I have the feeling that most of the world--not to mention the computer industry--has been following the Net neutrality issue in the United States. No wonder; seldom do we see an issue so fundamentally difficult to grasp and so clouded with confusing rhetoric. It may not seem like this to the fierce partisans (paid and unpaid), such as those marching the halls of Congress on behalf of one interest group or another. Nor may it seem confusing to those whose knees jerk in synchronicity with calls for Internet fairness or the unfettered free market. However, though confused as any, I did suspect that under the raveled and sometimes intentionally twisted Net neutrality arguments there are things worth thinking about. Foolishly treading where angels fear, I'll try to do some unraveling in this column and the next.

I'd start with an acceptable description of the Net neutrality issue, but unfortunately such descriptions have so often been embedded in ways to cast, frame, or bias the issue (take your pick) that almost any approach will be seen as falling into one camp or another. But what the heck, I'll identify the Net neutrality issue with three things: A principle that guides the Internet, the technical reality of the Internet, and making money.

Principle Interest

It's generally agreed that one of the reasons the Internet has been so successful is that it was founded on a set of goals and principles that guided and enabled its growth. One such principle is the idea that the network itself is "dumb." In other words, it knows not and cares not what is contained in the packets of data that whiz across the network. Packets should be moved, by whatever route, as rapidly as possible, from whatever source to whatever destination. In short, the Internet is a packet-data transport mechanism that is content-neutral, which these days is what most people recognize as Net neutrality.

Current use of the Internet is dominated by an enormous volume of relatively small transmissions (a small number of packets)--Web pages, e-mail, search requests, and so forth. The original content is broken up into packets, the packets travel by the best available route to the destination, and at the end point the packets are reassembled into the original content. With small transmissions, the Internet does this quite efficiently.

However, we are at the beginning of using the Internet for streaming data--movies, radio, television, voice (phone). By definition, streaming involves a massive number of packets that must be delivered in a timely manner. Here latency of the Internet comes into play. The more packets in a transmission and the farther they have to go, the more likely some packets get lost. The system has to request replacement of lost packets and they are retransmitted. Of course this takes time and holds up the assembly of a transmission at the destination--creating a lag (latency in the system). Lags may not be a problem for e-mail, which sits around anyway, but for movies, TV, or telephone conversations, latency degrades the quality of what you see or hear, sometimes a lot.

There are many ways latency can be addressed, but the two most mentioned at the moment are faster transmission--usually by installing fiber-optical transmission lines--and preferred transmission, which means that some packets move through the system faster than others.

Who Pays, Who Profits

Much of the U.S. Internet, especially the fiber-optic, nation-spanning backbone transmission lines, was built with government subsidy (about $280 billion worth). The current performance upgrade to the system is installation of fiber-optic lines on the "last mile" to homes and businesses. While it was understood that backbone lines would be used by everybody and controlled by nobody, there is no such understanding for last mile lines. Those who spend the money to put in the last mile lines, typically the cable or telephone carriers such as Verizon, AT&T, or Time-Warner want control of those lines and profit from their investment. There is some dispute about this, but it fits within the current practice of carriers owning the cable and telephone lines and charging both users and providers for access to the Internet. We've also generally accepted paying more money for more bandwidth.

What is relatively new (and the source of the fiercest debate) is the notion of stepping up the speed of transmission by giving some packets priority. New router technology makes it possible for Internet carriers to differentiate packet by source, content, and destination and treat some preferentially. These routers turn the Internet into a "smart" network capable of analyzing and acting on the packets. Network carriers are suggesting that like Internet access bandwidth, content providers, particularly those with streaming data, should pay for preferential treatment of their packets. The carriers say this will pay for redressing the inherent bias of the Internet against streaming data.

Packet Priority

It is on the issue of paying for prioritizing packets that the rhetoric becomes thunderous and the confusion endemic. Content providers such as Google, Yahoo, and Microsoft, naturally, hate the idea--it complicates their business model. Obviously, analyzing and handling packets based on their origin, content, or destination is not what the principle of net neutrality is about, so this is where the defenders of net neutrality join the battle. They and the content providers want Congress to ban or regulate the ability of Internet carriers to charge for packet prioritization. The carriers and their supporters countercharge that a Congressional mandate for net neutrality is unnecessary government intervention, and that in a number of ways will lead to no good.

I'll plunge into some specific arguments in the next column, but I'll note here that there are big problems with the debate: Almost everyone is arguing on the basis of speculation about the extreme cases: what might happen if net neutrality is totally abandoned, or if a government ban on packet prioritization is applied. Also, when the subject is complicated and assertions are difficult or impossible to prove, then there is a counter-argument to almost everything. For example: "Internet carriers will have too much power over what content will be provided," countered by "Competition will give people ways around the control." Finally, as important to the future of the Internet as this issue probably is, it has become quite politicized, which has kicked what is almost a technical debate into deep philosophical waters.

Under the circumstances, is there any way to cut through some of the clutter and confusion? I think so, and I'll spell out how next time.

Nelson King writes Pursuits bimonthly for ComputerUser.