Internet Service Provision – A Business of Averages

The whole basis of the Internet and packet switched networks (as opposed to circuit switched networks) is that a single ‘pipe’ can be shared by many users. If it were necessary for every user of the Internet to have their own exclusive ‘pipe’ to the Internet – then the internet as it exists today would simply not exist.

Contention is good.

The business then of ISPs is to manage this contention and this averaging of usage between users. The lower the average usage of ISPs customers, the higher the contention can be, without affecting those users and therefore the lower the cost of provision and in turn the lower the cost of the product to the customer.

Because Internet users tend to use the service at similar times, and not use it at others, the averaging that matters is peak average usage. Thus an ISP’s job is to provide enough capacity to handle this peak average usage. Their job is to match the contention in the different parts of their network to the ‘optimal’ level for the service they provide.

In any business that is based on such ‘averaging’ there will always be heavier (peak) users and lighter (peak) users. The job of the ISP is to balance these different usages against each other and provide a service that delivers the optimal balance of capacity and cost. The more ‘light’ and ‘average’ users there are the easier it is to support a subset of heavy users.

Of course this means there is room for a range of different Internet service providers that offer different balances. There could be and are services aiming at ‘low usage’ users, which have higher contention ratios and lower cost as well as the opposite.

So we finally come to the problem of NTL’s attempt to limit the usage of their services. NTL do not market their service as a ‘low average’ usage product, or at a price that is compatible with such. Yet they are attempting to reduce the average usage of their users and thus their costs of provision, without offering the benefits of those price reductions to users. Rather than talk about ‘optimal contention’ in an open and honest way, they seek to characterise their heavy users as ‘abusers’ of the service. Rather than perform their role of matching and balancing customer’s usage in optimal ways, they are seeking to increase the profitability of their product, with no cost benefit to the users of the service. They claim the cap will only affect a tiny proportion of their users and yet the smaller this proportion is the less there is a need for such a restriction. They also make a false assumption that a heavy ‘off peak’ user of the service is automatically a heavy user during ‘peak’ usage times. Further they also make the false assumption that by ‘modifying’ a heavy users' usage across peak and off peak times, they will reduce it during peak times. In reality it is much more likely that such users that do modify their behaviour will simply use the service less in the less popular times of usage but keep their usage the same in peak times, thus gaining no benefit for NTL at the same time as restricting the service for such users.

Even ignoring the woeful, and in many respects downright dishonest, way in which NTL has ‘introduced’ these new limits to a previously unlimited service, the above shows that such limits have been badly considered by NTL. It shows that they will be very unlikely to create the benefits claimed for NTL, let alone gain any benefits for users. They show that NTL seeks to turn it’s BB service in to a ‘budget’ service aimed at and suitable for ‘low usage’ users, whilst continuing to market and price it as a ‘premium’ service.