And Akamai pays the big eyeball customer providers like the telcos, to put servers directly on their networks to get them close to the customers. The Akamai model *does* lead to the content providers paying the telcos, indirectly through Akamai.

Should that arrangement be prohibited? If not, then why should a direct arrangement between content providers and eyeball customer providers be prohibited?

And Akamai pays the big eyeball customer providers like the telcos, to put servers directly on their networks to get them close to the customers. The Akamai model *does* lead to the content providers paying the telcos, indirectly through Akamai.

Should that arrangement be prohibited? If not, then why should a direct arrangement between content providers and eyeball customer providers be prohibited?

First, I don’t recall exactly what I said, but I was trying to avoid getting too technical, so I wouldn’t be surprised if I did conflate QoS and bandwidth at some point. But I’m actually pretty sure I *did* explicitly make the distinction between speed and reliability or quality, precisely with respect to voice. In fact, unless I’m misremembering, the only time I really focused on VoIP was to make just that distinction. If I got sloppy somewhere, mea culpa. But I don’t think it’s fair or accurate to say I entirely neglected it.

Second, the relevance of Akamai is that the way at least some neutrality advocates are presenting their case is in terms of this idea that companies with a lot of money to throw around shouldn’t have any advantage vis a vis some small startup in terms of getting their content to users faster. And insofar as that’s not actually the case under the status quo, I think it was germane to *that* point, whatever other distinctions one might draw on other grounds. I wasn’t invoking Akamai as some kind of all-purpose answer to every issue people have with non-neutrality, only as a small counterexample to this idea that the net in its current form is some kind of utterly level playing field where money doesn’t play any important role in determining how fast your content gets to users.

First, I don’t recall exactly what I said, but I was trying to avoid getting too technical, so I wouldn’t be surprised if I did conflate QoS and bandwidth at some point. But I’m actually pretty sure I *did* explicitly make the distinction between speed and reliability or quality, precisely with respect to voice. In fact, unless I’m misremembering, the only time I really focused on VoIP was to make just that distinction. If I got sloppy somewhere, mea culpa. But I don’t think it’s fair or accurate to say I entirely neglected it.

Second, the relevance of Akamai is that the way at least some neutrality advocates are presenting their case is in terms of this idea that companies with a lot of money to throw around shouldn’t have any advantage vis a vis some small startup in terms of getting their content to users faster. And insofar as that’s not actually the case under the status quo, I think it was germane to *that* point, whatever other distinctions one might draw on other grounds. I wasn’t invoking Akamai as some kind of all-purpose answer to every issue people have with non-neutrality, only as a small counterexample to this idea that the net in its current form is some kind of utterly level playing field where money doesn’t play any important role in determining how fast your content gets to users.

You cover why Yglesias’s arguments don’t necessarily make sense, but neither do Sanchez’s. He fails to distinguish between bandwidth and QoS, mixing up the two quite frequently. He claims VoIP needs a lot of bandwidth — which is flat out wrong. It needs QoS, not very much bandwidth. Voice is not a high bandwidth app.

He also confuses the issues with his Akamai example. In that case, it’s about the web service provider choosing to improve performance for all visitors, no matter who their service provider is. It’s an extension of what happens now. Right now, they pay their own service provider for the bandwidth used. And they pay Akamai to make that bandwidth more efficient.

What the telcos are trying to do is on the other side, get the web service provider to pay for speeding up access just for certain customers (so, paying AT&T; to improve performance for AT&T; customers, for example).

What it boils down to is the question of what you pay for when you buy internet access and bandwidth. Are you paying for just the connection to the center that the ISP provides, or are you paying for the connection to everyone else on the network? Traditionally, it’s been believed that you pay to connect the ends. After all, the value is in the network effects of connecting all those ends. Without that, the center isn’t that useful. The telcos are now saying that’s not true, that you only pay to connect to the center, and the “free ride” is the other half — getting out to the ends. Now, they want people to start paying for the other half. That’s totally separate from the Akamai situation, and still seems an awful lot like double-billing to me.

You cover why Yglesias’s arguments don’t necessarily make sense, but neither do Sanchez’s. He fails to distinguish between bandwidth and QoS, mixing up the two quite frequently. He claims VoIP needs a lot of bandwidth — which is flat out wrong. It needs QoS, not very much bandwidth. Voice is not a high bandwidth app.

He also confuses the issues with his Akamai example. In that case, it’s about the web service provider choosing to improve performance for all visitors, no matter who their service provider is. It’s an extension of what happens now. Right now, they pay their own service provider for the bandwidth used. And they pay Akamai to make that bandwidth more efficient.

What the telcos are trying to do is on the other side, get the web service provider to pay for speeding up access just for certain customers (so, paying AT&T to improve performance for AT&T customers, for example).

What it boils down to is the question of what you pay for when you buy internet access and bandwidth. Are you paying for just the connection to the center that the ISP provides, or are you paying for the connection to everyone else on the network? Traditionally, it’s been believed that you pay to connect the ends. After all, the value is in the network effects of connecting all those ends. Without that, the center isn’t that useful. The telcos are now saying that’s not true, that you only pay to connect to the center, and the “free ride” is the other half — getting out to the ends. Now, they want people to start paying for the other half. That’s totally separate from the Akamai situation, and still seems an awful lot like double-billing to me.