Friday, June 01, 2007
Treating Different Types of Communications Differently
A friend who read my Creating Sustainable Network Neutrality paper wrote to say, "Help me understand what is so bad about treating different types of communications differently."
That's a really good question! My answer:
1. If you want to offer vertically integrated services on special purpose networks, such as video entertainment or pager service or telephony, I do not have a problem with that, provided you don't use your market power to impede Internet applications that offer competing services. (In fact, many of these are extremely useful; consider for example the Cisco technician who deals with router outages, you don't want to depend on the Internet to reach her, because the Internet service at her location is by definition on the blink, but instead you'd better use a non-Internet-dependent pager or cell phone.)
2. On the Internet, you don't need to treat different types of communications differently. In every case, simply adding more capacity is more cost effective and more reliable in getting the packets where they need to go than treating different types of communications differently. This is a conclusive finding dating back to Andrew Odlyzko's work at Bell Labs in the early 1990s. So if carriers want to treat different types of Internet traffic differently, they must be doing it for other than operational reasons -- e.g., because discriminating different kinds of traffic is more consistent with their business model.
3. On the Internet, treating different types of communications differently is anti-innovative. One of the strengths of the Internet is its stupidity, that is, its ability to carry all kinds of traffic without "asking permission" for special kinds of carrier services. So we have this great Internet thingy that you and I can use for our own private application, where we can be as inventive as we want without having to pay for some special service, without waiting for clueless execs to approve anything, without having to fit our new service into some pre-existing category, etc.
4. On the Internet, if you treat different types of communications differently, it is a slippery slope. Once the machinery for treating different types of communications differently is in place, the people in charge of that machinery might want to give their financial partners an advantage. Or they might want to give ideologies they like an advantage, and suppress the ones they think are dangerous.
There's more, but I think these are the high-order bits.
That's a really good question! My answer:
1. If you want to offer vertically integrated services on special purpose networks, such as video entertainment or pager service or telephony, I do not have a problem with that, provided you don't use your market power to impede Internet applications that offer competing services. (In fact, many of these are extremely useful; consider for example the Cisco technician who deals with router outages, you don't want to depend on the Internet to reach her, because the Internet service at her location is by definition on the blink, but instead you'd better use a non-Internet-dependent pager or cell phone.)
2. On the Internet, you don't need to treat different types of communications differently. In every case, simply adding more capacity is more cost effective and more reliable in getting the packets where they need to go than treating different types of communications differently. This is a conclusive finding dating back to Andrew Odlyzko's work at Bell Labs in the early 1990s. So if carriers want to treat different types of Internet traffic differently, they must be doing it for other than operational reasons -- e.g., because discriminating different kinds of traffic is more consistent with their business model.
3. On the Internet, treating different types of communications differently is anti-innovative. One of the strengths of the Internet is its stupidity, that is, its ability to carry all kinds of traffic without "asking permission" for special kinds of carrier services. So we have this great Internet thingy that you and I can use for our own private application, where we can be as inventive as we want without having to pay for some special service, without waiting for clueless execs to approve anything, without having to fit our new service into some pre-existing category, etc.
4. On the Internet, if you treat different types of communications differently, it is a slippery slope. Once the machinery for treating different types of communications differently is in place, the people in charge of that machinery might want to give their financial partners an advantage. Or they might want to give ideologies they like an advantage, and suppress the ones they think are dangerous.
There's more, but I think these are the high-order bits.
Technorati Tags: NetworkNeutrality, StructuralSeparation
Comments:
How does point #2 apply to a "wireless internet" situation? Wireless capacity can face real constraints due to inability to acquire more spectrum or inability to put up more towers. Under these conditions, would you support treating Internet traffic differently in order to e.g. ensure that voice calls could get through?
Post a Comment
