But a cable does not, by itself, do the job. There are a whole series of nationally controllable conditions that have to be met for the promise of the cable to be realised. To understand these, it is useful to look at Nigeria, a country that got connected to a submarine cable for the first time, like Bangladesh, just three years ago.

Why Nigeria? Why not Sri Lanka? Why not Pakistan? Both Pakistan and Sri Lanka were connected to submarine cables many years ago: SEA-ME-WE2 as well as SEA-ME-WE3. The only non-landlocked South Asian countries that were not connected to submarine cables until now are Bangladesh and the Maldives.

Even, Mynamar, which now occupies the absolute bottom in the world's telecoms rankings, was connected to submarine cable (SEA-ME-WE3) before Bangladesh. The Maldives, which has the highest telecoms rankings in South Asia, is a micro-state with very high telecom rates, and is dependent on satellites for international and intra-national connectivity.

Nigeria, a member of the Commonwealth and rich in natural resources like Bangladesh, is Africa's most populous country. It compares well with South Asia in terms of telecoms indicators, with fixed telephones per 100 population slightly higher than Bangladesh and mobile telephones per 100 population higher than Bangladesh, India, and Pakistan.

Nigeria was dependent on expensive satellite connectivity until the 28,800 km long SAT-3/WASC/SAFE cable landed on its shores in 2002. Like today's Bangladesh, hopes were high of what the submarine cable would deliver. Regrettably the cable has not delivered.

This is because neither the Nigerian government nor its industry and civil society made the effort to establish the right national conditions for the cable. As a result, it did not help the ITES sector; it did not improve Internet connectivity; it did not contribute to fulfilling the needs of the people of Nigeria for reliable, cost-effective connectivity.

The SAT-3 cable did not increase Internet traffic from Africa, including Nigeria. Indeed, the year-on-year growth slowed in the year after the cable (71 percent in 2002 and 53 percent in 2003). In the case of Nigeria, one reason could have been delays in completing the national infrastructure necessary for full use of the cable.

While the landing station was completed in December of 2001 and the cable was inaugurated in May of 2002, the traffic started flowing from Nigeria only in April 2003. Those familiar with the repeated delays in contracting the "dry" segment of the cable from Cox's Bazaar to Chittagong are likely to see the similarities. Harry Goldstein quotes the SAT-3 administrator as saying that "in many of these countries . . . backhaul network is quite not accessible or may not be fully in place or may not have the capacity to support international access" (Surf Africa, IEEE Spectrum, February 2004).

And the prices were not conducive to use by information technology enabled services (ITES) entrepreneurs. More than a year after the landing of SAT-3, a US-returned Nigerian ITES entrepreneur, Dr. Aloy Chife, CEO of Socketworks, told Goldstein of his frustrations. He had spent $13,000 in capital costs and was spending $1,000 a month in operations costs for a VSAT connection to run his business, while cable capacity lay unused: "The worst thing that happened to SAT-3/WASC was that the Nigerian people were represented by the most incompetent and most dysfunctional company in the world, NITEL."

It was only after the competitor to NITEL, Globacom, got frustrated by the obstructionism of the incumbent and decided to build its own $150 million cable from Lagos to London that NITEL saw sense and launched a wholesale Internet service in April 2005. But that was too little, too late. Faced with the fiasco of an underutilised cable, the Nigerian government is trying to repair the mess; and finding it difficult because vested interests have got locked in.

The Nigerian government is now considering spinning off the cable operations as a company separate from the about-to-be privatised NITEL and regulating it as an essential facility. But now, the unions are protesting the hiving off of the potentially lucrative cable segment that can be used as a weapon against competitors. Whether the government will succeed in doing the right thing, even three years later, depends on its ability to make the unions see sense.

A cable is an opportunity
Sometimes, there are advantages to starting late. Bangladesh is possibly the largest country lacking submarine connectivity as late as 2005. But, it has the advantage of learning from the mistakes of others. The model of one operator being fully responsible for international connectivity is obsolete. In this obsolete model, the "owner/operator" of the cable station builds reliable connectivity from the cable station to customers' premises (or nearby points) and determines the prices to charge from the customers.

In an environment where domestic competition exists, the customers include not only end-users but other operators. Therefore the regulatory agency has to step in, to ensure that the capacity is made available at cost-effective and non-discriminatory terms and to ensure high reliability. This requires a strong and competent regulatory agency.

A model more appropriate for a quasi-competitive environment like Bangladesh, is to treat the submarine cable station as a common facility where different operators can co-locate their equipment at non-discriminatory and cost-oriented rates. This is what is theoretically optimal for a competitive environment. This too requires a strong regulator to ensure that co-location is properly provided.

Because not all operators may wish to incur the high capital costs of directly interfacing with the submarine cable, especially at the outset before demand has built up, a pragmatic solution would be a synthesis of the solutions described above. That is, the incumbent operator Bangladesh Telegraph and Telephone Board (BTTB) builds capacity to major population/commercial centers and offers connectivity at non-discriminatory and cost-oriented terms to other operators (the "buy" option).

But all operators are also given the "build" option in that they can, if they so wish, build the link to the cable station, co-locate equipment, etc. The offering of both build and buy options reduces the burden placed on regulation. The fact that the incumbent's customers (the other operators) are free to build their own capacity and enter into direct competition at the facility-level tends to moderate the incumbent's anti-competitive urges.

The above arrangement can ensure better utilisation of Bangladesh's fibre lifeline to the service economy; but there remain the problems of implementation. As long as the operator of the cable segment is also active in the downstream service sectors, it has incentives to use its control of the bottleneck to harm its competitors.

It will not be driven by the desire to sell as much capacity as possible and optimise revenues from the cable business. These incentives will be overridden by the incentives to gain advantages over competitors in downstream markets. Opacity of BTTB's accounts will make it difficult for the regulator to find out whether it is discriminating against its downstream competitors.

Implementation also requires a competent and committed regulator and a modern, competition-centred regulatory regime written into the governing legislation. Without lessening the urgency of reforming Bangladesh's regulatory framework, the immediate problem can be addressed by structurally separating the cable segment (the share of the SEA-ME-WE 4 cable, the cable station, the fibre connecting the landing station to major population centers, the redundancy channels and related facilities) from BTTB, vesting its ownership in a fully government owned company.

To ensure that the new company is truly separate from BTTB and that it is efficiently managed, it is necessary to concession out its management to a competent international operator through a transparent bidding process. The management contract must include provisions for avoiding conflict of interest (operator cannot be involved in running downstream businesses within Bangladesh), incentivising greatest use of capacity, and ensuring non-discriminatory pricing to operators (e.g. mandating web-publication of capacity sale terms and prices as well as quality of service indicators).

The implementation of this institutional arrangement will, no doubt, take some time. It should have been done as soon as Bangladesh decided to join the SEA-ME-WE 4 consortium. But that is water under the bridge. Given the current delays in completing the Chittagong-Cox's Bazar dry segment, the government can expeditiously implement this solution without wasting more time.

The choice before the government today is no different from that facing a home builder who discovers that the foundation of his house is fundamentally flawed. Fixing the foundation will take time and money, but those costs less than those of perpetually repairing the house with the flawed foundation or of being unable to make full use of the house.

If the opportunity is seized
The significance of connecting Bangladesh to the world through an optical fibre cable must be emphasised. A densely populated country like Bangladesh has little alternative but to develop its services sector in the globalised economy of today. It is unlikely that its agriculture and industry sectors can adequately employ its youth and pay them well.

Already, Bangladesh is exporting services on a massive scale, in the form of low-paid workers to the Middle East and elsewhere. This is disruptive to families, dangerous and demeaning to the workers, and non-optimal for the economy, in that the wages that are earned do not nourish other Bangladeshi businesses. The challenge is to shift from exporting workers to exporting work. The key to this is telecommunications.

Today, Sri Lankan accountants are selling their services over the SEA-ME-WE 3 cable; Indian doctors are reading the x-ray images of patients in the US. They are together with their families while making a good living. The income they earn benefits the local economy, because they buy their groceries from the shop down the street and not in some foreign land.

If the government of Bangladesh and the relevant stakeholders learn from the mistakes of others and create a policy and regulatory environment that will allow the greatest participation of multiple players in the large task of cost-effectively connecting the people and companies of Bangladesh to the burgeoning service markets of the world, and also help the country's agriculture and industry sectors to be more competitive, the promise of the service economy will no longer be a distant dream.

Rohan Samarajiva is the Executive Director of LIRNEasia, Chairman of the Lanka Software Foundation, and Member of the ICT Subcommittee of the Ceylon Chamber of Commerce.