Congestion becomes a hot topic as Sri Lanka’s telecom sector goes through another growth spurt.

Every time the telecom sector goes through a growth spurt in Sri Lanka, congestion becomes a topic of conversation. People begin to grouse that the companies are cheating the customers by giving them telephones but not the ability to make calls.

Every time the telecom sector goes through a growth spurt in Sri Lanka, congestion becomes a topic of conversation. People begin to grouse that the companies are cheating the customers by giving them telephones but not the ability to make calls.Because of the growth spurt in fixed telephony driven by the issuance of CDMA frequencies to Lanka Bell, SLTL and Suntel in mid 2005, the congestion issue is reappearing on conversations again.

Road congestion

Congestion is not limited to telecom. We live with traffic congestion everyday on the streets of Colombo, Kandy and elsewhere. The causes of congestion in telecom are a little more complex and a little easier to remove, but the fundamental dynamics are the same in telecom and roads.

The road is designed at a certain time, based on certain assumptions about traffic. If it is assumed that there will be a lot more traffic in the future than at present, congestion can be postponed, if not avoided altogether.

However, this will drive up the costs considerably; a six-lane highway costs a whole lot more than a two-lane highway. If costs go up, the charges for using the highway should also go up, theoretically at least. Since no one pays for the use of roads in Sri Lanka, this translates into a greater burden on taxpayer funds. Since we don’t actually build highways using taxpayer funds but rely on external loans and grants, this means that that a greater burden is placed on future generations. So the congestion-avoiding action is not cost-free.

every infrastructure is a compromise between building adequate capacity for future growth and keeping costs at reasonable levels.

The result is that every infrastructure is a compromise between building adequate capacity for future growth and keeping costs at reasonable levels. When private firms are supplying infrastructure services under competitive conditions, cost minimization tends to play a greater role than when government is supplying infrastructure under monopoly conditions.

In other words, there should be less congestion on government-supplied roads than on private-supplied telecom networks. But this does not mean that private firms benefit from congestion. A call that is not completed does not generate revenue. There are strong incentives to generating revenue. What is said above is that because private suppliers are likely to be more conservative in their growth assumptions than government suppliers, there is likely to be less congestion in government-built infrastructure in the short term.

But that is obviously not true in Sri Lanka. The ability of government to properly plan, dimension and build infrastructure assumes an effective government, not a dysfunctional mess of the kind we have. The theoretical propensity of government to supply infrastructure services with lower propensity to congestion is washed out by the multiple institutional factors that afflict our government.

This was why I stated above that the congestion problem is simpler in roads than in telecom, but fixing it is more difficult.

Telecom congestion

Congestion in telecom is, in essence, the inability to connect to the called party. From the perspective of the user, congestion takes several forms, including:

  • Inability to get dialtone (which indicates that the telephone is ready to commence dialing)
  • Hearing various tones and stutters for extended periods before one hears the signal indicating that the called phone is ringing (if one gets that far)
  • Perpetual engaged tones.

It is caused by multiple factors. At the simplest level, it could be because the called party has her phone off-hook because it is being used to answer another call or to make one. It can be because there are more users than there is capacity either at the calling party’s base station/exchange or at the called party’s base station/exchange.

In the case of a caller from one network (e.g., Lanka Bell) connecting to a party in another network (e.g., SLTL), the problem could be in either network (more users than capacity) or in the interconnection links (inadequate dimensioning for the traffic).

People figure out simple ways of dealing with congestion. If they hear the engaged tone immediately, they call again after a few minutes…

People figure out simple ways of dealing with congestion. If they hear the engaged tone immediately, they call again after a few minutes, assuming the called party was on another call. When they notice that calls are difficult to make at particular times, like the transition from a high calling rate to a low calling rate (especially at night) they wait a while and try again.

But they cannot deal with interconnection-caused congestion. This type of congestion can be diagnosed by using different phones to make the same call. Assume connection to difficult to an SLTL number from a Suntel phone. If it is possible to make the same call from an SLTL phone, the problem is an inadequately dimensioned interconnection link between Suntel and SLTL. If even the SLTL phone fails to connect, the problem could be in the SLTL network itself, and so on.

Only the regulator can deal with interconnection related congestion. The Indian regulator is paying serious attention to interconnection caused congestion: http://www.trai.gov.in/pr17jan06.pdf.

The recent complaint to the Telecom Regulatory Commission by Lanka Bell about difficulties in getting adequate interconnection capacity from SLTL (http://www.lankabusinessonline.com/lanka-bell-stuck-in-interconnection/) suggests that the time has come for greater attention to the problem by the Sri Lankan regulator as well.

Anti-competitive congestion

Interconnection-related congestion is only partly technical. It is natural to have to increase capacity rapidly when demand is booming, as at the present time in Sri Lanka. But increasing capacity in telecom networks, especially the point-to-point interconnection links between networks which comes in small, flexible increments, is much easier than increasing capacity on roads. Delays of more than a week or two are likely to be driven by anti-competitive motives and must be carefully scrutinized by the regulator.

A customer cannot easily differentiate between within-network congestion which is entirely the responsibility of his operator and interconnection-related congestion which is outside the control of his operator. He will blame his operator even when a competitor has caused the problem. As word gets around, the operator causing congestion will get the new business, especially if he is the incumbent and has a majority of the telephones in the system.

Congestion is a symptom of a mismatch between supply and demand. Some forms of congestion are normal in rapid-growth telecom scenarios. Depending on how the quickly suppliers react, congestion can be managed. Those that do not may be punished because potential customers go their rivals. What is important is that this should not be allowed to be used as an anti-competitive tool, especially by the incumbent.

The remedy to congestion is to seek to increase capacity, either through the existing operators or by introducing adequate competition.

Fixing congestion in telecom is relatively easy. It simply means that the suppliers increase capacity to match demand. Because uncompleted telephone calls do not yield revenue there is no particular advantage from congestion to suppliers in a competitive market. The remedy to congestion is to seek to increase capacity, either through the existing operators or by introducing adequate competition.

-Rohan Samarajiva: samarajiva@lirne.net