Thursday, May 5, 2011

Spectrum Reforms: Good & Bad News




[with additional diagrams + last two bullet points]

A good initiative is under way, but needs changes to work out complex issues

Shyam Ponappa / May 5, 2011


There’s some good news, and yes, some bad news… The good news is that momentous developments are under way in spectrum and telecom policy:

  • The Ministry of Communications & Information Technology held consultations with service providers, then posted the transcript on the Department of Telecommunications (DOT) website.
  • The Wireless Planning & Coordination Wing (WPC) disclosed data on all commercial spectrum allocations – frequencies allotted by geography and service provider or operator – on its website.

Terrific first steps in a constructive approach. There’s more: the ministry’s report of 100 days states: “We will hold consultations with key stakeholders to evolve a clear and transparent regime covering licensing, spectrum allocation, tariffs or pricing, linkage with roll out performance, flexibility within licenses, spectrum sharing, spectrum trading, MVNOs, unlicensed bands, M&A, etc, in a technology agnostic environment after due consideration of Trai recommendations in this regard. Interest of the ‘aam aadmi’ would be the prime consideration.” That’s comprehensive alright, which is good, though the ‘aam aadmi’ bit is either confused or manipulative. Elected governments should act in the public interest, no more, no less. While the private sector is exhorted not to play games, the government at all levels – politicians, administrators and agencies – must also focus on results, and avoid populism.


Display & Presentation

The presentation of information could be more effective for the patterns and structure to be easily accessible. The WPC display is of voluminous raw data. There is no overview, with the ability to drill down to details, nor to aggregate details by operator or frequency. The full set runs into 32 pages of tables (Figure 1).

Compare this with a display in colour from the US’ National Telecommunications and Information Administration (Figure 2). Similar information from the WPC runs into many pages.

Figure 2: Fragment of Allocation Chart (USA)

Source: http://www.ntia.doc.gov/osmhome/allochrt.pdf

However, the US display contains not as much detail, and has no interactive capabilities (these are possible extensions). For an interactive graphical interface, consider the “market map” by Moneycontrol.com for stocks (Figure 3, left).

One can drill down in any sector by clicking on the rectangle. For example, “Telecommunication”, which opens a map with the listed companies, each colour-coded to reflect more detail (green for gains, red for losses).

Clicking on a company shows its daily price and volume chart (Figure 3, right). In a variant (at Smartmoney.com), it opens a menu with access to details like news, financials and so on. Similar spectrum displays could show, for example, information by operator for network rollout and subscribers by frequency.

Figure 3: Market Map of Stocks (Sectors) Companies

Source: http://www.moneycontrol.com/mcplus/marketmap/nse/marketmap.php

An alternate display format is the “Topics most commented on” on The Economist's website.

When the cursor hovers on a topic, related comments are displayed. Clicking on a topic realigns the clusters based on content around that topic, as for India in Figure 4.


Figure 4:

This would work well for aggregating comments on related issues in the consultation transcripts.

Imagine what such a graphical interface to a relational database could do for effectiveness and transparency in spectrum policy. It could be extended to telecom and broadband next, and, eventually, to all of government.


The Bad News: Process Limitations

Judging from news reports, process inadequacies might render the ministry’s grand intentions unachievable. The following examples show why.

- Spectrum sharing is an obvious solution for high demand with limited supply. The DoT has reportedly considered it for years, but discussions so far have been superficial and on “excess spectrum”. Also, the statements of intent on sharing or trading are confusing. “Spectrum trading” implies exclusive rights to spectrum, unless otherwise specified. “Spectrum sharing” means aggregating spectrum for redeployment, with Dynamic Spectrum Allocation. This is analogous to “common carrier access” and “big pipes” for railways, roads, oil pipelines, or airways. Therefore, from a policy perspective, spectrum sharing and spectrum trading are mutually exclusive.

Spectrum and airways or flight paths coexist in the atmosphere. Imagine if airways were auctioned to each airline for its exclusive use, instead of being available to all airlines for similar aircraft through Air Traffic Control. That’s what we have with spectrum auctions in communications. The logic for spectrum auctions is based on old technology with no allowances for improvements in managing interference in the last 60-70 years. Also, allocating spectrum in this way means that aggregate capacity is constrained for two reasons. One is that each operator uses only part of allotted capacity. A study in Singapore in 2008 found that only two bands had a utilisation rate of 50 per cent; the overall utilisation rate for 80-5,850 MHz was about five per cent (Figure 5).

Figure 5: Average Occupancy of Frequency Bands in Singapore


Source: http://www.pwtc.eee.ntu.edu.sg/News/Documents/Spectrum%20survey%20in%20Singapore_%20Occupancy%20measurements%20and%20analyses.pdf

Second, a large band provides much greater capacity than the sum of smaller bands.

Our spectrum predicament arises primarily from inappropriate allocation policies. Therefore, forward-looking policies need the incorporation of a technical understanding of spectrum occupancy, of the effects of spectrum aggregation versus fragmentation, and of technologies like multiple antenna effects (multiple-input and multiple-output, or MIMO), which enable more effective spectrum use and improve functional attributes of higher frequencies. A backward-looking audit of historical data will not serve these purposes.

- Another damaging effect is the move to extract spectrum from Defence to auction to the private sector. The rationale apparently is the high revenues the government can collect. This cannot be in the public interest, especially since the alternate optical fibre network to have been built by BSNL is still not ready.

- Decisions on issues like the desirable number of operators per circle need an objective rationale. No data have been offered contrary to the UK Ofcom’s findings of maximum welfare at three to four operators.

An inherent limitation of the consultation-and-pronouncement approach (as opposed to a collaborative-stakeholder-workout) is that external expertise in technology and process consultation, sorely needed in India, has to be brought in only by the government. This must be done before formulating new policies, because the issues are too complex to resolve without objective expertise.



shyamponappa at g mail dot com

No comments: