Paid Priority: Devil or Angel?

iStock_000032459330_LargeNow that the dust has settled around the FCC’s latest round of Internet regulations – or “Internet access regulations” if you prefer – people are struggling to digest the likely consequences of the rules. There are a lot of details in the 400-page order, but the most important is the pre-emptive ban on something called “paid prioritization”. The backlash against this obscure practice – which was proposed by the FCC itself in the May 2014 Notice of Proposed Rulemaking – was instrumental in justifying the FCC’s use of Title II in the final rules.

Paid prioritization is actually a slippery notion. In the popular telling, it means a pay-to-play “fast lane” that allows some content to move through an ISP’s network ahead of other content. That sounds like a bad thing because it would create a two-tier system of content delivery, where the rich can pay for expressways and the poor are relegated to dirt roads. Such a system might very well shield successful companies like Google and Netflix from competition, all other things being equal.

The FCC cited the opinion of Chris Riley, a policy analyst currently working for Mozilla (he formerly worked for Free Press) to the effect that priority is a zero-sum game: “Paid prioritization has a distinct degrading effect on other access service traffic, an effect that creates complex incentives for network operators… Prioritization is inherently a zero-sum practice, and inherently creates fast and slow lanes and prevents a level playing field.”

That seems to make sense: if the post office delivers my mail before it delivers yours because I paid extra, that means your mail is delayed and you suffer. Or does it?

The judgment that such a practice is always bad and discriminatory assumes that every application has the same needs and wants. The postal analogy only makes sense if you and I are both hovering over our mailboxes with bated breath waiting for a life-saving medication, and that’s not the typical case. What if I’m waiting for medication and you’re on vacation?

In this case, we have entirely different needs, obviously. You don’t care when the mail is delivered as long as it’s there when you come back home but I care a lot. This is the situation that arises on the Internet when one person in a home is on a phone call with Vonage and three others are streaming Netflix programs.

Vonage and Netflix both communicate over the Internet in terms of packets, units of information, but they don’t do the same thing with packets when they arrive. Vonage converts packets to sound as soon as it gets them and plays that sound to the ear in a few thousandths of a second. Netflix, on the other hand, saves incoming packets in a buffer and plays them to your screen an average of five seconds later.

In effect, Vonage is hovering over the mailbox and Netflix is on vacation. So Riley’s claim about “distinct degrading effects“ is simply untrue. Because applications are different, moving one packet from Vonage ahead of a hundred Netflix packets does not degrade the Netflix viewing experience AT ALL. Vonage has a slight measurable effect on Netflix because of the volume of data it delivers, but prioritizing that data adds no additional problems for Netflix.

The FCC believed that Riley’s analysis was correct and enacted a pre-emptive ban on prioritization on the basis of that belief. Because this is a clear analytical error, the question of how, when, and why it will be corrected is interesting.

These questions are worthy of analysis.