Akamai achieves its performance gain by degrading the performance of the non-Akamai web. It does this by putting its packets ahead of the non-Akamai web’s packets on the last mile. The trick that it uses to accomplish this is location, and the effect is exactly the same as prioritizing one stream over another stream on the backbone.

If the effect is the same as non-neutral service tiering, the regulatory framework should also be the same. The only real difference between what Akamai does and what AT&T might like to do is who’s doing the privileging/degrading. We don’t have one set of rules for one company and another set for another company, do we?

Well their claim is that the effects are not the same–that AT&T might achieve higher throughput for some users or applications by degrading the performance of others’ network connections. Whether that’s a reasonable concern or not (I tend to think it’s unlikely) it’s certainly an intelligible concern, and not difficult to distinguish from the Akamai case.

Obviously, any traffic on the Internet can affect any other traffic by increasing congestion. But I think there’s an intelligible argument that this generalized increase in congestion is less troubling than an ISP deliberately degrading disfavored traffic.

Again, I’m not saying I agree with this argument, but I don’t think there’s anything hypocritical about it.

Right, and that’s silly because both Akamai and creating the Akamai effect by other means have exactly the same effect on the network.

If you never visit a site that employs Akamai, your Internet access won’t be affected at all by its existence.

That’s not at all true. The effect of Akamai is to put more traffic on my ISP’s network faster, and that’s exactly the same as a system that brings traffic from distant shores into my ISP’s network faster. Everything that anybody does on the Internet affects everybody else in some degree because packet networks use shared media.

So the Akamai analogy is technically sound, even if civilians don’t understand it.

But the objection isn’t about the effect, but about the means used to achieve the effect. What they’re afraid of isn’t that peoples’ Internet service will work better. They’re afraid that in the process of making some peoples’ Internet service work better, they’ll degrade the quality of other peoples’ Internet access.

The reason that’s a concern is that consumers have very few options when it comes to “last mile” broadband service. So if your cable and telco broadband providers both do something stupid with their network, you’re stuck with it.

Now, I think those concerns are overblown. But it’s not difficult to see how Akamai is a different case. The use of Akamai is purely optional. If you never visit a site that employs Akamai, your Internet access won’t be affected at all by its existence.

No one is opposed on principle to faster or more reliable Internet connections. Their concerns are that in the process of trying to improve some peoples’ Internet access, they’ll degrade other peoples’. Again, I don’t think that’s likely to happen, but let’s not throw around spurious charges of hypocrisy.

But maybe I’m also lacking in mental agility and integrity.

Here’s a simple example: network A provides users limited delay for streams up to a certain limit, where the limit is sufficient for 10 phone calls. Network B provides variable delay but accepts a much greater load. It would be reasonable to use network A for voice and network B for Bittorrent and web surfing.

1. For prioritization to work on the Internet, it’s not necessary for everyone to adopt it. Rather, it’s simply necessary for there to be distinct paths for adopters and non-adopters. These exist today.

2. The Akamai example is interesting precisely because nobody sees it as an illegitimate service, at the same time that neuts see a Telco offering that has precisely the same effect as illegitimate. It’s a double standard.

To bring it closer to the neutrality issue, consider the case of WebEx, a private network overlayed on the public Internet that provides QoS to paying customers. I don’t hear anybody saying it’s either dysfunctional (as you claim any such service must be) or illegitimate (as QoS provided by Telcos is under NN law.) Why is that?

I believe it’s because NN advocates lack the mental ability, integrity, or technical awareness to pick up the contradiction.

It seems to me that Prof. Yoo does a reasonable job of explaining NN to the non-technical audience. That’s always a challenge, of course, but it’s a noble one and he does it as well as anybody and better than most.