Saturday, November 3, 2012

Super WiFi & Shared Spectrum

                                         A Time to Start Sharing

Look into sharing spectrum and 'Super Wi-Fi', instead of auctions, refarming and exclusive allocation.

Shyam Ponappa / Nov 01, 2012

Amidst our preoccupation with internal problems of misgovernance, we’re losing track of long-term technical developments elsewhere.  For instance, there’s a buzz about “Super Wi-Fi” technology in other countries that's missing in India. Yet this could make spectrum abundant, while avoiding the problems of private allocation.  Here’s why India with its floundering, beleaguered telecommunications sector should stay abreast.

Super Wi-Fi using TVWS

The technology for unused TV spectrum bands, or TV White Spaces (TVWS), is referred to as “Super Wi-Fi”, although it doesn’t conform to earlier Wi-Fi standards, nor does it use the 2.4 GHz or 5 GHz licence-exempt spectrum. Super Wi-Fi has its own standards (IEEE 802.22, and 802.11 af in draft form) using 470-810 MHz, the “digital dividend” after conversion from analogue broadcast TV. It can be used for long-range rural broadband, and to improve short-range coverage. In the US, where it was pioneered, access is available without a licence to devices registered with a proximate geolocational database. Like regular Wi-Fi, Super Wi-Fi expands the use of available spectrum by sharing access.

The US is also permitting exploration of shared use in defence and security bands from 1755-1850 MHz, extending the potential for sharing spectrum.1

TVWS trials

Earlier this month, Singapore’s Infocomm Development Authority organised a workshop on TVWS with government and private entities. Organisers included the Institute for Infocomm Research (I2R) and other local participants, with presentations from companies from the US, Europe, Asia, and Africa, including Microsoft, Google, Spectrum Bridge, Adaptrum, Ericsson subsidiary Telcordia, Neul, Japan’s NCIT, and so on.2 Completed or ongoing trials are shown in the diagram below.

Super WiFi Trials

TVWS and shared spectrum vs refarming

These developments should be of vital interest in India to policy makers, operators and users — not only for TVWS as a shared resource, but as an approach that could be extended to other bands, so that limited spectrum availability doesn’t constrain reasonably priced, high-speed data services. This is a serious limitation in India, unlike in other countries where a few operators have sufficient spectrum; in this sense, the need to share spectrum is much greater in India. For example, sharing could provide a better alternative to refarming of the 900 MHz band, allowing for both 4G and legacy uses.

One difficulty is that dominant operators may oppose sharing because their spectrum holdings provide a competitive advantage: possibly Bharti, Reliance, or an aspirant like Vodafone with access to inexpensive offshore funding. Our collective interests here, however, are likely to be best served not by constraining access through limited, exclusive spectrum, but by making spectrum abundant through sharing, allowing for wide bands (2x20 MHz or 2x40 MHz) that can be aggregated for much higher throughput for data, not just for voice.

For this to happen, (a) the government has to explore spectrum sharing in TVWS as well as in other bands, and (b) stakeholders must be receptive, to co-operate effectively on a workable plan on the lines of revenue sharing after NTP-99, extending to broadband delivery. Everyone will gain: users will get better access, operators can thrive, and the government will collect much more revenue over time. However, dominant operators will need to give up their spectrum for the greater common interest including their own, and for this, they will need compensation — as in production-sharing agreements in the oil sector.

The advantages of spectrum sharing

There are a number of advantages of sharing spectrum. First, and important, it can be non-discriminatory. Second, it avoids private allocation; shared spectrum can be accessed without allocation to private parties. Then there is the fact that capital cost is reduced. There is no deadweight loss from capital tied up in auctions, freeing it all for network development and service delivery. Finally, there’s the general misinformation about auctions, which become academic if spectrum is shared. If spectrum is instead auctioned, the public interest – of users, operators and producers – will be adversely affected. (On producers, while local manufacturing is currently insignificant, there is considerable scope if it is set up right, as telecommunications equipment imports are expected to exceed energy imports in a few years.)

The Supreme Court was misinformed about auction fees exceeding revenue-sharing collections, and not informed of its detriments. As evidenced after NTP-99, networks and services proliferated, resulting in much higher collections than auction fees foregone. The Supreme Court’s opinion on the Presidential Reference clarified that auctions were not mandatory for other resources, but not for spectrum, although the reasoning is the same. This needs rectification if spectrum is not shared, because revenue-share collections and tax revenues on profits from more extensive networks and services are likely to far exceed the estimated auction fees of Rs 40,000 crore over three years, quite apart from the major public benefits of access.

Space for constructive decisions

Another requirement for constructive resolution is that policy makers be given the requisite space to frame solutions that are genuinely in the public interest. These solutions can be premised on abundance if it is possible, rather than artificial scarcity and rationing. At present, the Telecom Regulatory Authority of India, the Department of Telecommunications, and other authorities including the Empowered Group of Ministers are under immense pressure to favour aggressive government collections, instead of what might be genuinely beneficial. This is an odd consequence of the government’s increasing loss of credibility, resulting in the rise of populists, “profit haters”, and ignorant-yet-opinionated sceptics. Uninformed attacks on constructive approaches and alternatives need to be presented and seen in a more balanced way by an informed media, press and public, instead of being fuelled by indiscriminate hype.

Also, we have to learn to distinguish between problems of ideological conviction – those that can be solved through political accommodation – and engineering problems, like network design and service delivery at least cost. Resorting to political accommodation for engineering requirements results in malfunction and/or collapse.

A good way to proceed is to ensure sharing solutions are worked out without incurring exorbitant cost — not only for TVWS but also for legacy operations, such as in the 900 MHz band. These can induce new network build-outs for data services in urban as well as underserved rural areas, and broadband service delivery across the country.

shyamponappa at gmail dot com 

1: Other bands being considered for sharing in the US are:
1695-1710 MHz & 3550-3650 MHz; Unlicensed: 5350-5470 MHz & 5850-5925 MHz.  For details, see:


Details on the UK (Cambridge) trials at: White Spaces Trial - technical findings-with higher res pics.pdf