EU Reconsiders Network Neutrality

One of the problems with  “network neutrality” continues to  be that nobody agrees about what we are talking about. Now there is a potential change in European Union policies regarding “network neutrality” that essentially recognizes that it cannot be defined in a way that all 28 EU members would accept.

A new document issued by the EU president in fact suggests the “removal of the definitions of ‘net neutrality’ and ‘specialized services’” from any future policy documents.

The proposal suggests that, instead of defining network neutrality, policies only point to the objectives of net neutrality. It isn’t clear how a government agency can create or enforce a policy about a practice that is not defined, some might argue.

The inherent difficulty of the subject is further illustrated by the handling of traffic management principles--which always mean the potential for managing traffic flows to protect overall network performance under conditions of congestion. “Clear principles for traffic management in general” are required, the document argues.

That is all well and good, with the obvious caveat that clear policies will be hard, but not impossible, to craft. Especially contentious will be proposed rules about Internet service providers being required to “maintain sufficient network capacity for the Internet access service regardless of other services also delivered over the same access.”

The example given in the document is sufficient bandwidth to handle simultaneous use of email by one user while another watches streaming video.

In principle, that sounds simple enough, but requires assumptions about what apps, with what bandwidth and latency characteristics, are used at any location, by what number of devices and users, at any particular time, in an environment where end user demand changes continually.

The problem is that best effort Internet access is, by definition, conditional, a matter of  statistics, since the service is shared by many users.

And that means Internet service providers essentially have to guess at peak hour demand, in the same way that voice providers had to estimate peak demand. Maintaining sufficient network capacity” therefore requires quantifying estimated demand at peak hour. Sometimes planners are going to guess wrong.

Also, the document allows for transparent, non-discriminatory and proportionate traffic management that is not anti-competitive. Some might argue that is among the fundamental problems with network neutrality: networks have to be managed, which means sometimes it is necessary to groom (“throttle or block”) traffic, or simply let it all slow down.

Of course that should happen in a way that is not anti-competitive. And there are legal remedies for anti-competitive behavior. But it often does not make sense to groom traffic in a “non-discriminatory” way.

Video and voice are highly intolerant of packet delay. When networks get congested, better user experience is obtained if those types of apps--video and voice--are given priority, while other apps more tolerant of delay can be delayed. Some strong forms of network neutrality make that illegal.

There is a reason “network neutrality” is so hard to understand. Defining it is difficult. And defining it is difficult because we are asked to suspend thinking about network congestion issues under conditions of scarcity.

And make no mistake: networks are risky, expensive bets, so  they will remain relatively scarce. No network can be built for infinite demand, thus necessitating hard choices about capacity. Congested networks must be managed. The only issue is how they are managed.