The Facts and Fiction of Broadband Caps and Congestion
The argument being put forward in favour of metering our internet use, as ISPs frame it, is that some users hog bandwidth and cause a slow down and congestion on the network for everyone else. However, this is not necessarily the case.
The US model of fixed-cost broadband contains little connection between the costs of providing a service to the user and their levels of consumption. In fact, particularly heavy users could well be the biggest revenue earners for providers. This is because the income of providers depends on when people use the service rather than how much they use it.
The issue could be compared to the use of motorways. Road networks have to be designed for rush-hour levels of traffic. This peak usage sets the costs for construction of the road. Thus, when rush hour expands greatly, new lanes must be added to cope with the extra demand. The cars that cruise the empty motorways in the early morning do not create the extra demand. It is caused by the number of cars using them in the morning and evening rush hours.
This is exactly the same for ISPs dealing with different levels of usage. Peak time for internet use comes between 7.00 and 10.00 in the evening. ISPs use this high level to set their costs and prices. This means that limitation of use on networks does not address the problem. It is not over-use by certain individuals that causes the problem, but rather moderate use from large numbers of people.
Cable internet providers will not relieve the problem by penalising heavy users by applying caps on their bandwidth.
The over-arching problem, therefore, remains. No one wants a slow, clogged internet, especially those users streaming video, playing games, or dependent on cloud resources for their business. In essence, these high-demand services have been a shot in the arm for ISPs. They drive user demand and helped the shift from dial-up to broadband.
So, what is the answer if caps will not do the job? BitTorrent, a company with more than 100 million users a month, has to get to grips with these challenges. Because of this, they understand and have in-depth knowledge of network speeds and those horrible times when the whole thing grinds to a near halt.
The CEO of BitTorrent, Eric Klinker, has suggested that high-bandwidth services should be the best they can, while admitting that something needs to be done to limit the congestion during the peak-usage periods.
Internet traffic at non-peak times, using a system designed to cope with heavy traffic, essentially costs nothing. A single car on the M25 incurs very little cost. Web applications can play their part in prioritising traffic.
The companies such as BitTorrent that generate high traffic must, therefore, help to find a solution. The company, more than twelve months ago, established a new protocol, µTP, which gauges the congestion and limits itself accordingly.
The company claim µTP never contributes to network congestion and so adds no extra elements to ISP costs. Cable internet providers will still have maintenance costs for the upkeep of the system, but congestion would no longer be a problem.
Such developments, Mr Klinker claims, are inherently friendly to the ISPs’ business model. He also invited other application and cloud services to help develop and build on the open-source technology.
This would go some of the way to solving the problem without penalising the user and applying limits. His second idea, which may be less appealing to some, is based around the London congestion charge and peak pricing. Bandwidth-hungry services could be set to download during off-peak times and if that is not possible then the user should pay a premium.
This may be a difficult system to actually put into practice and would rely on users reporting when they encountered congestion.
The congestion problem is not just the preserve of the ISPs, but also web developers. Forcing users to pay for it could well be an uneven and draconian approach.



Recent Feedback