4 ms·

As far as I can tell, net neutrality is an interesting mythology that probably is not reducible to specific technical requirements in a long-term-meaningful way

by shimon 18y ago

As far as I can tell, net neutrality is an interesting mythology that probably is not reducible to specific technical requirements in a long-term-meaningful way. One one hand, there have been good rulings from the FCC, e.g. preventing ISPs from (anticompetitively) slowing down VOIP traffic that isn't handled by their own VOIP service. But on the other hand there are intrinsic physical limitations to the way networks are designed, so it's perfectly reasonable that if you have a Comcast internet connection your VOIP packets will have a faster route to Comcast VOIP servers than to Vonage. There's also a tricky distinction going on here. If you have a Comcast cable line going to your house, your network connection is already non-neutral. That cable carries video signals over IP on a certain share of the wire's bandwidth; this share is always dedicated solely to Comcast IPTV and no matter how many blip.tv videos you have going on your laptop your digital TV will always have the same dedicated bandwidth. Is that unfair? Comcast did a lot of work putting up wires to your house, navigating a bunch of annoying state and/or municipal laws, and investing in a lot of network infrastructure, for which they clearly deserve some reward. Is it OK to dedicate bandwidth on the wire but not over IP? Seems like a distinction that will get meaningless pretty quickly. On the other hand, there are many good reasons to be suspicious of telecom conglomerates, and it might be a reasonable policy position to say that comcast-the-internet-provider and comcast-the-cable-service should be separate companies. All I've heard from the Net Neutrality heroes, though, is vague and inactionable invective. It worries me not because I think it might work, but because it's (1) likely to generate incompetent legislation and (2) crying wolf now when we might well need coherent opposition to specific industry actions or proposed mergers in the future.

wmf 18y ago
Although it doesn't affect your argument, I feel compelled to mention that cable TV isn't over IP; it uses MPEG transport stream directly over QAM.
- eru 18y ago
  He did not say it's over IP.
- shimon 18y ago
  Good point; I meant to say "IPTV" somewhat generically. The important point is that it's a digital signal on a dedicated portion of the wire's bandwidth. I think Verizon's FiOS TV might be over IP, although I really have no clue.
acgourley 18y ago
Awesome, thanks for the well thought out post. It definitely seems like legislation is a solution in search of a problem right now. It seems like NN activists should instead focus on keeping the market place(s) competitive. Everyone needs more than one viable ISP, and every ISP needs more than one viable upstream network to resell. I don't think we're there.
davidjoho 18y ago
I favor NN regulation. I agree that it does not consist of specific technical requirements. It is more like (but not exactly like) free speech: Once you've agreed that it's desirable, you still have lots of arguments about how to apply it. And as with a principle such as free speech, the best way to explore its topology and applications is to look at examples. Your example of preventing an ISP from slowing down another company's VOIP traffic is a great one. Paradigmatic, even. NN, in my view, can be seen as a generalization from that example. I'd also want to exclude ISPs from deciding that, say, VOIP traffic gets priority over WoW traffic, because I don't want ISPs deciding what the Net is for. Rather than having to detect similar such slowdowns and then petitioning the FCC for redress in each specific case, I'd like to see the FCC adopt (through legislative mandate) a _general_ principle that ISPs are not allowed to prioritize packets (with some set of hard-to-work-out and ever-arguable exceptions). That would leave room for ISPs to offer plans whereby users can specify which types of traffic and/or which sites _they_ want to receive priority (for the last mile, at least), and would certainly leave room for ISPs to offer tiered service. My interest in NN is not theoretical, if only because I don't know enough -- and I greatly appreciate shimon's link and post. My motivation is exactly what shimon expresses in his third paragraph: The economic reality is that the large ISPs have a structural and just about unavoidable interest in treating the Internet as a way to give preferential treatment to their own content and services. For that reason, I'd rather have exactly the sort of structural separation that shimon mentions, in which case we wouldn't need NN regulation. This is a position shared by Yochai Benkler, Susan Crawford, and David Isenberg, to name three highly competent, reasonable, informed proponents. As for shimon's claim that NN is all "vague and inactionable invective," well, there are lots of highly technical, reasonable, and smart people who are talking exactly about actionable specifics, along with more than occasional invective. Tim Wu, Harold Feld, and David Reed spring to mind.

by shimon 18y ago

wmf 18y ago

Although it doesn't affect your argument, I feel compelled to mention that cable TV isn't over IP; it uses MPEG transport stream directly over QAM.

eru 18y ago
He did not say it's over IP.
shimon 18y ago
Good point; I meant to say "IPTV" somewhat generically. The important point is that it's a digital signal on a dedicated portion of the wire's bandwidth. I think Verizon's FiOS TV might be over IP, although I really have no clue.

acgourley 18y ago

Awesome, thanks for the well thought out post. It definitely seems like legislation is a solution in search of a problem right now. It seems like NN activists should instead focus on keeping the market place(s) competitive. Everyone needs more than one viable ISP, and every ISP needs more than one viable upstream network to resell. I don't think we're there.

davidjoho 18y ago

I favor NN regulation. I agree that it does not consist of specific technical requirements. It is more like (but not exactly like) free speech: Once you've agreed that it's desirable, you still have lots of arguments about how to apply it. And as with a principle such as free speech, the best way to explore its topology and applications is to look at examples. Your example of preventing an ISP from slowing down another company's VOIP traffic is a great one. Paradigmatic, even. NN, in my view, can be seen as a generalization from that example. I'd also want to exclude ISPs from deciding that, say, VOIP traffic gets priority over WoW traffic, because I don't want ISPs deciding what the Net is for. Rather than having to detect similar such slowdowns and then petitioning the FCC for redress in each specific case, I'd like to see the FCC adopt (through legislative mandate) a _general_ principle that ISPs are not allowed to prioritize packets (with some set of hard-to-work-out and ever-arguable exceptions). That would leave room for ISPs to offer plans whereby users can specify which types of traffic and/or which sites _they_ want to receive priority (for the last mile, at least), and would certainly leave room for ISPs to offer tiered service. My interest in NN is not theoretical, if only because I don't know enough -- and I greatly appreciate shimon's link and post. My motivation is exactly what shimon expresses in his third paragraph: The economic reality is that the large ISPs have a structural and just about unavoidable interest in treating the Internet as a way to give preferential treatment to their own content and services. For that reason, I'd rather have exactly the sort of structural separation that shimon mentions, in which case we wouldn't need NN regulation. This is a position shared by Yochai Benkler, Susan Crawford, and David Isenberg, to name three highly competent, reasonable, informed proponents. As for shimon's claim that NN is all "vague and inactionable invective," well, there are lots of highly technical, reasonable, and smart people who are talking exactly about actionable specifics, along with more than occasional invective. Tim Wu, Harold Feld, and David Reed spring to mind.