In the wake of the FCC’s net neutrality Order, published on December 23, several of us have focused on the Commission’s confused and contradictory treatment of “paid prioritization.” In the Order, the FCC explicitly permits some forms of paid priority on the Internet but strongly discourages other forms.
From the beginning — that is, since the advent of the net neutrality concept early last decade — I argued that a strict neutrality regime would have outlawed, among other important technologies, CDNs, which prioritized traffic and made (make!) the Web video revolution possible.
So I took particular notice of this new interview (sub. required) with Akamai CEO Paul Sagan in the February 2011 issue of MIT’s Technology Review:
TR: You’re making copies of videos and other Web content and distributing them from strategic points, on the fly.
Paul Sagan: Or routes that are picked on the fly, to route around problematic conditions in real time. You could use Boston [as an analogy]. How do you want to cross the Charles to, say, go to Fenway from Cambridge? There are a lot of bridges you can take. The Internet protocol, though, would probably always tell you to take the Mass. Ave. bridge, or the BU Bridge, which is under construction right now and is the wrong answer. But it would just keep trying. The Internet can’t ever figure that out — it doesn’t. And we do.
There it is. Akamai and other content delivery networks (CDNs), including Google, which has built its own CDN-like network, “route around” “the Internet,” which “can’t ever figure . . . out” the fastest path needed for robust packet delivery. And they do so for a price. In other words: paid priority. Content companies, edge innovators, basement bloggers, and poor non-profits who don’t pay don’t get the advantages of CDN fast lanes.
So important are CDNs in today’s Internet architecture that the FCC felt the need to explicitly exempt them. So much for a “neutral” policy.
In footnote 235 and elsewhere, the FCC seems to think CDNs are all about geographic advantages (to overcome speed-of-light delay) and server-consolidation (yielding infrastructure efficiencies to content and app providers). True enough, but the FCC ignores what Sagan believes is his key service — real-time prioritization.
The net neutrality Order allows CDN prioritization but says paid priority by broadband service providers on the customer link “would raise significant cause for concern,” and “as a general matter, it is unlikely that pay for priority would satisfy the ‘no unreasonable discrimination’ standard.”
Numerous legal infirmities are likely to kill the entire Order when reviewed by the courts. But the substantive technological problems and inconsistencies embodied in the paid priority section need to be examined. Parsing these items should help us all understand how this very complex thing we call the Internet works . . . perhaps chasten the FCC when it attempts to enforce its new rule . . . and hopefully inform a better policy should this Order be vacated.
The FCC fails to explain persuasively why CDN prioritization is much different from last-mile broadband prioritization. Even if a meaningful distinction were demonstrated, we’d still be left with a probable ban on last-mile prioritization, which is a problem because last-mile prioritization will likely prove essential for a variety of real-time communication services.
The FCC’s case really rests on Section II.B, which is all about the supposed motives of broadband service providers. Apparently, the FCC thinks CDNs should be exempt because they have no malign intent whereas broadband service providers are awash in “incentives to limit Internet openness.”
The FCC says broadband providers may have incentives to:
(1) disadvantage some edge providers by blocking or controlling the transmission to end-users;
(2) charge edge providers for access or priority to end-users; and
(3) “degrade or decline to increase the quality of the service they provide to non-prioritized traffic.”
The “no blocking” principle that all parties agreed to way back in 2005 would seem to take care of ominous incentive (1). And a much simpler reliance on existing consumer and/or antitrust law could easily defang ominous incentives (2) and (3).
But these specifics fail to address the larger point: Is it even true that most incentives steer broadband service providers toward a less open Internet? Did the FCC even consider incentives that point in the opposite direction — toward a more open Internet? I can’t find any evidence they did.
The FCC apparently cannot see that the vast bounty of content and apps on the Web created an entirely new market for telcos and cablecos. Namely, broadband Internet access. (Of course, in the typical two-sided coin of positive-sum innovation, it was broadband that created a new market for content and apps.)
As I wrote in The Wall Street Journal way back in 2006:
Blocking and degrading Internet access would quite simply be business suicide for incumbent service providers. Compared to cable’s other content operations like basic and premium TV channels, its broadband cable modem services are more than 50 times as profitable per unit of bandwidth consumed. This means that with just a tiny sliver of the usable bandwidth in its pipes, cable’s Internet services supply about 20% of the revenue and the majority of their net income. Does anyone really think the bandwidth providers are going to kill their golden goose?
Yes, it’s true that popular broadband Internet services also cannibalize some existing products. Long distance voice went away, and Web video is now beginning to compete with cable TV. But the telcos and cablecos know they will never be the key creators of content and apps — not compared to the rest of the world. They get the wonderful Web for free. Compared to hefty sums they must pay for TV content (e.g., ESPN) or movies, broadband access is a simple business. Why decimate the value of one of their three basic products (the others being TV and mobile) by closing off big chunks of the Web to their customers? It would be dumb. It won’t happen. It hasn’t happened.
So the FCC twists itself in pretzels trying to use a high-minded “net neutrality” or “open Internet” policy to do what it really wants — the much more vulgar task of regulating the telcos and cablecos — while exempting (for now) technologies and services that it knows are absolutely essential to a well-functioning Internet but are not at all “neutral” or “open” according to the FCC’s own criteria.
I’m glad CDNs are not covered by the rules, but in its feeble attempt to explain the broadband-CDN distinction, the FCC overlooked the fact that the no-priority rules could limit or block innovation in key real-time communications services.
CDNs are good for speeding static content and even putting some broadcast content (like live sports events) onto fast-lanes to consumers. (Akamai’s record traffic day to that point was its Web broadcast of the Masters golf tournament last spring.) But when it comes to decentralized, real-time, interactive, unpredictable, transactional content — like high-resolution video conferencing or online gaming — CDNs won’t do. We will need some form of in-stream prioritization.
But this is a matter of engineering and economics. There are several ways to achieve quality-of-experience for multimedia applications. We can create virtual (logical) channels using packet priority technologies. We can dedicate frequency (analog)channels using cable-TV like banding or, in the optical realm, wavelength division multiplexing (WDM) within a wire. Or we can deploy more wires (or wireless spectrum). There are technical and financial tradeoffs between using computer power (switching) and communications power (bandwidth) to achieve these ends. They depend on the state of the existing infrastructure, the cost-performance ratios of the technologies and resources (which change over time), and the strategic architecture and business model of the network. In the end we will either prioritize digitally on the last-mile link or prioritize incoming/outgoing traffic onto/from a frequency channel just outside the FCC “no priority” zone. But that will then raise the question of whether providing such channels is itself paid priority. Do we begin to see why these are not questions that can be answered by crude politics?
By exempting CDNs, the FCC acknowledges their herculean task in delivering multimedia to the masses. Akamai cogently described the state of play when unveiling its new HD video service:
Why is it so difficult to deliver an optimal end-user experience? Simply put, it’s challenging to reliably deliver high-throughput data streams across the Web – especially to large audiences.
The fundamental challenges of online video delivery lie within the Internet itself. Given that the Internet is made up of over 13,000 competing networks, it works surprisingly well. But its many bottlenecks and capacity limitations are unpredictable and difficult to manage – lying outside the control of any single entity or group. While these problems affect the delivery of all types of Internet content, they are particularly challenging for video, which requires the transfer of large data volumes at very high rates.
But if delivering static and broadcast content is a challenge, doubly so for the coming wave of real-time interactive cinema-quality video. It’s a challenge CDNs cannot conquer and will require new technologies, architectures, and business models — many of them requiring some form of broadband prioritization — to master.
It would be a shame if the FCC’s new Order interrupted this essentially technological and economic story because of a dubious analysis that narrowly focuses on the supposed political motivations of broadband service providers and completely ignores powerful incentives pushing all parties toward Internet openness.
[Cross-posted at Digital Society]