The calls for a US “network neutrality” rule get louder and louder and louder.
I’ve covered this topic at length before. But here’s the real issue in a nutshell.
Would you want to make it illegal for at&t to offer a $5/month plan to poorer households that only allowed access to services by Yahoo!?
The proposed neutrality rules would do just this, hurthing the weakest in society most. Perhaps the Internet is supposed to become a polite, middle-class over-educated ghetto, but that’s news to me. (Personally, I’d prefer to make typographically-challenged corporate identities illegal first!)
Network neutrality rules, apart from being a fuzzy undefined concept, throw out the best of price discrimination and entrench the false “natural duopoly” for access.
There are deeper, more philosophical reasons too for being suspicious of such rules. Underlying the idea of “neutrality” is the idea that Internet Protocol embeds no assumptions or values, and the mesh of interconnect agreements we call the Internet is likewise “neutral”.
You probably won’t fall off your chair with surprise if I tell you this is nonsense. The Internet is riddled with design and political assumptions, and we should be open to competing architectures emerging.
Exhibit A. The Internet takes the “end-to-end” philosophy of distributed intelligence well beyond the application layer. There is no equivalent of the GNU General Public License for the Internet that you have to agree to in order to join. Repressive, filtering China can happily peer with anyone else. Everyone is free to negotiate their own interconnect at the financial and political layers.
You might want to contrast this with the PSTN, where common carriage agreements flatten the possible set of interconnect agreements. There are plenty of “managed” networks in telecom with standard rulesets to be abided by. The resulting networks have a more homogenous nature. This might be a good thing, or it might not. Let the customers decide. Would an “IMS-net” with universal rules for interconnect create more value than the deliberate emergent, chaotic Internet? I have an opinion, but I certainly don’t want to make the experiment with other people’s shareholder capital illegal!
Exhibit B: The semantics of IP addresses follows a particular “lowest common denominator” format. The databases held by ARIN, RIPE etc. allow you to trace back from an IP address the ISP to which it is assigned. This is probably the minimum possible assumption that can be embedded in the network, so follows the “end-to-end” philosophy, although is outside the scope of its original concept. But we can imagine networks where these databases offer much greater detail on the nature of the end nodes. Is it a good anti-fraud device for me to know you’ve been personally assigned that address for 3 years? For me to be able to verify with your telco (for a fee) that the delivery address you gave for the plasma TV is the same as the premises address of your IP address?
Exhibit C: The Internet doesn’t allow “topological introspection”. OK, protocols like BGP in principle allow this. But you don’t get access to that data as an edge node. Yes, you can traceroute your way around all day. But as a matter of universal fact, you can’t from the edge determine the real topology of anything in the middle. Internet Protocol assumes the world is, indeed, flat. Which is fine, until you hit the limits of that abstraction. I’d rather stream my P2P IPTV from someone else the minimum possible number of hops away. We have kludge it, but it’s not pretty.
In fact, there are lots of ways in which the current Internet design could be improved, and many people working on doing that just now.
Having a “two tier” network is something we should look forward to. We want more Internets! Plural! They may continue to interconnect; they may decide that the Internet Mk1 is more a source of digital pollution than valuable content. I just can’t see how any “stop the clock — we’re all just comfortable as we are!” neutrality rule helps us reach new and better places. Even monopolies have an interest in deploying new and more valuabe (monopoly) stuff. Kinda hard when by law you’re only allowed to offer Net service as experienced via dial-up modem c. 1997, only faster.
Indeed, this is all reminiscent of the arguments about socialised medicine in the UK. “We must spend more on the National Health Service to prevent a two-tier system emerging!” Yet lots of people who can afford it opt-out and but private medical care. Likewise, ossifying and constricting the Internet’s rules of engagement will just result in a hidden transfer of traffic onto other, completely private networks outside of the neutrality rules. Do you really think the Baby Bells won’t be able to buy some finessing get-out clauses? This is a much easier lobbying problem that undoing the whole unbundling regime of the ‘96 Act!
So neutrality rules that entrench our “Internet Mk1” as somehow sacred, hallowed and for all time are just totally counter-productive. Better to allow Verizon to screw over their customers and make it worthwhile for someone to bypass them entirely using newer technology. Or just swallow your pride and copy the unbundling rules that work just fine over here. BT can deploy a two-tier walled IMS garden, if they like. Just they have no way to make me buy it unless it creates some compelling value.
Posted by Martin Geddes at 09:02 AMTrackBack URL for this entry:
http://www.telepocalypse.net/cgi-sys/cgiwrap/mgeddes/MT/mt-tb.cgi/655.
Listed below are links to weblogs that reference Neutrality, schmootrality: a heretic speaks:
»
Cerf lobbies Senate (and Trek UI) from Computerworld Blogs
In today's IT Blogwatch, we look at network neutrality. Not to mention is it art or is it a Star Trek UI? ...
[Read more]
»
Neutrality, schmootrality: a heretic speaks from Skype Journal
The calls for a US “network neutrality” rule get louder and louder and louder. I’ve covered this topic at length before. But here’s the real issue in a nutshell. Would you want to make it illegal for at&t to offer...
[Read more]
Actually, you've solved a problem, but it's U.S. education, not the Internet. If we simply take the common sense measure of allowing Monsanto to sponsor our Biology curriculum, allow Snapple and Pepsi to be not just sold in schools but put onto the lunch menu, then we'll have no education problem.
I applaud your intelligence on this issue! Anyone who opposes you hates our freedoms, um, hates poor people.
Posted by: at February 8, 2006 07:12 PMUntil there's competition between broadband access providers (today I can only get BB from cable, can't even get DSL as the copper's too impaired), the monopolists need to be regulated. I'm sure you can trump up esoteric examples of benevolent services that would be blocked by "fair playing field" regulation, but I'm more concerned about the Verizon's and AT&T's giving unfair network benefit to their voip and video over Internet applications.
I find humor in your example as I worked for AT&T. Your sample offer would slog through the AT&T product process and result in service likely more costly than unregulated broadband.
There is nothing wrong with a two tiered network, just make sure the "fast tier" is implemented in a fairly priced manner. A carrier's own fast tier applications must have the same cost burden as those from third parties.
Posted by: at February 9, 2006 02:57 PMIf unfair network benefit translates into lower consumer welfare, then doing it is the best way to get more competition. We can either have price-regulated monopolies or we can let entities with market power try to exploit it and in doing so provide greater incentives for entry. Regulators are not smart enough (and maybe not incorruptible enough) to come up with rules that prohibit price discrimination that reduces consumer welfare while allowing price discrimination that enhances consumer welfare.
Posted by: at February 10, 2006 03:08 PMShorter Geddes: Censor the poor - it's for their own good!
Posted by: at February 13, 2006 02:43 PMCan someone explain to me why telecom stockholders need to be the ones to fund your (undoubtedly noble) social program goals?
Or try this one on for intellectual size: we'd like everyone to have an open, symmetric, 100mpbs fibre connection. So we'll make that the only legal retail broadband product. After all, why stop at only one social objective?
How quickly do you think capital will come flooding into the sector as a result?
Price, quantity, quality. Fix any two.
Posted by: at February 13, 2006 03:12 PM