Stephen Temple

When the broadband Internet first emerged the vast majority of users down-loaded Web pages and the vast majority of content was in the form of Web pages. This generated sporadic data usage that the IP nature of the Internet was very efficient in handling. At this time there was a rough industry rule of thumb to install capacity that was around 50 times less than might theoretically be needed if all customers were to have been active at the same instant in time. Very attractive flat rate monthly fees resulted from applying this rule and everyone was happy.

The only threat to this ideal state of affairs came from a very small percentage of users shipping vast quantities of data per month with peer to peer applications – one of the most popular was exchanging DVD movies.  This was the context in which traffic management was first applied to the broadband Internet along with caps, fair usage conditions, threatening letters etc..not that any of this had much effect on this minority of users.

What has changed is the growth of video streaming coupled with higher delivery speeds which is stimulating demand and the three factors (continuous data stream, higher delivery speeds and larger numbers of users) is leading to an exponential growth in the demand for network capacity. What might have appeared a fair use data cap of say 4 Gbytes per month 10 years ago has gradually become the average data being shipped per month per customer. ISP’s are struggling with the consequence of this trend. Where traffic management has historically been applied to deal with a small minority of “bandwdith hogs”  we now find the traffic management essentially being applied to one group of customer/applications in order to protect the quality of service for another group of customers/applications…within the constraint of the installed capacity. The capacity itself is limited by ISP’s wanting to keep subscriptions low.

As the numbers no longer add up something has to give and the choice boils down to:

A. Net Traffic Neutral Options

1. Let all monthly prices rise in order to install enough additional capacity so as to hold up the quality of service for all users/applications.

2. Keep subscriptions/investment low but let the quality of service crash, which everybody equally shares the misery of irrespective of their usage/application.

B. Net Traffic Discriminatory Options

3. Keep subscriptions/new investment in capacity low  but discriminate technically against the most data demanding users/applications in order to preserve the quality of service for modest data usage users/applications…ie technical discrimination by application

4. Let subscriptions rise for the high data usage users/applications to pay for the corresponding additional capacity but keep subscriptions low for the modest data usage users/applications…ie price discrimination by application

Most people looking at the best solution would start from the premise that ideally consumers should always have a choice.  This makes options 2 and options 3 the least desirable as consumers are being left with no choice. This neatly boils the Net Neutrality debate down to simply who pays for the extra network capacity needed to support the data hungry applications. Is it only those using these applications or are these incremental costs shared by all consumers – even the light broadband users who only surf a few web pages and send some e-mails.

We have many examples of public services within which cross subsidies exist between different classes of users – the urban to rural services cross subsidy is a good example so we should not be opposed in principle to a cross subsidy where it serves a wider public interest.

In the cross subsidy model the subscription for all users rises and usually when prices rise it leads to a correspondingly smaller global population of Internet users. If only those using the data hungry applications, such as video, pay the extra cost – it is a bigger jump in price for those users and the market for these data hungry applications would not grow nearly so fast.  There is also a sub-choice in the latter case as to whether the consumers of heavy data applications pay directly for this additional capacity or whether the applications companies pay and retrieve the money via their business model (subscription, advertising etc).

So far I hope my analysis has been neutral but it is time to come off the fence:

In my view it is in nobody’s interest for the quality of service to simply crash due to an impasse over who pays for additional investment to support heavy data usage applications.

The Net Neutrality camp have a fair argument that it is not right for bandwidth management to be applied by a dominant ISP to heavy data usage applications without that ISP also offering a part of the capacity that is free of bandwidth management constraints ie providing a genuine consumer choice.

This deals with the two wrong answers. Where I am more tentative is the core of the debate – who pays for the extra capacity needed for heavy data usage applications. Option 4 has some merit as surely it is more important to have some low entry price subscriptions to the broadband Internet to extend its reach into all part of society – even if this means that there is a slower growth of heavy usage applications, such as video. On the other hand, there may well be a competitive advantage for a country that creates a better climate of innovation for video applications over the broadband Internet? A difficult choice.

I also see it as finely balanced as to whether consumers of heavy data usage applications pay directly for the necessary extra capacity or whether a more convenient arrangement is for ISP to charge application companies for it, who in turn recover the cost from consumers indirectly. Perhaps charging those content companies with a more compelling interest in holding up delivery speeds creates choice and therefore is worth considering.

This debate will run on…and at the end of the day, if governments put in place pro-investment regulation, that will accelerate the arrival of fibre to the home…then the net neutrality argument becomes a thing of the past by virtue of the sheer surplus of network capacity that fibre generates…now there is a debate worth having!