2003 July 1 – Yochai Benkler

(entry last updated: 2003-07-01 13:41:22)

(Hubris, I know, but I’m going to try it again today….) (Donna’s links)

(Yochai’s now at Yale, according to Larry’s intro)

Title: The Technical is Political: Access to an open information environment. An extension of the architecture of communication and its implications for the political economy of communication.

  1. Models of Communication – Who gets to say what to whom and who gets to control that exchange of information. At the extremes we have the Broadcast model (special presenters speaking from the core to a receptive core) and the internet model (multiple presenters, throughout the system, distributed capital) – with the telephone in between, where content is at the adges, but the distribution follolws specific sorts of rules mandated by the capital owners.

    Yochai then puts up a painful graphic, mostly to depict complexity in the modalities of communication. The chain is (a) noise/signal conversion; (b) intelligence produiction, (c) message production, (c) transmission and (e) reception. Yochain shows that the number of these elements that “belong” to carriers vs users varies widely along this spectrum of communication models.

  2. The stakes of architecture – at stake is democracy. In particular political democracy. Jonas of IDT is quoted as pointing out that he wants to control content by controlling the pipe. (Note that not everyone agrees that this happens). Versus the internet as a domain where everyone can be a pamphleteer – access makes it possible to change/form opinion from the bottom up.

    Another stake: personal automony. Because you can work the arhitecture, you can manipulate quality of service to move people from one source of information to another – essentially offer preferential capability to make one source more attractive than another. (This page takes forever to load – let’s try another site that is faster)

    Another stake: innovation. I don’t need permission to innovate in an open network, but I do need to in the controlled networks

    Another stake: efficiency – in the economic sense that monopoly is less efficient in the marketplace than competition. Also, the efficiency of flexibility – by optimizing a network to achieve one task, it’s now suboptimal for another task. So, deadweight losses as we wait for the network to change.

  3. The State of play at the physical layer – one way to think about the policy problems is tothink of the layered model of communication. Yochai chooses three layers to construct his policy discussion: the content layer, the logical layer, the physical layer. (Shades of the class we’re putting together for CMI! This is the working model employed in many communications policy research activities at MIT) (Yochai points out that the logical layer is also the standards layer – the other two layers should be self-evident; the wires at one and the content at the other)

    At the physical layer in the transition to broadband, we have a little wireless, a little more satellite, a little more DSL and a lot of cable – on Yochai’s plot, it looks like about 70% of the bar – not to scale according to Yochai. An FCC Third 706 Report, Feb 2002 statistics are cited. Looks like growth, especially in cable and DSL. But maybe not.

    Lets look at homes – now it’s entirely cable and DSL – all the other wired methods (T1, etc) are used by bigger institutions. If you go to SOHO – Cable is dominant, followed by the Bells’ DSL, then a sliver of other sources of broadband. Is this competition? Not necessarily, according to Yochai. Just the phone companies and the cable companies run the physical layer that runs into the home and small office.

    Historically, of course, communications has been considered to be a natural monopoly – in particular, multiple providers would be more costly than a single, because of the capital costs of constructing parallel networks. Thus, we had monopolies, but we need regulators to make sure that the monopolist doesn’t exert its power to extract monopoly rents.

    In the 1990s, a move takes place to change telecommunications law to accommodate the new technologies of wired networks – cable TV and telephones. We need to get them to upgrade their networks to get them to provide broadband. How to do this – with competition or monopoly? Given that the regulators seem to have failed to save us from the inefficiencies of a regulated monopoly, the experiment is to try a little competition instead of an imperfectly regulated monopoly. (glossing of much of network economics here)

    The 1996 Telecommunications Act. Aggressive regulation to require sharing of bottlenecks in the network – to construct intra-modal competition in telephone companies. However, we won’t require cable to share the same way that we require telcos to share.

    The phone companies have fought this in the courts, and have managed to slow the degree to which the Bells have relinquished market to other participants. But the cable companies get a free ride, even though they have to deal with local requirements as cable is deployed. Local jurisdictions tried to set rules in these agreements that match the requirements that the telcos face. This too was fought in court, and the rulings have tended to distinguish cable from communication – thus constructing a loophole that keeps the cables out of this sharing regime.

    In the past year and a half, the requirements to share are bing increasingly gutted by the FCC and the DC Circuit Court of Appeals. While there was a vision that there will be competitors on each wire, there are going to be two parallel networks – the telephones and the cables. The shift is to say that competition between the two modes (cable and telcos) is all that’s needed – we don’t need competition within the distribution mode – duopoly.

    So – cable will be controlled by one company in each location; DSL will be controlled by one company in each location; competition will be between the two modes (assuming you have two to choose). Is duopoly enough?


    1. A refinement of Yochai’s claim of efficiency in monopolies – it’s more efficient in that the cost to each user is as low as it can be, assuming the regulators do their jobs to keep the monopolist from extracting excess rent. Eventually, this becomes an explanation of natural monopoly – a consequence of the nature of the technologies available to supply the good.

    2. You’re talking about the cable and the telcos, what happened to the internet? Can’t that be regulated too?

      Yochai points out that the internet is not the physical layer – it’s the logical layer. So there’s something else we have to worry about.

  4. Issues

    Let’s talk about another possible physical layer – the open wireless network. Wi-Fi plus something. The plus is that the end user devices have to route, as well as receive and transmit. An ad hoc infrastructure that can scale without necessarily relying upon substantial infrastructure – the end-user needs drive the development of the last mile (the OWL network)

    As you look at the topology, a key point is that there have to be at least some members on the OWL who are also connected to the hardwire internet network.

    How to consider whether this is a good thing to do or not. Note first there are no owner – the end users make the network by buying the hardware they need. Moreover, there is no license – just equipment that can be acquired which does not require a license.

    Some consequences: the end user equipment will probably be more expensive than it might be otherwise. Because the equipment does not require the user to subscribe to an ISP, the consumer will be willing to pay more for the equipment than they would for something that connects to leased network lines. Also a new notion of valuation is needed to assess the worth of wireless.

    Which brings us to open spectrum – (I’m only going to summarize, because this is pretty widely available online already). Interference in spectrum is partly an issue of the “dumbness” of the receivers. Licensing is set up to protect stupid radio devices, in this scheme.

    Cheap processing means that we can make smart receivers, so they can think harder about what the signals mean. Also, new theories, especially Shannon’s Information Theories (look up link). So, interference can be tolerated.

    Also, cooperative transmission routing can also improve efficency, by load balancing and by working to minimize transmission inside the range of other transceivers – again minimizing the effects of interference.

    A counter intuitive effect, the addition of users does not reduce the capacity of the network – rather, it increases the capacity – a new repeater in the network and a new cooperator/path opportunity. And there are even more technological strategies out there – multiuser detection, spatial diversity, and sharing information about signal structure to ease filtering.

    Note: Much of this is still at the theoretical/model stage. Lots of things to learn to find out whether this will really work.

    So what does this do to displacement in the network? In elements of this techniology, we reach a point where there is no displacement of the network resource with a particular communication – if there is no displacement/loss of netork function, then the marginal cost of the communication is $0 – and it should be priced accordingly, i.e., free.

    IOW, spectrum can only be valued once you think about the technology and the local deployment of the technology – meaning that markets in spectrum are possibly far too expensive for what they sell and introduce too much inefficiency into the market for communication.

    The transactions costs in this OWL network are stupendous; Yochai argues that this means there shouldn’t be a market. (But there are also arguments that would say that this means that we need a property/market construct to reduce these transaction costs.)

    Yochai argues that there are open and closed pieces of the logical layer – TCP/IP open, MS O/Ses are closed, Linux open. Trusted systems introduce a closed layer; as might CBDTPA. Similarly, there are pieces in the content network that are open or closed. Copyright, KaZaA, censorship, free sharing.

    Yochai argues that we need an OWL or something like that to ensure that there is at least one path, from the physical layer to the content layer, for communication that cannot be blocked though ownership/permissions.


  1. Where does the connection to the global internet fit into this? Doesn’t that still mean that there’s a potential block?

    I don’t know; it might work, but we are currently facing the last mile as the ugliest problem. So far, we don’t see abuses at the backbone, so we may just get lucky.

  2. How might the current providers block this development?

    One thing is the push to make spectrum property; thus prohibit this technology from being legal. So far, the FCC seems to have made some substantial changes in the direction toward open spectrum, rather than away from it. Several windows in spectrum are being opened up, and we will see whether this really works.

    But there will definitely be regulatory tricks. Look at the Verizon activity in NYC, where their pay phones are going to be wireless access. This may lead to some interesting market plays, either giving Verizon ownership of the OWL (thus allowing them to make you one of their subscribers) or getting you used to having this sort of access, so you’ll start paying more for (Verizon) hardware that will give it to you in the form that you like most.