Today’s editorial in the San Jose Mercury News in favor of Internet regulation is a real belly-buster:
Network neutrality isn’t new. Its basic tenets — that all users can access all legal content on the Internet and that all content providers are treated the same on the network — have been in effect since the birth of the Internet through regulations governing the old telephone network. But a series of court decisions and a vote of the Federal Communications Commission last year have voided those rules. And that has opened the door for phone and cable companies, which control Internet access, to change the rules of the game.
Bzzztttt – wrong. Cable Internet access has never been covered by common carrier laws, only dial-up and DSL have. And who provides the better service? Lawmakers freed DSL of its disadvantage against cable a year ago. The Internet itself, the backbone network, has never, ever been covered by common carrier regulations, and it’s this very thing that the search and software monopolies seek to change.
Consider the nascent world of Internet video, which promises to be a free-for-all of ingenuity and creativity. With enough bandwidth, CNN, a public access channel or an amateur video producer could put up content for the entire Internet to enjoy. Scores of innovative start-ups are coming up with business models to exploit that creativity, by organizing the new content, making it searchable and delivering it effectively to millions of users.
This stands reality on its head again. Real-time video isn’t practical over Vint Cerf’s Internet because it was engineered for e-mail and downloads. The Internet needs to grow the ability to deliver multi-media streams with very low delay to support the next generation of applications. Net neutrality criminalizes this practice, and along with it most of the advances that have been made in network engineering since the Internet was turned on in 1983.
Outside the Bay Area, few lawmakers seem to understand that by not enacting network neutrality legislation, they’d be subverting the basic principles that have made the Internet into such a powerful force for economic growth. Perhaps, it’s because they’ve been worn down by armies of lobbyists from the telephone and cable industries.
What Bay Area lawmaker understands how the Internet works, the one who gave us the V-Chip or the ones who line their campaign coffers with contributions from the companies that gave us the Internet Bubble? Our politicians are like politicians everywhere, captives to a group of interests who have positions on issues the lawmakers and their staffs scarcely understand. The legislation that opens up rights of way for fiber to the home is sensible, practical, and good for America as a whole and the high tech community in particular.
Network neutrality would be nothing short of a disaster for all of us who create or use networks.
A point about internet regulation history and some questions:
(1) The internet access market was competitive when internet access was considered a telecommunications service, which constrained the cable companies from offering services that discriminated on the basis of content. Who would buy internet access from the cable companies (when other choices are available) if you couldn’t get the content you wanted?
The problem now is that the choices for internet access are dwindling to one or two sources. The FCC’s DSL Order last August removed the restrictions imposed by the common carrier paradigm that made the internet access market competitive. The order grandfathered in exitsing ISPs for another year, so the demise of the common carrier paradigm hasn’t been immediate. That (and the media attention “net neutrality” is getting) is probably why we haven’t seen content discrimination yet. However, the telecoms have announced their plans for it. I’m not buying your “cable wasn’t regulated” rebuttal. The internet was regulated, and neutrality principles were promoted, albeit indirectly through competition among ISPs. Mercury News appears to have that part of the story right, even if it is oversimplified.
(2) Some net neutrality proponents advocate for a pure version of the concept: no discrimination among bits of data such that routers continue to use a best efforts approach. However, softer versions have been suggested that would allow tiers for bandwidth management but prohibit discrimination within a class of data. For example, emergency services would receive priority over video services, which would receive priority over email. But the telecoms and cable companies would be prohibited from giving priority to their own video services over a competitor’s video services. Video services would have to be given the same priority. This softer version apears to be the goal in Ed Markey’s bill: the Net Neutrality Act of 2006. What do you think of this softer version of net neutrality? How would it impact the ability to improve the network itself?
(3) If the telecoms reserve broadband for their own video services in areas where they lay down new fiber, will the remaining broadband be sufficient to handle IPTV once it’s available?
The fact is that cable company Internet access has always been regulated as an “information service” not as a common carrier, and it has out-performed the more heavily regulated DSL and dial-up. If you claim that it was de-facto common carrier because the options were, you have to explain why there’s no access to independent ISPs over cable but there was over DSL (by law.)
You assume a latent desire on the part to ISPs to engage in content discrimination, but we’ve never seen them do it, even though cable has been perfectly capable of doing it all along. Their performance is so much better than DSL that they could easily offer standard access that’s faster than DSL and then extra-special access to favored sites that’s faster still. They haven’t done that, and the whole net neutrality argument depends on the assumption that they want to.
I believe you misstate Markey’s amendment, which as I read it criminalizes third-party billing for QoS, that Skype would not be able to buy QoS at wholesale and pass the savings along to their customers. This is actually anti-competitive.
Finally, fiber has a lot of potential bandwidth; the issue is what you connect to it.
(1) Cable still had to compete with ISPs offering DSL or dial-up services over the telephone lines, so it was economically constrained to offer a similar service. No one would switch to cable for internet access if cable was going to discriminate based on content, when the ISPs were not (or could not because of the amount of competition). Now that competition is dwindling (ISPs are consolidating or disappearing, especially in light of the DSL Order), and most people expect to have two primary providers of internet access: the local cable company and the local telecom. They will provide access to the internet backbone and largely perform all the ISP services in the future. Independent ISPs will be few. In this environment, a market with only two real players, econmic thoery generally predicts that the firms will behave strategically because of their inherent interdependence. It seems likely in such a scenario, that signaling–like the advance announcement of discriminatory practices–will lead to coordination–here content discrimination. The change in market structure is the key is why content discrimination seems likely.
You correctly recognize that there is no need for regulation if the telecoms and cable companies won’t engage in content discrimination. However, the telecoms have made public annoucements that they intend to, and the cable companies probably would follow that lead given the economic dynamics of the market.
The notion that the telecoms and cable companies will discrimniate against competing content is reinforced by their profit incentive. This profit incentive comes through in their arguments about net neutrality hurting build-out. The rationale for this argument is that net neutrality legislation will prevent content discrimination, which would give the telecoms and cable companies some additional profits, some of which would be re-invested in build-out. No regulation, more build-out. For this argument to have any merit, there must be a profit incentive for the telecoms and cable companies to discriminate in content (and by that I mean against competing content/application providers).
What we have is removed regulatory paradigm, a loss of competition (for the foreseeable future), and in their place a duopoly with a history of anti-competitive tactics. (Indeed, they are now complaining about each other’s anti-competitive tactics.)
(2) Ed Markey introduced a stand-alone net neutrality bill called the Net Neutrality Act of 2006 (HR 5273). Section 4(a) provides: “Each broadband network provider has the duty to . . . (7) if the broadband network provider prioritizes or offers enhanced quality of service to data of a particular type, prioritize or offer enhanced quality of service to all data of that type (regardless of the origin of such data) without imposing a surcharge or other consideration for such prioritization or quality of service. . .” That sounds like a QoS provision to me–along the lines that Professor Wu has advocated. What do you think about it? What kind of effect would this have on network innovation (i.e. improvements to the network itself rather than the applications, content and services on top of it).
(3) Thanks. I assume you mean that the fiber could probably handle IPTV and probably lots of other applications simultaneously. For example, Verizon’s announced FTTP build-out would leave 20% for non-Verizon services, which is about 19Mbps per household. I assume you mean by “a lot” is that IPTV and most other anticipated services would fit comfortably in the 19Mbps? (I am very aware that 19Mbps is plenty for almost all current household uses–it’s the future stuff I’m concerned about.)
One point of clarification: You are right that the Net Neutrality Act would not permit billing for QoS but it would permit broadband management through QoS tiering. Discriminatory billing, however, is not necessarily good. It would allow the telecoms and cable companies to effectively price QoS out of reach for all but the biggest firms, hurting innovation and start-ups in the services and applications markets. And ultimately consumers pay for it all, either (1) directly from the telecoms and cable companies or (2) directly and indirectly from the application/service providers who pass on the QoS fees. What is bad is loss of innovation from having more internet end points available to compete. I don’t really understand your “wholesale” comment.
There’s never been a law prohibiting ISPs from engaging in “content discrimination”. The common carrier regulations that used to apply to the telephone network simply said you could dial whoever you wanted and that you could choose your ISP, and they said nothing about transactions on the backbone.
Comcast, Earthlink, and others have portal pages with links to a preferred search engine, e-commerce sites, etc., but you don’t have to use them. So why haven’t ISPs engaged in content discrimination when it’s perfectly legal?
There’s a misunderstanding of QoS that runs throughout the Neutrino scare sites, to the effect that QoS speeds up the loading of web pages. According to Christian Coalition and others, the phone company wants to make some web sites faster than others. But the real value of QoS is simply to reduce the jitter for fixed-rate streams so that audio and video are more solid and better synchronized.
I can’t support a political movement that’s based on a lie.
As to the speculation that Comcast and AT&T will collude to reduce competition, this is why we have anti-trust laws, isn’t it? And we don’t apply them prospectively, we wait for someone to break the law and then prosecute. That seems like a perfectly functional model, and one that doesn’t need extra special billing mode restrictions to be operational.
There didn’t need to be law prohibiting content discrimination (perhaps until now) because in the competitive market (one with many players) that we had, consumer demand prevented it–your free market at work. The cable companies and the many ISPs were competing with each other and had to be responsive to consumer demand, and consumers presumably preferred not to have content discrimination. But the market structure is changing from many players to only two.
Game theory predicts that an interdependent relationship in a multiple game scenario is likely to yield a cooperative outcome, over time at least. In that environment, announcements about intended conduct can help facilitate coordination, which is what we have here with the telecoms announcing their intentions to discriminate.
The reason for regulatory intervention is that the antitrust laws are not equipped to prohibit interdependent coordination. You must have collusion, which requires an agreement. In this case, one probably would have to prove parallel conduct and “plus factors” because the internet providers are too well counseled to enter into an explicit, collusive agreement. We might not have the “plus factors” here (which is an antitrust notion that doesn’t square with economic theory all that well anyway). Judge Posner has advocated for a change in antitrust law to prohibit interdependent coordination, because its effect is the same as an explicit agreement, but the courts have not accepted his invitation. Thus, interdependent coordination (as opposed to a collusive agreement) usually escapes antitrust liability.
Of course, the telecoms and the cable companies might not discriminate against content, and we could wait and see. We could simply turn a blind eye to their announcements to the media and the FCC. But there are costs. Most notably, there is the loss in innovation–which has a cumulative effect–until legislative action is taken. And legislative action can be a very slow process.
Which brings me back to a question you still haven’t answered: What do you think about the soft version of net neutrality? You have argued here and elsewhere that content discrimination will help create an efficient allocation of bandwidth so that content needing priority gets it. Doesn’t the softer version of net neutrality, by permitting discrimination based on type of content, address this concern while preserving the innovation potential because of lower barriers to entry?
You have also argued that net neutrality will hurt innovation of the internet itself. I understand that argument as applied to a rule prohibiting any discrimination. Does the argument hold with the softer version of net neutrality? If so, how? Thanks.
(I doubt that you think the misunderstandings of the Christian Coalition or the misinformation on how the internet works is itself a reason to oppose the softer version of net neutrality.)
The soft version of neutrality requires the network’s owner to give priority service away to competitors for free, right? So all he can do is charge his own customers at either the basic rate for best effort service or a higher rate for QoS, but he can’t enter into any agreement with a third party to provide QoS to customers of the network who didn’t pay for it themselves. Thats’ the idea, right?
I don’t get that it’s a better alternative to free billing options for high QoS. It makes no difference to me whether the customer pays At&T for QoS or Skype pays AT&T for the same customer’s QoS. It’s a wash.
Now if you’re arguing that the network has to let anybody tag their packets for low-jitter service regardless of their service plan, that’s a poison pill. The whole concept of QoS depend on the resources being available to handle the QoS streams, and some sort of Admission Control is necessary to make that happen.
Regarding your comments on interdependent collusion, that’s a much larger issue that on-line service plans, and should be dealt with in the context of anti-trust law. Putting a band-aid on COPE isn’t a sufficient or proper solution.
If we develop inter-provider QoS, then we can have settlement-free inter-provider peering at different levels of QoS, which will get Google Video access to AT&T’s eyeballs at the higher level of service without paying AT&T directly. They can do that, or they can pay another provider that provides transit to AT&T at the higher level of service, either a tier 1 that does settlement-free peering with AT&T at the higher QoS, or a lower-tier provider that pays AT&T for transit. The definition of “tier 1” will become more complex, as a current tier 1 might not have the right traffic to exchange with another current tier 1 at a higher level of QoS and might have to pay for the privilege, while continuing settlement-free peering for the best-effort class.
That’s what I would expect to happen *without* any government regulation. Net neutrality would say that the evolution of these kinds of contractual arrangements between providers is prohibited.