Because the existing internet architecture doesn't guarantee that your packets will necessarily arrive, and because streaming video needs high speeds and some predictability, lots of sites work with middlemen (like Akamai, as in "One Out Of Every Five Global 500 Companies Trusts Akamai").  These middlemen in turn buy up strategic bandwidth from ISPs all over the world.

So the streaming site (say, Fox News) sends its content to a cache.  Akamai takes the content and distributes it to servers that are close to ISPs.  From a user's perspective, all of this is invisible.  He makes his request of Fox, and, unbeknownst to him, he gets (for example) text from Fox and streaming video from a different route (a closer route) coming from an Akamai cache.  Akamai helps Fox create a better user experience. 

This all costs money.  Akamai has to pay for servers and additional ISP connections.

The argument from at least one telephone company in favor of a "tiered internet" is that paying for good service is nothing new.  Fox pays now to make sure that its users are happy.  If, say, a telco starts only prioritizing video packets of its partners, in order to get this value-added service Fox will have to pay for it.  What's the big deal?  Fox and many other sites already do all kinds of things to ensure that their users are happy.  They don't actually rely on the standard end-to-end internet model -- they cache and adjust and pay through the nose.

The key semantic question seems to be:  if services won't or can't pay for the telco special value-added services, are they being degraded? The telco would say "no -- you pay for first class on airplanes, you pay for tolls to enter cities, but those who don't pay aren't harmed."  Their point is that they should be allowed to charge for special services, that Fox and others pay for these kinds of services all the time, and that no one will be hurt by the mere offering of these services.

The risks are greater than just hurt feelings and misunderstandings.  If a telco makes an exclusive deal with any video streamer, then the rest will be second-best.  Maybe not worse off than they are today, but not as zippy as the premium service.  And start-ups may not be able to pay for any of this.  Plus, there may be future start ups that are unrecognizable to us as start ups -- amorphous net thingies without a home or a terrestrial bank account.

Harold Feld was at the same meeting with me today, and he's focused on what comes out of all of this prioritizing.  Is the result "the internet"?  In his words, "What the $#@! is the 'Public Internet'"?  This term (another semantic struggle) is being used to refer to the non-prioritized offering of an incumbent.

It's all pretty difficult.  It does seem to me that choosing to pay Akamai to buy cleverly-situated market-price ISP service for you is different than having to pony up to be carried to particular subscribers.  It also seems to me that, yes, services that aren't "prioritized" are necessarily, over time, degraded -- user expectations will change to be satisfied only with the highest speed access.  (I get impatient ordering a sandwich in DC; this has happened to me only since I moved to NY where the deli-guys scare and intimidate you with their alarming speed.)  I'm very worried about the risks to unborn technologies. 

And, as Harold says, the whole "public"/"private" new distinction seems very strange indeed.