.com>
 
* ‘The mother of invention’, The Economist, September 24, 2009

Link to the original article

India Study Tour Report

2009-10-17 to 2009-11-01
Sagie Chetty, Masters of Management ICT Policy & Regulation
Student Number 0617514V
Supervision: LINK Centre
Graduate School of Public and Development Management
University of the Witwatersrand

Presentations were held at a number of universities, namely the Indian Institute of Technology (IIT), Chennai and IIT, Mumbai; the International Institute of Information Technology (IIIT), Bangalore; and the Indira Gandhi National Open University (IGNOU), the National Institute of Science, Technology and Development Studies (NISTADS) and the Jamia Millia Islamia University – all based in Delhi. The visit concluded with meetings with officials from the Telecoms Regulatory Authority of India (TRAI).

The presentations were well attended and discussions were robust and thought provoking. The South African telecommunications sector was seen as being non-competitive with unnecessarily high ownership by government in the telecommunications sector. From the information provided, students concluded that the SA telecommunications regulator was weak and lacking in the commensurate skills to manage this highly technical sector.

On the other hand, students gravitated between having admiration for India’s own telecommunications regulator, TRAI and criticism of TRAI’s inability to improve broadband take-up in India. Students commended TRAI’s technical skills, independence and its courage in standing up to powerful mobile companies and incumbent telecommunications companies. However, lack of policy direction with regard to broadband rollout is seen as a major failure.  Comments regarding this failure are attributed to TRAI’s driving down of telecommunications prices to levels that do not allow for infrastructure investment. 

The future for broadband in India lies in mobile technology and some predict that fixed line will be defunct by 2025. Some academics also believe that there are too many players in the telecommunications sector in India making spectrum allocation highly competitive and therefore, very expensive. These costs will have to be recovered and the end users will pay dearly for this. Therefore, the model that the Department of Telecommunications (DOT) is using for spectrum auctions is being questioned by students and academics.

The innovation that I observed in India relates to CIS’s early work in projects assisting the visually impaired to read; the writing of 4G standards at the IITs and the innovation with regard to interconnection usage charges (IUC) at TRAI.  These are some of the lessons that I have taken back to South Africa.

My observation of students in India is that they are highly motivated and eager to learn. Entrance to the universities is highly sought after and universities have high standards and are generally difficult to get into. The IITs certainly are increasing the requirements for students to get into them. The institutions are vibrant and are fertile grounds for thought leadership and innovation. India is producing a veritable number of PhDs and institutions seem to offer funding for capable students. South Africa needs to re-examine the funding model for students here. My impression is that students in South Africa do not have similar support as their counterparts in India.

The talks generally concluded with a re-affirmation of the strong historical and cultural links between South Africa and India.  Mahatma Gandhi’s time spent in South Africa developing his notion of non-violent protest is well known in India and will always bind our countries together.

India is a vibrant country with an economic engine that is gathering revolutions. Its future is bright and its institutions are producing bright young minds to take their place in this awakening economic giant. South Africans can do well in learning from this super power in the making.

Report

Videos

Shyam Ponappa November 12, 2009
Distinguished Fellow
Centre for Internet & Society
Bangalore/New Delhi
cis-india.org

shyamponappa@gmail.com

Telecom Regulatory Authority of India
Attn: Sh. Sudhir Gupta, Advisor (MN)
Mahanagar Doorsanchar Bhawan
Jawahar Lal Nehru Marg, New Delhi-110 002
Tel. No.011-23220018 , Fax No.011-23212014

E-mail : advmn@trai.gov.in 

TRAI Consultation paper No. 6/2009- October 16, 2009

"Overall Spectrum Management and review of license terms and conditions"

Sir,

It would help to have a logical framework that defines overall objectives, prioritizes issues, and structures and organizes issues and questions. This would facilitate analysis and response, as we have attempted below.

We begin by responding to Question 57 as a preamble to all the questions:

57. What in your opinion is the desired structure for efficient management of spectrum?
[This question addresses only one of two essential criteria, efficiency. The other criterion is effectiveness; both need equal emphasis.]

Please see separate attachment for answers to Questions 1-56.

Status

Currently, communications services in India comprising Internet, voice and SMS have the following attributes:

1. Low broadband usage, with relatively high prices: eg, direct satellite TV subscriptions at Rs. 200/month, compared with 512 kbps Internet at Rs. 1,000/month.
2. Fragmented spectrum allocation for exclusive use by each operator in a service area.
3. Very high intensity of spectrum use by operators compared with international norms because of constrained availability.
4. Too many operators per service area (11-14 or more [15-16 with all potential operators with GSM and CDMA counted separately], versus the global average of 4-5).

[For details on (2), (3) and (4), please see: 'An assessment of spectrum management policy in India', David Lewin, Val Jervis, Chris Davis, Ken Pearson, Plum Consulting, December 2008
http://www.plumconsulting.co.uk/pdfs/GSMA%20spectrum%20management%20policy%20in%20India.pdf]

Needs

 Our needs are:

• good services for Internet, voice and SMS,
• at reasonable prices, eg, comparable pricing for TV and broadband,
• accessible from/to most households across the country.

The need is especially great in rural areas, as broadband can be the medium for delivery of essential services like education (from basic to advanced to vocational training and Continuing Education at all levels, including high-level professional CE), health (again, from basic diagnostics and monitoring at home, to advanced care at adequately equipped centres), and security and law-and-order services at significantly higher levels than is possible without excellent communications infrastructure.

In view of the above, we suggest that the Government of India consider adopting the following policy goals in the public interest ( and therefore, that where appropriate, the TRAI set these objectives/make appropriate recommendations to the GOI).

Suggested Policy Goals/Objectives [based on needs]

1. Adopt the criteria of long-term net benefits in the public interest for decisions, eschewing short-term cash collections from auctions and fees.
2. An approach to policies for telecommunications services (not for broadcasting) that limits the number of operators per service area in line with international experience, because of the economics of networks.
   [This implies an explicit reversal of prior policies to maximize competition, and requires allowing for consolidation through mergers and acquisitions.]
3. Access to broadband (to be defined as at least 512 kbps in keeping with international norms) at all feasible locations in the country for all users.
4. Develop incentives and penalties favouring good rural service provision, with the emphasis on broadband: an Administered Incentive Pricing mechanism.
5. Explore ways to structure policies to reduce costs/maximize utility through facilities and resource sharing, so that prices can be reduced while maintaining good scope for investment from growth and profits.
   
   This implies two areas of exploration:
   a) Shared use of facilities and equipment/networks;
   b) Shared use of spectrum.
   
   (i) This is best done by collaborative consultations between experts (from the GOI, private sector and academia), operators, equipment providers, and government. Without the requisite interdisciplinary skills combined with operating expertise and investment capability, the effort is too complex for an iterative, serial consultation process.
   (ii) Even within the GOI, this requires interdisciplinary and cross-jurisdictional convergence, both to develop solutions as well as to implement them.
   (iii) This also needs GOI initiatives to invite companies like Ericsson, Nokia, Motorola and Qualcomm as well as Google and Intel, possibly cable companies like Liberty Global, and electricity companies that deliver Internet through their networks.
   (iv) The GOI also needs to depute experienced representatives from various ministries and departments including the WPC, the Defence Services, and specialist agencies such as the DRDO/NTRO.
   [Please see ‘Managing Spectrum’ in the Business Standard November 5, 2009, and related references: http://organizing-india.blogspot.com/2009/11/managing-spectrum.html]
6. Monitor operations online and intervene actively where revenues (the totality of rates/tariffs) are far above total costs, i.e., profits are unreasonable. This is a necessary adjunct to accepting a monopolistic/oligopolistic market structures.

Suggested Approach

The use of a decision tree as in the ‘Issue Map for Spectrum & Broadband’ below (please see Exhibit) facilitates a logical sequence and prioritization in exploring alternatives. (Please note that this is for broadband, voice and SMS, and not for broadcasting.) A similar exploration process for networks and facilities (sharing versus exclusive use for delivery) could follow. However, stakeholders should be free to use any analytical process to improve on this in the common interest.

Once decisions are taken on these two issues (spectrum and network/ facilities sharing), other issues like pricing and consolidation can be logically addressed based on these decisions, probably within the scope of existing laws and regulations.

New regulations or laws should be considered only after comprehensive analysis on the lines of Project LARGE (Legal Adjustments and Reforms for Globalising the Economy by Sh. Bibek Debroy).

 Exhibit: Issue Map on Spectrum & Broadband 

Shyam Ponappa
Centre for Internet & Society
cis-india.org

Attachment – Question 1-57

TRAI Consultation paper No. 6/2009 – October 16, 2009

Overall Spectrum Management and review of license terms and conditions

Chapter 1
Spectrum requirement and availability

1. Do you agree with the subscriber base projections? If not, please provide the reasons for disagreement and your projection estimates along with their basis?
   Do not disagree.
2. Do you agree with the spectrum requirement projected in ¶ 1.7 to ¶1.12? Please give your assessment (service-area wise).
   Agree if exclusive bands of spectrum are used by different operators, and the spectrum requirement is linked to subscribers. Disagree if common use of spectrum is adopted. Please see preamble (reply to Question 57) for details of shared/pooled spectrum approach.
3. How can the spectrum required for Telecommunication purposes and currently available with the Government agencies be re-farmed?
   (a) By rationalizing usage, as advocated in the preamble for commercial operators, by pooling spectrum for common use where possible.
   (b) By inducting equipment that allows more efficient usage and usage of other bands.
4. In view of the policy of technology and service neutrality licences, should any restriction be placed on these bands (800,900 and 1800 MHz) for providing a specific service and secondly, after the expiry of present licences, how will the spectrum in the 800/900 MHz band be assigned to the operators?
   (a) Please see suggestions on shared/pooled spectrum as above.
   (b) In the event that common use of spectrum is infeasible/not accepted by the Government of India, and exclusive bands of spectrum are assigned to operators as is the practice now, work out ways to consolidate fragmented bands (other than through M&A) for operators, to enable operators to hold contiguous bands for greater efficiency, and explore shared use of pooled spectrum.
5. How and when should spectrum in 700 MHz band be allocated between competitive services?
   Preferred method: for common use (can be pooled or shared even if assigned for exclusive use, immediately).

6. What is the impact of digital dividend on 3G and BWA?
   Should extend its reach and access because of lower costs.
   
   Chapter 2
   Licensing Issues

7. Should the spectrum be delinked from the UAS Licence? Please provide the reasons for your response.
   If spectrum is treated as a common resource, the logical requirement is for a linkage that is not dependent on ownership, but to access for service delivery, i.e., common access.
8. In case it is decided not to delink spectrum from UAS license, then should there be a limit on minimum and maximum number of access service providers in a service area? If yes, what should be the number of operators?
   Follow global practice: do not exceed five operators in any service area unless there are compelling reasons to do so.
9. What should be the considerations to determine maximum spectrum per entity?
   Minimum contiguous band for effective rollout and efficient delivery, i.e., inexpensive capital outlay for equipment and towers/network while maintaining Quality of Service.
10. Is there a need to put a limit on the maximum spectrum one licensee can hold? If yes, then what should be the limit? Should operators having more than the maximum limit, if determined, be assigned any more spectrum?
    This depends on the overall approach to spectrum management, i.e., common use, or exclusive use. The logic for a limit is effective delivery capability at ‘normal’ cost. There is no logic for assigning more than this. However, if spectrum is for common/shared use, the only criterion is throughput/capacity.
11. If an existing licensee has more spectrum than the specified limit, then how should this spectrum be treated? Should such spectrum be taken back or should it be subjected to higher charging regime?
    As in No. 10. If common/shared spectrum use is adopted, there needs to be a transition worked out, as in the transition to revenue sharing.
12. In the event fresh licences are to be granted, what should be the Entry fee for the license?
    The principles followed should be:
    (a) Low license fees to minimize access costs.
    (b) Provided licenses are delinked from spectrum and few in number, there need to be strict rollout requirements.
    (c) Incentives for broadband and rural coverage in the form of a structured Administrative Incentive Pricing mechanism.
    (d) Penalties for failure.
13. In case it is decided that the spectrum is to be delinked from the license then what should be the entry fee for such a Licence and should there be any roll out condition?
    As in No. 12.
14. Is there a need to do spectrum audit? If it is found in the audit that an operator is not using the spectrum efficiently what is the suggested course of action? Can penalties be imposed?
    (a) Operating attributes should be monitored online on a continuous basis.
    (b) Spectrum use probably needs to be monitored as an operating attribute.
    (c) Penalties and incentives are needed, including forfeiture for continued transgression.
15. Can spectrum be assigned based on metro, urban and rural areas separately? If yes, what issues do you foresee in this method?
    This needs to be considered only if common/pooled usage is decided against. With common use or sufficiently large blocks/bands of spectrum, no problems are likely to arise.
16. Since the amount of spectrum and the investment required for its utilisation in metro and large cities is higher than in rural areas, can asymmetric pricing of telecom services be a feasible proposition?
    Yes. 
    
    M&A issues
    If the common/shared use approach is adopted, M&A can be under existing laws and regulations.
17. Whether the existing licence conditions and guidelines related to M&A restrict consolidation in the telecom sector? If yes, what should be the alternative framework for M&A in the telecom sector?
18. Whether lock-in clause in UASL agreement is a barrier to consolidation in telecom sector? If yes, what modifications may be considered in the clause to facilitate consolidation?
19. Whether market share in terms of subscriber base/AGR should continue to regulate M&A activity in addition to the restriction on spectrum holding?
20. Whether there should be a transfer charge on spectrum upon merger and acquisition? If yes, whether such charges should be same in case of M&A/transfer/sharing of spectrum?
21. Whether the transfer charges should be one-time only for first such M&A or should they be levied each time an M&A takes place?
22. Whether transfer charges should be levied on the lesser or higher of the 2G spectrum holdings of the merging entities?
23. Whether the spectrum held consequent upon M&A be subjected to a maximum limit?
    
    Spectrum Trading
24. Is spectrum trading required to encourage spectrum consolidation and improve spectrum utilization efficiency?
    
    At present, trading is required to allow consolidation. However, if a comprehensive approach is taken to spectrum use, and especially if common use through common access is established, this set of problems will no longer exist after a transition period. Nor will there be any shortage of spectrum.
25. Who all should be permitted to trade the spectrum ?
    As in No. 24.
26. Should the original allottee who has failed to fulfill “Roll out obligations” be allowed to do spectrum trading?
    There should be penalties and forfeiture for failure to meet rollout obligations, and clawbacks as an interim measure during the transition.
27. Should transfer charges be levied in case of spectrum trading?
28. What should be the parameters and methodology to determine first time spectrum transfer charges payable to Government for trading of the spectrum? How should these charges be determined year after year?
29. Should such capping be limited to 2G spectrum only or consider other bands of spectrum also? Give your suggestions with justification.
    
    This question assumes there is a difference in “2G spectrum” and other spectrum, which is incorrect. The difference is in equipment that has evolved in different phases along different bands. Spectrum should be treated as technology-neutral for the purposes of service delivery. Any service should be deliverable on any band, subject to interference limitations.
30. Should size of minimum tradable block of spectrum be defined or left to the market forces?
31. Should the cost of spectrum trading be more than the spectrum assignment cost?
    
    Spectrum sharing
    
    These questions are addressed in the preamble in the cover note.
32. Should Spectrum sharing be allowed? If yes, what should be the regulatory framework for allowing spectrum sharing among the service providers?
33. What should be criteria to permit spectrum sharing?
34. Should spectrum sharing charges be regulated? If yes then what parameters should be considered to derive spectrum sharing charges? Should such charges be prescribed per MHz or for total allocated spectrum to the entity in LSA?
35. Should there be any preconditions that rollout obligation be fulfilled by one or both service provider before allowing the sharing of spectrum?
36. In case of spectrum sharing, who will have the rollout obligations? Giver or receiver?
    
    Perpetuity of licences
37. Should there be a time limit on licence or should it be perpetual?
38. What should be the validity period of assigned spectrum in case it is delinked from the licence? 20 years, as it exists, or any other period
39. What should be the validity period of spectrum if spectrum is allocated for a different technology under the same license midway during the life of the license?
40. If the spectrum assignment is for a defined period, then for what period and at what price should the extension of assigned spectrum be done?
41. If the spectrum assignment is for a defined period, then after the expiry of the period should the same holder/licensee be given the first priority?
    
    Uniform License Fee
42. What are the advantages and disadvantages of a uniform license fee?
43. Whether there should be a uniform License Fee across all telecom licenses and service areas including services covered under registrations?
44. If introduced, what should be the rate of uniform License Fee?
    
    License fees should be treated as part of the overall scheme of Administered Incentive Pricing.
    
    Chapter 3
    Spectrum assignment
45. If the initial spectrum is de-linked from the licence, then what should be the method for subsequent assignment?
    Please see comments on common/shared use in the preamble in the cover note.
46. If the initial spectrum continues to be linked with licence then is there any need to change from SLC based assignment?
    The SLC basis for spectrum assignment gives rise to many distortions and is not in line with international practices.
47. In case a two-tier mechanism is adopted, then what should be the alternate method and the threshold beyond which it will be implemented?
48. Should the spectrum be assigned in tranches of 1 MHz for GSM technology? What is the optimum tranche for assignment?
49. In case a market based mechanism (i.e. auction) is decided to be adopted, would there be the issue of level playing field amongst licensees who have different amount of spectrum holding? How should this be addressed?
50. In case continuation of SLC criteria is considered appropriate then, what should be the subscriber numbers for assignment of additional spectrum?
51. In your opinion, what should be the method of assigning spectrum in bands other than 800, 900 and 1800 MHz for use other than commercial?
    
    Spectrum pricing
52. Should the service providers having spectrum above the committed threshold be charged a one time charge for the additional spectrum?
53. In case it is decided to levy one time charge beyond a certain amount then what in your opinion should be the date from which the charge should be calculated and why?
54. On what basis, this upfront charge be decided? Should it be benchmarked to the auction price of 3G spectrum or some other benchmark?
55. Should the annual spectrum charges be uniform irrespective of quantum of spectrum and technology?
56. Should there be regular review of spectrum charges? If so, at what interval and what should be the methodology?
    
    Structure for spectrum management
57. What in your opinion is the desired structure for efficient management of spectrum?
    
    Please see the preamble in the cover note.

Shyam Ponappa
Centre for Internet & Society
cis-india.org

November 12, 2009

In communications services, the high demand for spectrum compared with limited supply is well established. The Telecom Regulatory Authority of India (Trai) estimates demand in five years at 580 MHz, with current assignment to commercial operators at about 160 MHz. In this limited amount, fragmented spectrum holdings reduce efficiency, and broadband
growth and availability have been abysmal. Therefore, the policy alternatives evaluated should include ways to maximise utility through conserving resources and facilitating broadband Internet. The Empowered Group of Ministers (EGoM) needs this analysis to make informed decisions. The related issue of maximising utility from facilities, i.e., sharing networks for maximum benefit while conserving capital, thereby resulting in lower prices, likewise deserves serious consideration. For this, they need inputs on the benefits and costs of coordinated policy reform to promote broadband through incentives and penalties.

Having said that, it is for the officials providing support to the EGoM to structure, analyse and prioritise issues and provide the requisite information to facilitate informed decisions on complex choices. This requires appropriate inputs on technology as well. Efforts on all these aspects seem inadequate, with the EGoM being simply not adequately informed.

Trai recently began a consultation process, addressing a host of issues relating to 3G, Broadband Wireless Access (BWA) and licensing. A major deficiency is that no purposive goals and objectives are indicated, nor is there a facilitating logic to the structuring of issues (57 wide-ranging questions, with roughly three weeks for comments).

This is because Trai has posed issues built up over the years in one burst, resulting in the equivalent of a “flash flood”. Instead, structured consultations on discrete sets of questions, as in the indicative example below, are likely to yield better results. However, given where we are — the usual how-far-to-go-in-how-little-time — an organised, logical presentation with relevant inputs would improve the chances of good decisions and outcomes. Here is a suggested road map.

GOALS & OBJECTIVES
The first requirement for the consultation process is clear objectives based on needs. As Trai has not provided this, here are indicative constructs:

Our policies for infrastructure should be in public interest. In communications, these are:

• Ready access anywhere in the country to: (a) good services and (b) at reasonable prices.  
• The services can be thought of as “Broadband Internet” and “Voice and SMS”.

(Note: There are very different objectives for broadcasting, which is outside the scope of these comments.)

DECISION TREES & ISSUE MAPS

A decision tree is an alternative to wading through a welter of unstructured questions, starting with fundamental objectives, using a logical decision map/issue map as a framework (see graphic). This requires judgment in selecting, organising and prioritising issues. The example assumes that the least capital and operating costs (while maintaining high quality) are appropriate criteria for services in public interest.

These decisions will determine how issues of licensing and consolidation/acquisitions pan out. Questions on pricing remain, e.g., per cent of revenues for licences and spectrum charges, and the timing of fees (i.e., cash flow from a fiscal perspective). If the decision is to pool spectrum, there are critical questions on Administered Incentive Pricing. The same principles of concessions and incentives (i.e., subsidies) as for sectors like power and highways need to be applied. Finally, there needs to be rationalisation in non-commercial uses, e.g., governance and defence.

SPECTRUM & NETWORK EFFICIENCY=LOWER COSTS

Given our fragmented spectrum holdings, perceived scarcity and economic efficiencies of limited competition in networks, there is reason to explore an approach to conserving spectrum and consolidating facilities. Spectrum can either be given or licensed for exclusive use in bands to separate operators as is done now, or be made available in large (at least 20 MHz) blocks to all operators for common use. Alternatively, operators can be given incentives to pool licensed spectrum to create a common capacity. The same approach can be explored for networks (facilities that use spectrum); these too can be pooled and shared if individually owned. Operators do this in a limited way, e.g., sharing towers, but pooling can be organised and extended much further.

Ill-considered policies that increase competition for its own sake because of the predominance of doctrinaire “free-market” notions have displaced more appropriate market structures. In India, this has resulted in 12-14 operators per service area, compared with the global average of three-five. The economics of networks favour limits to competition, because networks lend themselves to a limited-player (monopolistic or oligopolistic) market.*

Interestingly, an economist at the US Federal Communications Commission has this to say: “…For what should competition be promoted? Promoting competition for particular services can have major implications for the evolution of regulation and the long-term competitive structure of the industry. Unfortunately, the ‘competition for what?’ question has not received adequate consideration.”**

The benefit of using contiguous bands of spectrum is that costs could be significantly lower for equivalent voice and data capacity because of less advanced technology and less density of towers and equipment. Likewise for shared networks. With competition and good regulation, the likely result is lower costs, both for Broadband Internet and for Voice and SMS.

CONCLUSION

An inter-disciplinary consultation with stakeholders and specialists is essential to consider spectrum and sharing of facilities. Companies like Ericsson, Nokia, Motorola and Qualcomm as well as Google, Intel and possibly cable companies (Liberty Global?) should be invited. The EGoM’s goal should be nothing short of a broadband revolution. We need this for
education and vocational training, health care, governance and economic productivity across the board.

shyamponappa@gmail.com

*A rational spectrum allocation policy, BS, July 2, 2009

** Douglas A Galbi, Senior Economist, US FCC