Archive for the 'Policy' Category

3G: it doesn't really matter – does it?

February 17, 2010

About the auction to rollout 3G services in India:

Cabinet secretary K M Chandrasekhar today indicated that all the outstanding issues pertaining to the 3G spectrum auction had been resolved, however, the auction would happen at a time when maximum revenue could be generated from it.

“I don’t think it really particularly matters if it happens this year or next year… they (DoT and the finance ministry) will take the decision based on how the revenue can be maximised,” cabinet secretary K M Chandrasekhar told reporters…

He said the issues that delayed the auction process in January had been resolved but refused to give any definite timeline for the auction. He also did not specify the issues.

I don’t think the secretary for a moment considered actual subscribers who, of course have been waiting for the best part of five years.

Update:
Delicious, dark irony. The same government that thinks nothing of losing several thousand crore rupees a year on subsidies that have been repeatedly shown to be ineffective thinks nothing of denying 3G services to its people until it can ‘maximize revenues’.

The problem with free wifi in the U.S. – and how not to solve it

September 9, 2007
Readers of John Dvorak’s column on MarketWatch are going to be astounded by some extremely poor reasoning this week. Dvorak talks about why municipal wifi is turning out to be a pipe dream in most American cities.

Dvorak starts off well enough, alleging that vested corporate interests, especially those of commercial ISPs AT&T and Comcast, are successfully sabotaging attempts by local governments to implement free community wifi.

The argument falls apart when Dvorak lists common arguments made against municipal wifi,and how they ought to be countered.

Argument 1: It’s not good enough. It’s not fast enough.You’ll have to pay more for higher speeds and the lower speeds are unusable except for e-mail and Web surfing.
Dvorak’s suggested counter-argument: “Since 99% of the usage would be for email and web surfing, I wonder how is this not fast enough? Not fast enough for what? All of the systems I’ve seen are plenty fast.”

Argument 2: It stifles competition. Nobody else will be able to compete in a market where something is free.
Dvorak’s suggested counter-argument: “These are the same people that say governments can’t do anything right. But apparently they do this right.”

The last one takes the cake, though.

Argument 3: Government shouldn’t be in the connectivity business.
Dvorak’s suggested counter-argument: “Another standalone bromide. Why shouldn’t it be in this business? Nobody knows.”

It’s fashionable in India in these neosocialist times to spew invective at private service providers and advocate a larger role for Government, but coming from an American, it’s surprising (although I’m not sure if Dvorak is known to have Leftist leanings).

Dvorak observes the symptom corectly, but misreads the problem, and therefore advocates a horrendously wrong solution.

Symptom: Municipal (community) wifi is dying
Dvorak’s reading of the problem: powerful commercial ISPs are derailing government’s efforts
Dvorak’s solution: greater role for Government in providing community wifi (possibly through public money)

The actual problem: there are powerful commercial interests in the first place
A possible solution: lower entry barriers to local wifi providers. Since, as Dvorak mentions, there are only two dominant ISPs, they have enough bargaining power at both the national and local level to block other entrants. The U.S. has had a history of being plagued by vested corporate interests (the alternative energy sector in the U.S. has been chronically starved of funding, not least because of subtle but powerful opposition from Big Oil).

Finally, mandating government to play ISP is an invitation to inefficient use of public money, regulation skewed in favor of the government-owned ISP (think MTNL/BSNL in India), and a sick, wheezing, slow network that people will eventually tire of. It will probably lead them to paying for reliable connectivity from AT&T and Comcast anyway.

A striking parallel to this is the Maharashtra government’s insistence on providing farmers with free electrcity and water – another example of government getting into the business of providing public services. Farmers are by-and-large so fed up with how ridiculously unreliable these utilities are, they’d much rather pay for guranteed supply.

Back in 2005, I wrote about why the government should not be in the business of running business. The Indian pseudo-socialist government, of course, finds it extremely uncomfortable acknowledging that this school of thought even exists. But it looks like two years later, the American Government needs to read that article – more than ever.

When Telecom & IT babus just don't get the Market

May 17, 2007

For all who go to town declaring that India has an entrepreneur-friendly, liberalized Telecom sector, here’s a dampener. TRAI’s recommendations on “ Review of Internet Services“, a report dated May 10th 2007, show just how much (or little) babus in the ministry understand the Internet. We’re light-years away from a truly liberal Telecom policy, because the DoT hasn’t even grasped the concept of creating a free, healthy market.

After having read through the report, I think I have a better understanding of what ails the DoT (and by extension, most ministries at the Center and the states). The Telecom ministry thinks it has to juggle different objectives which, in its view, are mutually incompatible. Therefore, to achieve all of these objectives, each stakeholder has to compromise to some degree. In reality, though, these objectives are _not_ incompatible, unnecessary compromised are made, and it results in a policy full of caveats, which ends up pleasing no one.

There are plenty of examples of these misplaced assumptions in the report. To demonstrate, here are a few sentences from the voluminous, bloated 126-page report, with my comments inline, in italics.

Under the “Scope of ISP license” section:

Web hosting by certain foreign companies within Indian domain who have significant market share in global market should form part of our developmental agenda and necessary policies to encourage such web hosting need to be evolved.

How in the world can “foreign companies” with “significant market share” hosting their data here going to help our “developmental agenda”?!

Strong views were expressed to permit IPTV under ambit of ISP license as it has potential to drive market, easy to provide using ISP backbone and can encourage Internet penetration .

So is this – “encouraging Internet penetration” – going to be the compass by which permission for other services will be granted in the future? What about TV on mobile phones? This is an extremely niche market, but does not in any way increase Internet penetration. Will this be allowed then?

Under the “Grey Market Operations” (?!) section:

Some entities located abroad are offering unauthorized Internet telephony services in our country for making calls to and from abroad on Public Switched Telephone Network (PSTN) and Public Land Mobile Network (PLMN) ….The licensing, legal and technological issues arising from such services need to be urgently addressed by DoT.

Some companies are also providing Software through their websites that enables the user to have free chat with anyone using the similar software anywhere in the world while logged on to the server of such service providers. Software can be downloaded free of charge from their website.

The Internet telephony call through such unregistered entities using PCs/IP access devices in India to landline or mobile phones abroad and vice versa result in a revenue loss to the government . Such calls can escape the eyes of law enforcement agencies also.

Here is a classic case of the DoT’s misplaced priorities. Is the objective of TRAI to “increase Internet penetration” and ensure that the Indian consumer gets the best and cheapest service, or is it to ensure that the Government – comprising the DoT, BSNL and MTNL – earn the maximum possible revenue? The DoT’s most glaring – even criminal – policy failure is to disallow such an interconnect. If the DoT has to stifle innovation and consumer benefit in favor of PSU revenues, we must then question the very reason for the existence of these PSUs. Indeed, in an article I wrote back in December 2005,

If the state is not running telecom companies for “social service”, or it put it more correctly, “social benefit”, then what are they running it for? If profit is the only motive, then the company should be privatized, fully, and right away. Think about it: If there is no “larger good” for running BSNL and MTNL, their motives are just like any other private company. Why, then, should they be at such a huge advantage compared to a private player?

By escaping regulatory levies such unlicensed foreign entities are able to provide cheaper services to lure the subscribers. The Authority is aware that large number of Indians are availing such services. Stopping access to such services is technically difficult. There is serious revenue implication for the government. DoT may address this development on priority. One of the possible options could be to ask such companies to register in India, seek permission from DoT and host their website in India.

Again, the ridiculous premise that “hosting their website in India” is going to guarantee compliance! Besides, why should “cheap services” have such negative connotations as to invite use of the term “lure the subscribers”? And yes, the issue of “serious revenue implication for the Government” raises its ugly head again.

Under the “Revamping and Restructuring of Internet Services” section:

It is extremely important to regularly analyse the business data of the operational ISPs to ensure that the licensed ISPs are contributing to the growth of Internet and engage in legitimate business.

What is “legitimate” and “illegitimate” business? What “business data” will help identify such legitimacy? Should ISPs contribute to the “growth of Internet”? What if an ISP grows in an area only by managing to convert users of competing ISPs to its services? It has been a resounding success, but has not resulted in a single extra subscriber. Is that legal or illegal?

The Authority therefore favours a uniform FDI cap /equity of 74% across all telecom licenses. ISPs who have more than 74% FDI cap /equity at present shall be required to bring down their FDI cap/ equity to 74% within two years.

What is the magic number of 74% supposed to signify? Why not 65%? Why not 84%? As long as the foreign entity has a controlling stake in a firm, any additional restrictions on its holding are meaningless. If 50% FDI is allowed, why not 100%? In fact, what is any restriction at all on FDI going to achieve in a field like telecom?

Annexure II – Recommendations of “stakeholders”:

Stakeholders also felt that the ISP licensing should only be linked with the vanilla bandwidth provisioning whereas all other services based on video & voice applications, www-hosted applications should be freed from licensing. It was also stressed upon that in a multi-tasking, multi-function, convergent nature of Internet; it would be illogical to consider regulating isolated applications.

Ah. Finally something which makes sense. But…

However, it does not appear possible to expand the scope of ISP license to cover all services as it will infringe on the rights granted to ILD/NLD/ Access Providers. Therefore, for the ISPs to move up the value chain, there is no option but to obtain one or more licenses as per the services planned to be offered by them.

There – senseless restrictions again. The same issues with interfacing with PSTN that we dealt with earlier.

Stakeholder commented that since Skype / Google type service providers are not licensed to provide such services in India without having facility for lawful interception, therefore, the vigilance and monitoring efforts are required to be beefed up, as these applications not only bypass the laws and regulations of the land, but also pose a threat to security. As such these services should be blocked.

I have nothing to comment on this, except that I would dearly like to know who this particular “stakeholder” is, and that I’d like to hear his views on how Google Talk poses a “threat to security”.

Finally, the US has the most liberal laws I’ve seen:

In US, ISPs do not require license or authorization. Instead, e-mail, data and Internet services are treated as “information services,” and ISPs are permitted to operate unfettered in a competitive and free market, subject only, with a few limited exceptions, to general business laws.

Ah, this is an entrepreneur-friendly policy. Here is a Government that believes in leaving anyone who wishes to offer services “operate unfettered”, and step back from the scene, as opposed to one that gets involved in grand-scale centralized planning to the extent of mandating the “target for broadband users in India” in a particular year!

The Paradox of India's Telecom Growth Story…

July 29, 2006

… is that while the country is the world’s second-fastest growing telecom markets (an astounding 5 million new connections per month), its adoption of high-end mobile technology is far behind the rest of the world. Here are two contrasting articles from BusinessWeek and The Financial Express, that highlight the growing divide between growth at the low end and the high end.

Part of the problem lies in incomplete reforms in the telecom sector: New Delhi has allowed private players to offer both fixed line and wireless services, and has recently permitted upto 74% foreign ownership of telecom companies. That has resulted in the world’s cheapest call rates – 2 cents per minute in India, as compared to 33 cents in Japan, 11 cents in Brazil, and 24 cents in Australia. At the same time, there appears to be an inexplicable delay in policy formulation and spectrum allocation for high-end 3G communications technology. As the Financial Express article points out, other economies are building high-speed wireless networks, entrepreneurs are leveraging this to construct new business models, and exciting new services (Mobile TV, mobile gaming) are becoming widely available. India risks being left behind with only voice and text-based services.

The other part of the problem might be India’s own spectacular growth. At this point in time, there is very little incentive for India’s telecom providers to provide 3G services: the existing (and potential) voice market is too compelling in terms of sheer numbers. In the foreseeable future, it is rural India which will drive these growth numbers. It is a high-volume, razor-thin margins game. Urban markets are fairly saturated now. Telecom service providers,then, are better off investing in setting up new network infrastructure, bringing more of the country under wireless coverage (only 30% today), than in high-end services.

Dithering on 3G policy by the Department of Telecom suits India’s telecom companies just fine -for now. There’ll be a time when these growth figures won’t be so rosy -and that’s when telcos will feel the pinch from lack of these value-added services to add to their margins.

Thoughts on mobile data access in India.

April 26, 2006

It’s expensive:
I use the Airtel GPRS service very heavily. Because in Pune, it’s unlimited (in terms of volume and time) as long as I pay Rs. 150 per month. I use it for checking email, catching up on blogs, and using it as a Bluetooth modem for my Thinkpad so I can dial into my corporate network. It’s cheap, I’m connected, and I’m happy.

However, I’ve learnt that almost all other GPRS plans (even Airtel in other locations) charge based on volume, typically a paisa per kilobyte. That works out to Rs. 10 per MB. And I think that’s prohibitively expensive, something like twice the average broadband tariffs. Besides, the speeds aren’t anywhere like broadband, in fact far closer to the dialup speeds circa 1998. In addition, the mobile phone doesn’t offer you the user interface that a PC can, so what would compel a user to sign up for a scheme like this?

This is the same problem that’s hampering the widespread adoption of broadband in India. I had referred to this once before , and I have personally seen plenty of families I know, who haven’t signed up for broadband because it’s “expensive”, or won’t fit into their exact usage pattern. The same is set to happen with the mobile date market in India.

We need to remember that the mobile phone is to India what the PC is to the United States – the most widely used medium for data access. There is a tremendous market for mobile data services. We need to stop thinking according to the scarcity mentality – that is, to try to extract the maximum revenue from a small market, and begin to take bold steps to expand the market. There is huge opportunity in the latter.

Best phone for the Indian Market?
Smartphones such as the Nokia 6600 are cheap today, and will be real cheap in the near future. I am of the firm opinion that it’s going to be this model that can be a game-changer as regards hardware. Once this piece reaches Rs. 5000, once there is a critial mass of people using it, mobile data usage will explode. This phone can do most things that a rich data experience needs – for connectivity, we need Bluetooth, USB – this phone has it. RealPlayer for videos, an MP3 player, FM radio (although I think this isn’t stereo output), decent amount of storage, document viewers, Java – the works. It even has a camera, but this is  exceptionally poor and isn’t any real use. But a little amount of tweaking can make this a dream phone. It could be the iPod of the masses, the Simputer of the masses, and the PC of the masses too.

Mobile Applications:
What kind of applications would people use? For one, we need a kickass web browser for the mobile phone platform. I’d expected the Open Source Community to put together a mobile Firefox for at least the Blackberry or an O2 or a Treo (since that would probably be easier – more computing power, more disk space, more memory) than a Nokia, but then Opera beat everyone to it with the Opera Mini. That simple application has the potential of being a total game-changer. It runs on any phone that can run Java apps – any phone! Also a similar Java-based, small footprint Instant Messaging application – something like Migg33. Another thing would be a service like iTunes. We have a Jurassic version of that with Airtel’s Easy Music service. But that requires you to walk into an Airtel outlet. What we need is true download-via-GPRS, just like iTunes. With more and more phones having lots of storage, a file management application that lets you use your phone as a portable drive would be real cool. Again, it needs to be Java-based so that the interface would be the same no matter what phone it was installed on.

So it’s clear that we have the building blocks in place. But for the market to really take off, it needs a big gamble from a player who’s willing to change the rules of the game by making sustained investments for some amount of time. That is what Reliance did in the early 2000s, and today it can afford a 40 paise per minute tarriff within the Reliance network. Right now I would think only Reliance and Bharti are in a position to make that kind of investment.

What ails unlimited Broadband in India?

February 25, 2006

Here’s something I’m unable to figure out – why are Broadband tariffs in India still metered either by time or by data transfer? What is preventing providers from launching fixed-rate plans on a monthly basis?

Is it that such a plan requires economy of scale, or a certain critical mass of customers to become viable, and that mass is not there in India yet? Such a case would be a bit of a chicken-and-egg situation, where subscribers won’t take up broadband plans because they aren’t attractive (either price-wise or feature-wise), and cable companies/telcos won’t provide these plans until there are enough subscribers. I think some vendor as large as Reliance Infocomm is in the best position to absorb short-term losses (which I’m sure they did for Reliance India Mobile) in return for a large (and captive) chunk of a new market.

And there is definitely a large market. In 2005, the number of Internet connections in India was 50.5 million (pathetically low, yes). But even in August 2005, the number of broadband connections were about 0.5 million! As opposed to China with 28 million, Japan with about the same, and the US with 42 million! Even tiny South Korea has well over 10 million (and broadband penetration is well over 70%!).

Is it that there is too much effort and risk involved in laying the infrastructure for providing Broadband to enough homes to reach that critical mass that’ll allow vendors to manage large enough margins with unlimited broadband? Given the state of roads everywhere in India, where your cables could be dug up, damaged, cut by the municipal corporation during road repairs, before you can say Jack Robinson, there isn’t too much incentive to lay cables,especially when they’re fibre-optic ones. Perhaps wireless could be a wonderfully clean alternative, but as I’ve mentioned in the past, it can be more messy than its worth.

Or is it that providers are used to reaping large margins on the (albeit small) base of broadband users in India, and would be loath to see those margins fall? Such a case would be sheer shortsightedness – trying to squeeze more out of a small market instead of attempting to expand the market.

3G v/s WiFi – Impending war between telcos and ISPs?

January 18, 2006

This announcement by Lenovo to include Cingular’s 3G implementation with its notebooks throws up some interesting questions. Right now, 3G only supports data transfer speeds of about 400-700kbps, and it is primarily used in high-end mobile smartphones. However, as this standard evolves, and it is able to support transfer speeds in Mbps, it’ll begin to compete with WiFi itself.

What that means is that telecom companies will find it more and more attractive to offer internet services through 3G, and not limit it to mobile subscribers. In fact, if a company such as Airtel (a major telecom company in India) were to start offering an “Airtel 3G connection” to virtually anyone who was willing to pay for internet access, it could end up becoming an ISP in itself. That’s where the competition heats up.

It’d be a cinch for Airtel to start offering wide-range wireless Internet services, for three reasons:
1.) Airtel could utilise its existing towers/current mobile infrastructure.
2.) Even if a separate parallel infrastructure were to be set up, the maintenance cost of a wireless network would be extremely low, even though the up-front investment would probably be higher when compared to cable.
3.) Existing subscribers could become ready-made, at-your-doorstep customers for Airtel. They could have an “Airtel account”, and access the internet either from their mobile phone, or their desktop/laptop, once they get a 3G network card.
4.) ISPs have already sunk in a huge amount of money into their wired network infrastructure, and would be loathe to cast that aside and setup a fresh wireless network.

What could go wrong?
1.) Simple – 3G fails to live up to its hype, cannot support high-enough data speeds.
2.) The vast majority of mobile phones begin to offer WiFi. In that case, it would be the ISPs who could start offering voice services, in addition to data, instead of the other way round.

Which way we go will depend on who builds better relationships with device manufacturers. The tipping point, in my opinion, will be mobile devices. Whoever (a telco or an ISP), does a better job of convincing mobile handset manufacturers to build 3G/WiFi into their most significant models, and then offering the best access plan.

We are now taking the first baby steps towards True Unification – the convergence of mobile devices (phones/consumer electronics) and notebooks/desktops. Whether the underlying infrastructure is going to be WiFi ( 802.11b/g/n) or 3G (WCDMA/EDGE), remains to be seen. Again, in any case, the real winners will be those companies which offer the best “unified” product-services. Google Talk is the prime contender at this moment.

When the state is in the business of running business.

December 11, 2005

India’s biggest problem today…
… is an interventionist State. Despite a brief start at reforms between 1991-93, the Government continues to be in the business of running businesses. Inevitably, when it is also the policy-maker, protectionist policies are bound to arise, and private players are bound to be put at a perennial disadvantage. Consider the Government’s decision on the 30th of November to deny private operators use of BSNL’s and MTNL’s “last mile” copper wire infrastructure. From the Economic Times:

On unbundling of last mile access of BSNL and MTNL Mr. Maran said the government wants private operators to build their own networks.

“If the state is not in the business of telecom, then their infrastructure can be shared and given to private players but in this case, the state is in the business and they are not in a social service segment,” said Mr. Maran at the Indian Economic Summit.

He also said both MTNL and BSNL are doing very well and there is no necessity to unbundle. “BSNL has done 2.5 lakh and MTNL one lakh in terms of broadband connections and from the peak price of Rs 5,000, today BSNL and MTNL’s connections are available at Rs 250 and Rs 199 respectively,” said Mr. Maran.

Reliance, Bharti and Tata Teleservices want DoT to allow them to use BSNL and MTNL’s last mile network for providing broadband services. BSNL and MTNL account for more than 90% of India’s fixed line subscriber base.

Maran’s quote says it all. If the state is not running telecom companies for “social service”, or it put it more correctly, “social benefit”, then what are they running it for? If profit is the only motive, then the company should be privatized, fully, and right away. Think about it: If there is no “larger good” for running BSNL and MTNL, their motives are just like any other private company. Why, then, should they be at such a huge advantage compared to a private player? First, they have access to tremendous funds (taxpayers’ money) which are not the result of revenue earned by the company. That’s what’s enabled them to build up this vast copper wire “last mile” network in the first place. Second, the parent ministry frames policy for the sector. This is almost always designed to protect the state-owned enterprises in the sector, which is but natural. But is it fair?

For instance, consider the Access Deficit Charge. This is a levy that private telecom operators have to pay when their subscribers make long-distance calls either to fixed-line telephones or to other mobile phones – basically, for any traffic that makes use of fixed-line long-distance infrastructure – and 95% of these revenues go to BSNL**, since it has a near-monopoly on fixed-line infrastructure in the country. Private operators have often suspected that the Access Deficit Charge is used to subsidize BSNL’s own cellular business. Or its International calls – just before TRAI ruled in late 2003 that all telecom operators would have to pay ADC (as opposed to only basic service providers earlier), BSNL cut its ILD (International Long Distance) call rates from Rs. 24 to Rs. 7.20 per minute***.

There is only one reason for the state to be present in any sector – for reasons of “National Security” or “Strategic Importance”. Examples of this include education and defence. The other reason offered by the Indian Government frequently is for “social benefit”. The implication here is that if the state was to exit that sector (from the business point of view), private enterprise would not run services in unprofitable (usually rural) areas, and that this would lead to unequal development.

I do not buy this argument. There are ways for the Government to ensure equal development, or at least prevent creation of a yawning gap between urban and rural areas. One of those ways is for the Government to focus exclusively on running services in non-profitable areas. The funding for such ventures would be generated using, most ironically, the Access Deficit Charge! Of course, it would be called something else!

A Solution:

Here’s how it could work: First and foremost, the Government must privatize both BSNL and MTNL, so that all telecom operations in the country are privately-owned, and there is a level playing field. It must then set up a fresh, fully autonomous company to provide services only to clearly demarcated “unviable areas”. As long as private operators and the State do not encroach into each others’ territories, there is no problem with levying a charge on private operators to fund the state’s efforts to improve connectivity in such rural or unviable areas. This “ADC” ought to be a fixed charge imposed upon every operator, regardless of their size or revenue. Since this ADC is now not based on the access of the fixed-line infrastructure owned by the state, by private operators, it would be unfair to charge the larger player more. (How the now-privatized BSNL charges other operators for use of its network is a completely different matter, and is out of the scope of this article). The only revenue for this company ought to be from the charges imposed on private operators, and 100% of the revenue from these operations must be pooled back into the company’s operations. No profits. In addition, if a private operator wishes to provide operations in a particular “unviable area”, it is free to do so, provided it satisfies certain minimum criteria for connectivity in that area.

However, there has to be a clear maximum time in which the Government must bring rural areas up to speed with urban areas in terms of communication infrastructure, hand over operations to the highest bidder among private operators. If, within this time, a private operator wishes to bid for the Government’s operations in an unviable area, it is free to do so. The more “unviable areas” a private operator takes over, lower is the ADC it has to pay. This is in line with the outcome that the Government-owned company now has that much of a smaller area to run.

Alternatively, and perhaps a more efficient scheme, would be for the autonomous Government enterprise to maintain each unviable area by forming a Joint Venture with a private operator. This would ensure that the Government enterprise yields results up to the mark, and prevents trade unionism from strangling performance. There would be a bidding process right at the start for selecting the private operator for forming a Joint Venture. As mentioned above, the Government must give up its equity after some prescribed maximum time frame to the private operator. In this Joint Venture, the role of the Government is to make sure that connectivity standards are met by the private operator, and that of the latter is to see that performance does not go down the drain.

The idea is for this Government enterprise to undertake upgrading rural telecom infrastructure in the maximum number of areas in the minimum amount of time, and then exit. At the end of that maximum time, regardless of whether or not the Government-owned company met its own targets, the ADC will be eliminated, and the whole country will be free for private players to start operations in. This will ensure performance.

This is a model that could be applied to every single sector – telecom, public transport – including railways, petroleum, electricity, and the like – where the Government is in the business ostensibly so that the poor can also avail those services. Summing up, any model for Government-private partnership in any sector has to be centered around the following tenets:

  1. The Government has to get out of “the business of running businesses”.
  2. The Government must chalk out clear policies for the sector under consideration and appoint a fully autonomous body to oversee the enforcement of these policies.
  3. Private enterprises are free to offer services in whatever areas they deem fit to.
  4. Private enterprises must fund, for a fixed amount of time, a Government programme to develop infrastructure in presently unviable areas.

At the end of the day, no model is going to be fool-proof, and cannot guarantee complete success. However, the one outlined in this article (or any one that incorporates the tenets above) has the best chance of eliminating distortionary pricing, ensuring a level playing field, and bridging the gap between urban and rural India. And that, surely, is what even the State wants.

** The Hindu Business Line, September 3, 2005.
*** Businessworld India, December 15, 2003.

Thoughts on wireless Internet access in India

February 6, 2005

The Indian Express has two articles today showcasing the state of wireless Internet access in India. They could not be more contrasting.

One article talks about the burgeoning use of WiFi by the upper middle class, especially in Mumbai and Bangalore. It’s exciting to know that entire residential complexes (Hiranandani – the real big builders – are profiled in here) are being provided with ready-to-use wireless Internet access. Further, the major vendors of traditional Internet access are WiFi-aware, and WiFi-ready too. For instance, Hathway and AirTel are already offering WiFi installation services. And at least in the major cities – Mumbai, Bangalore, Delhi and Pune, malls, cafes and bookstores seem to have wireless access enabled. Judging by the prices they’ve quoted in the article, WiFi internet access doesn’t seem outrageously expensive too.

There are only two bottlenecks to wireless internet access exploding in India.

One is the abysmally low percentage of computer owners who possess a laptop. According to the article, one percent of computer owners in India own a notebook computer. I think that’s about to change, though. Notebooks are now available in India around the Rs. 50000 mark, which is quite affordable for most. (I must also add that you can now purchase an amazingly powerful Pentium 4-based desktop PC for around Rs. 18000!) When I was researching notebooks for my eventual purchase of this ThinkPad, I found that most mid-range laptops, by local vendors like Zenith, and others like Acer, were available for as little as Rs. 40000. Certain Compaq models were next in line, costing about 50000-65000 (although HP has really high-end models too). So for an average middle-class family, buying a notebook computer should seem a natural choice.

By throwing open the doors for widespread rollouts of outdoor WiFi networks, Notebook computer penetration and wireless access ubiquity could piggyback on each other in a psoitive feedback loop once we’ve reached a critical mass of consumers.

The other problem is far more serious. That problem is the Indian Government. This other article talks about the ridiculous restrictions that have been imposed on outdoor use of WiFi. Here’s a quote:

Before buying equipment, he says he waits for an ‘in-principle’ clearance from the Wireless Planning and Coordination Wing of the Department of Telecommunications and a visit by the local police.

“Next, the standing advisory committee on radio frequency allocation must agree,” he says. “Sometimes they meet once in two months and your application doesn’t come up that day. We provide 30-35 application copies for all members.”

Then he waits for an operating licence.

It’s clear that the laws governing wireless Internet access are out-of-date. Here’s the current procedure for a licence for outdoor wireless access:

  1. “In-principle” approval from the Wireless Planning and Coordination Wing of the DoT,
  2. Visit by the local police,
  3. Approval from Standing Advisory Committee on radio frequency allocation,
  4. Issue of operating licence.

It takes up to a year for this kind of licence to be issued. For most businesses, that is simply too long. If the Government decides, once and for all, a spectrum for general outdoor wireless access, all of these steps could be eliminated. The Government must realise that the above method is simply not scalable. That is, once the number of applications increase, no committee is going to be able to scrutinise applications the way it is possible now. The police will be unable to keep visiting every locality. Actually, why the police need to be involved even today is uncertain.

Licences for wireless access smack of the licence-quota raj that was the hallmark of Indian business till the New Industrial Policy was announced in 1991. All the Government ought to be doing is defining “policy”, not “mechanism”. Incidentally, this a design principle for a lot of successful systems software, and it applies to this situation too. What it means is that the Government should simply formulate a set of guidelines that any WiFi internet access provider must adhere to. This should include the spectrum he/she must use, among others. Once this framework is in place, though, how the provider implements the wireless rollout, the tariffs for Internet access, et al is none of the Government’s business.

In a mature environment for wireless internet access, anyone with the equipment should be able to start a wireless network. It would simply be like a number of private intranets. To utilise the bandwidth offered by a wireless network, you’d need to log into the network. There are simple methods to implement this. Today, wireless network points trap http requests from hosts. If it is an unregistered host, a login/registration page is sent to it (the host), which will show up in a browser window. Very simple and elegant.

There are so many companies, notably Reliance and Hughes Telecom, which have begun digging up roads in most cities to lay their fibre-optic cables. While no one can fault these companies for the end aim – to provide cheap broadband Internet access, crisscrossing cities with subterrannean wires is madness. Especially when WiFi is an infinitely cleaner way to achieve the same thing. I wrote to the editor of the Indian Express two years ago, when Reliance first started this ambitious project. The letter was never published. It’s amusing to see this article so long after my letter, making the same points I had!

Finally, WiFi (or its long distance variant WiMax) is the best way to address the problem of rural connectivity. We have too many Government committees exploring how to “bring the information revolution to the underpriviledges masses” – in simple terms, ubiquitious rural internet access. But they’re all thinking in terms of expensive wired links. The only reason for doing that is to ensure that BSNL gets to play a major role here. BSNL has the largest wired network by far, in this country (much of it due to monopolistic restrictions, but that is not the point here). So it makes sense for the Government to make use of the infrastructure already available. Whether broadband over copper is actually feasible will be judged by how BSNL’s broadband initiative in the metros fares. But is it a good long-term strategy? A few villages in India still have only one phone connection for the entire village, and most have only a handful. How ubiquitious can Internet access get, with this kind of rural penetration?

Now I’m not aware of the kind of bandwidth that broadband over copper can provide. I’m thinking of a dual copper/WiFi infrastructure. We could have multiple broadband connections over existing copper infrastructure leading up to a village. These could then serve as starting points for a lot of WiFi connections. We could have a wireless hub/switch connected to the machine where these copper wires end, and enable multiple wireless connections from there. The “last mile” of the telephone network would become the “last but one” mile, with the actual end point as the wireless link. This is the kind of public/private collaboration the present Congress/Left Government would drool at.

The technology exists right now. So do the ideas, as do business models. The consumer’s been ready a while too. The only person that’s asleep is at the wheel – the Babu at the Department of Telecommunications. We need to sound a wake-up call before the WiFi moment passes us by.

The SQL Slammer Scourge

January 20, 2003

The SQL Slammer Worm hit businesses the world over in Jan 2003, causing, yet again, losses of millions of dollars. But this worm was just one member of a steady stream of viruses and worms to attack the Internet in the last few years. This brings us to the larger issue of the sheer vulnerability of current software, particularly Microsoft’s, to such attacks.

In this letter to the Editor of the Indian Express I highlight this problem and debate how much Microsoft is to blame for this threat.

Your editorial “When the worm turns”, on 28th January, brings to the fore a worrying aspect of computing that seems to have escaped the notice of most of us.

The SQL slammer worm, as you mentioned, exploited a gaping vulnerability in the Microsoft database program SQL Server, used by many many businesses around the world on their mission-critical systems. You have described the enormous damage it caused around the world, particularly in “wired” countries.

Now, at the end of your editorial you have written “Now, if only Microsoft would hurry up and plug that glitch in their software.” This is a crucial point to make and is the crux of my whole argument – should the world rely on just one company to keep its systems, worth billions and billions of dollars, up and running? By this I do not mean that Microsoft is to blame per se. It is perfectly right on its part to provide software that businesses need. The fatal mistake that businesses are making is, going in for software that is closed source, like all of Microsoft’s.

Closed source software is the kind that does not allow the end user to see or modify the source code (the files written in various programming languages that make up the end product). This means that Microsoft and Microsoft only can control SQL Server (and indeed all its other products). Any vulnerability in the program can be fixed through patches issued only by Microsoft, which may choose when to do so and whom to distribute these patches to. True, major Anti-Virus companies have issued fixes to guard against this worm and others, but as usual this is a reaction, these are steps taken only after the damage has occurred. The question we need to ask ourselves is “Why should there have been such a vulnerability in the first place?” This vulnerability remained because no one was able to review SQL Server’s source code in order to notice the problem. If the source code had been released to the public along with the product (or even as the product was being developed), many of the known and (God forbid!) as yet unknown bugs might have been revealed long back.

There is a software movement called Open Source Software. This movement believes in free access to the source code of products, and cooperative development of software. Through this model of software development, excellent products have emerged, such as the Linux operating system, and the Apache Web Server (a web server is the software that drives a website). Open Source Software gives complete control of the software to the end user, with the freedom to use, modify and redistribute the software as he wishes (with some restrictions to maintain the free nature of software). This software is the kind that we ought to be using, simply because of the control that it grants us. No longer does an organisation have to depend upon a handful of companies to maintain its software, and thus, its data.

This last point is important. Most commercial software, especially database products, store the customers’ data in their own format, which is readable only by that company’s software. The details of the format will never be made public. This renders the company totally dependent on the software vendor to guarantee access to the data. As we all know, data is the cornerstone of any organisation. Should businesses, whether small or collosal, leave control of their data in the hands of a single company?

As far as the question of security goes, Open Source Software undergoes intense testing and review by developers around the world. Most bugs in the software are reported immediately, and fixes are released in a matter of hours, rather than days or weeks, as with commercial software.

Will the businesses which lost money due to this worm , or the countless users who were unable to access their email, hold Microsoft accountable for this devastation that this worm has caused? No. It will be the insurance companies who will have to dole out any compensation. Time and money lost is lost forever. And after all this, we have no guarantee (and it is unreasonable to expect one) from Microsoft that its products will no longer contain such vulnerabilities.

Microsoft will continue making buggy products, which will continue to be a bigger menace to the world business, as computing technology makes further inroads into our lives. It is up to us to decide whether to risk our fortunes upon a single, unamenable entity, or upon Open Source products, which are freely modifiable, and thus more secure.