Starving the data hogs

Earlier this year Telstra’s Kate McKenzie urged the telecommunications sector to start a debate about how it should respond to the increasing congestion issues it is experiencing as data traffic continues to increase dramatically. It would appear the Australian Competition and Consumer Commission is keen to join that discussion.

McKenzie’s proffered solution to the congestion-driven network capacity and quality issues created by the video-driven doubling of data traffic each year was to use pricing to spread the demands on networks away from peak periods.

The ACCC chairman, Rod Sims, referred to McKenzie’s call in a speech to a conference today, saying that congestion pricing was a challenge that regulators had to confront.

In theory, he said, the advantages of congestion pricing were clear and even uncontentious. He asked, however, why there was such a gap between theory and practice and said one explanation might be consumer caution.

"The challenge for us is to understand this caution and see what types of safeguards might need to be put in place," he said, noting 'specifically' questions that would be addressed later in the conference about the role of consumers in determining how congestion would be managed as well as the role of the regulator and whether that differed when the network operator was wholesale-only or vertically integrated.

While Sims didn’t say anything conclusive, the obvious concern for regulators within issues of congestion pricing is whether it is designed to benefit consumers and improve their experiences or is used as a cloak for increased prices and profits.

The issue isn’t confined to telecommunications networks. Energy networks and roads are other obvious sectors where networks experience congested peak periods but the challenge is perhaps seen most starkly in wireless networks.

The take-up of smart phones and the sheer rate of growth in data traffic mean that despite the billions of dollars that Telstra, Optus and Vodafone are pouring into increasing the capacity and speeds of their networks they are facing increasing congestion issues.

As with electricity, the issue isn’t one of overall capacity but rather capacity at peak periods and the question is whether it is sensible to keep investing in adding capacity to cope with those peaks in demand or whether demand can be smoothed and shifted and the loads spread over the day.

As Sims noted, that can be done using ‘traffic management practices’ – assigning priorities to particular types of demand. Telstra, for instance, has been trialling 'shaping' of its loads, which means restricting the bandwidths available to particular types of services, mainly peer-to-peer services. As he also noted, 'managing' traffic could also be used tactically to disadvantage third party services like 'over-the-top' voice and messaging services.

The alternative approach, and the one McKenzie put up for debate, is to use pricing.

On the Telstra wireless network the top one per cent of users account for about 25 to 30 per cent of data traffic. The bottom 80 per cent of users account for less than 15 per cent. Price signals could be used to shift the bandwidth-intensive applications away from the peak periods and into those periods when the networks are under-utilised.

McKenzie pondered whether the sector could introduce a pricing regime which enabled the users themselves to decide on the quality and priority of the services they wanted, paying for larger data allowances at peak times, or not.

It is wasteful and inefficient for the networks to keep adding capacity to cope with ever increasing peaks in demand – and then have the networks under-utilised for much of the day. Increasing congestion driven by the behaviours/demands/needs of the few is also unfair to the great majority who don’t place the same demands on the networks.

Price, provided consumers are able to make their own decisions about what they are prepared to pay for, is a better and more transparent mechanism for managing and shaping demand than arbitrary decisions about traffic priorities made by network operators that may not be driven entirely by the interests of the end users.

If high-bandwidth users aren’t prepared to pay the price of peak period usage then the load on the networks would be spread and the operators would need to continually invest on the basis of peak demand. If they are prepared to pay a premium for large data usage at peak periods then investing in that increased capacity would be viable.

An issue which the ACCC will no doubt want to explore is the extent to which competition between the wireless operators and, to a degree, the open-access fixed line network that will (regardless of the election outcome) be constructed, is sufficient to discipline pricing and ensure that it isn’t just a cloak for profiteering.

Continuing increases in congestion, of course, would probably have its own impacts on the shaping of demand, albeit not ones that are necessarily in the interests of either consumers or the networks.

The alternative, which the operators wouldn’t be enthusiastic about, would be increased and intrusive regulation of a sector that has generally been quite lightly regulated.

Sims and McKenzie are both right. There is a much more intense and detailed debate that needs to be had and given the scale of the investments and the nature of the traffic management trials the sector has been making  (and would continue to have to make to try to respond to the congestion challenge), it’s one that should probably be had now.