(I've been meaning to post this for ages . . . apologies if it sounds old, it is.)
Internet regulatory policy that allows for vertically integrated "service" offerings based on QoS or QoS-like tiering is completely wrongheaded (I must sound like a broken record by now).
Some of the reasons are well known. Most obviously, tiering only makes financial sense to a small oligopoly of broadband providers who stand to profit from restraining broadband capacity, and the regulatory body that's laid claim to oversight (the FCC) of broadband has huge problems defining things like 'broadband' and 'competition' in terms that don't flagrantly and unjustly benefit the handful of incumbent providers. Their definition of broadband (200 Kbps) and the way they measure broadband penetration (by ZIP-CODE!?) are both woefully inadequate (at least according to the GAO); and if they actually applied sane measures of competition and concentration (like the HHI) they'd never have allowed the Terminator-esque merger-mania that's lead a return of Ma-Bell. The phenomenon of 'regulatory capture' is nearly synonymous with the FCC these days; despite the hard work of Commissioners Copps and Adelstein. But such arguments are well worn and well-known. They're arguments we're all familiar with by now (or at the very least ought to be) and they're very 'policy-minded'.
Instead of dwelling on these deficiencies of our regulatory system, I want to focus on a larger, more profound reason it's a mistake to allow such "tiered services." The reason I'm thinking of depends on basic understanding of what it is to "be digital" (apologies to Nicholas Negroponte). It is the not-so-simple observation that:
Computation IS Communication
The "services" model - the model we currently suffer - grossly fails to grasp what it is regulating. Here I need to make a point that cuts to the core of a number of issues that are being hotly debated - not least of which is the patentability of software, to which I'll return in another blog. The crucial point is that digital communications - as opposed to analogue communications - is by definition digital computation. I cannot emphasize the profound importance of this fact enough. It is the key to understanding why the BrandX decision was such a travesty and why the history of FCC regulatory intervention has been such a disaster. The insight is neither my opinion nor a veiled form of 'technological determinism', it is rather a FACT about any digital medium; it is a fact about "being digital".
Some legal thinkers and policy wonks have already touched on this aspect of "being digital", though haven't gleaned (or refuse to entertain) the full extent of it's impact. I'm thinking particularly of Jonathan Zittrain's influential article on "The Generative Internet" and Kevin Werbach's article, "The Federal Computer Commission" (North Carolina Law Review 84:1-75). Zittrain, on the one hand, argues that focusing on the end-to-end principle of digital networks is “myopic;” dangerously assuming the continued openness of the equipment attached at its ends. Werbach, on the other hand, argues that FCC regulation has in effect, if not in intent, already regulated the equipment at the ends of the network. At least, it’s done it more than most policy analysts would like to admit. I would contend that the breakdown at the ends of the digital network is actually more profound than either Zittrain or Werbach have entertained; that, in fact, regulation of digital communication IS regulation of computation, period. Put another way, the stuff going on at the "ends" of the network that Werbach and Zittrain are expressing concerns about IS EXACTLY what is going on in the network itself.
The basis for this claim comes straight from some very basic aspects of computation and the idea of 'computability' - the idea at the heart of ALL modern computers. The principle was set out by Turing (a response to Hilbert's Entscheidungsproblem) in his paper, " "On Computable Numbers, with an Application to the Entscheidungsproblem." In the paper, Turing describes what has come to be called a Universal Turing Machine (UTM), an abstraction instantiated in the hardware and software of the modern computer. A UTM has a state that is altered by data/instructions it reads-from and writes-to some kind of storage medium. That is the full extent of a UTM's functionality, and it is this reading-from and writing-to storage that is both communication AND the very activity of computation. When Sun Microsystems CTO, Greg Popadopoulos, says, "The world needs only five computers . . . Google [for example] runs a computer [that] happens to have hundreds of thousands of processors in it, and millions of disk drives, but it's a computer. The important distinction is there is a point of control that determines what software is going to run, and then the systems work collectively to provide some service . . ." he is invoking this idea of computability that undergirds the very idea of the computer. An idea also at the heart of the company's slogan since its inception in the early nineteen-eighties, "The Network is the Computer."
Put simply, computation IS nothing more or less than the storage and retrieval of information (data); i.e. computation IS communication. The boundary drawn between the two is increasingly an artifice of political and legal definition rather than a practical and technologically determined (by analogue technologies) reality. Being digital means a near total breakdown of this boundary.
Artificial boundaries
Popadopoulos' comment demonstrates the arbitrariness of the boundaries regulators impose on the digital world. From a computational perspective, it matters little if the bits travel thousands of miles across national borders (as they might if you're participating in a distributed application like SETI@home) or on the same mother-board (as they might on a dual-processor-PC-under-your-desk@home), yet our regulatory framework is founded on imposing such arbitrary boundaries.
While motherboard-centric communications are free-flowing - with little or no filtering between CPU and storage - communications over broadband networks are (post-BrandX) subject to just about whatever limitations and crippling the telecommunications and cable companies wish to impose (in order to extract additional profits for absolutely no additional investment in infrastructure).
Motherboard manufacturers build products that meet inter-operating standards and compete on reliability and speed performance in a competitive market more-or-less free of regulatory intervention. Whereas, broadband access providers perpetuate last-mile scarcity in order to generate additional profits by crippling and limiting bandwidth to some customers in favor of others in a non-competitive market presided over by an oligopoly of telecommunications and cable companies.
While software programmers invest time and assets in development - relying on the communications between the local CPU and the local hard-drive - they cannot safely rely on the communications between the local CPU and a remote hard-drive.
If we are to reap the benefits of digitalization we must enact laws that reflect digital reality; this means a complete overhaul of communications regulation.
