Differential Pricing

diff

The CRTC recently concluded its differential pricing (or net neutrality) hearing. If you weren’t glued to CPAC earlier this month, you can check out the transcripts while we wait on the Commission’s decision. Like any regulatory issue before the CRTC, this one has a long history. The hearings included several mentions Canadian Gamers Organization’s complaint against throttling of certain types of traffic associated with gaming, and the related ITMP regime that developed out of complaints that certain peer-to-peer traffic (like BitTorrent) was being throttled. These cases clearly implicated net neutrality, because they involved ISPs treating some kinds of traffic differently than others, making certain applications perform worse. The CRTC took a dim view of this sort of discrimination unless ISPs could justify its necessity. For example, blocking ‘malicious’ traffic (like DDoS) is acceptable under the ITMP regime because the reasons are deemed valid, but torrenting shouldn’t be blocked just because it is sometimes used to infringe copyright. In an alternate world of net neutrality absolutism we might have ended up with a regulatory regime under which all traffic is protected, and ISPs are legally prohibited from mitigating the sorts of DDoS attacks that have been knocking many services offline in recent years. However, most net neutrality advocates would not support such an extreme interpretation. Under the existing regulatory regime, Canadian ISPs can intervene when they can justify the need, but are not generally allowed to give some kinds of traffic preferential treatment over others.

Most recently, the question has been whether the practice of pricing certain types of traffic differently than others amounts to a similar kind of discrimination. For this, we owe a debt to Ben Klass, whose 2013 complaint (while he was an MA student in Manitoba) got the ball rolling. Klass is one of a small (but growing) number of individuals who have participated in a regulatory process that was really designed to serve the institutions that are being regulated (the ISPs). His work is a great example of how a regulatory system that depends on parties coming forward with complaints fails, when the stakeholders (ISPs) who are meant to come forward don’t want to complain, even though the issue is of public policy importance. Differential pricing is clearly an important public policy debate to have, and the CRTC has recognized as much with the recent hearings.

While an ISP may treat traffic related to Netflix, YouTube, and CraveTV the same way, if two of these services count against a subscriber’s data cap while one does not, then that is a form of differential pricing (known as zero-rating). I may be able to watch Netflix or YouTube without buffering, but if an ISP makes Netflix zero-rated, I will end up paying more at the end of the month if I watch YouTube and exceed my cap. In this example, distinctions are being made about traffic passing through these networks, and they will presumably affect the behaviour of subscribers. The ethical dimensions of these discriminations become clear in situations where ISPs start favoring services in which they have an interest, or when money starts changing hands between companies so that ISPs treat certain applications more favorably than others. Instead of blocking content, an ISP might simply make it unaffordable, with roughly the same effect.

Many ISPs (large and small) have argued that these policies are not nefarious attempts to control subscriber behaviour, but are all about offering choice to consumers, and differentiating themselves from their competition. Some have continued to claim these discriminations are about managing network congestion (much like the old rationale for throttling BitTorrent), but this argument took a beating at the CRTC hearings and isn’t likely to be very convincing. There are good business reasons why you might want to offer customers different options, including unlimited use of a particular app. However, if an ISP is concerned about the amount bandwidth people are using, zero-rating certain services and imposing caps on the rest seems like a silly way to address the problem.

The CRTC’s forthcoming decision has to grapple with some tough questions, and some easy ones. Vertically-integrated companies using internet pricing to discriminate against competing services has analogies with common carriage in the railway/telegraph era, and feels like the sort of unjust discrimination the CRTC is meant to prevent. But if we are going to accept the existence of data caps (and not everyone agrees we should) then should it be a matter of principle to subject all traffic to the cap? If we can discriminate against malware, maybe we can discriminate in favor of security updates, by zero-rating them, or zero-rate access to essential government services. Without data caps, these become non-issues, but a world without caps would have its own issues (which wireless and satellite providers are well aware of).

It’s times like these I don’t envy the regulator’s job.

Finally, we should remember that differential pricing, just like interventions against malicious traffic, presumes monitoring to accurately distinguish different applications and data usage. The ISP does not need to know exactly what subscribers are doing online, but it needs to be able to tell when subscribers are using a zero-rated service. Unless the ISP is somehow relying on the app to provide this information, this means using DPI technology to inspect and categorize traffic. For better or worse, differential pricing is part of the process of intermediation, in which ISPs play a growing and more refined role in governing our digital flows.