MI-80 De Oppresso Liber

Alright, let's think about filtering the Internet logically from AT&T’s perspective. It's easy to say the idea sucks, especially if one fears that he will be filtered or have his privacy invaded. The real question, though, is does this even make sense for AT&T?

First off, there is the real issue behind this debate: bandwidth. The current collection of American internet service providers simply does not have the bandwidth it promises to customers. This varies, but one major regional ISP gave the factor of 2MB of bandwidth per month per customer. Perhaps that was more true ten years ago, but the current generation of consumers use much more than that. For every few grandmothers who check their email once a weekend, there is someone else consuming GB's of data each month eschewing the graph. But the fact remains – the contract a consumer signs says (for example) 8Mb/s is the speed cap; nothing is stated about limiting how much one can download or how often he can utilize that cap. In theory one could use that all day, every day.

Unfortunately the providers have not built their networks to those specifications. They know some percentage of users will be off-line at any given time, and that some percentage of users will be light users. So, in order to grow the customer base, they've increased their bandwidth cap while only increasing the network infrastructure enough to support the "real" load instead of the amount promised to patrons. If every Comcast subscriber went online tomorrow and started downloading at his capped speed, the Comcast network would simply collapse.

So, from a real-world perspective, this debate has nothing to do with doing the right thing. Providers are looking for an economic way to extend the useful life of their existing infrastructure; they believe that they can begin doing this by blocking illegal traffic (or peer to peer traffic as some providers make no distinction). On the publicity side they can say, "Look at what we are doing for content providers!" However, on the consumer’s side, who is going to call and complain that they cannot commit a crime? With limited choice, if any, of competitors to choose from, the consumers are left virtually powerless in this situation.

Regardless of whether or not this policy is invading customers’ privacy, the long-term ramifications bring the rationality into question regardless of the legal sense it may currently make. If filtering was to be flipped on tomorrow, there is no doubt that it would be effective. However, technology to defeat filtering has existed for decades; it is called encryption. Encryption is already protecting everything from online banking to wireless networks from eavesdroppers. It is only a minor evolution to pair this technology with illegal downloading to hinder content filtering. Then it becomes a question of what the carriers will do with their costly filtering infrastructures, only able to monitor legitimate traffic and illegal users lagging behind in upgrading their software of choice. With reports of government monitoring in recent years such a carrier level investment would stand as a boon for organizations like the National Security Agency who desire to capture every communication, everywhere.

Encryption has not been widely used in peer to peer networks because there has traditionally been no need for it. However, newer clients and protocols have begun to integrate it. The simple fact is that a vast majority of users rely on the tools created by a few select developers. What this means is that as soon as developers find a need to integrate encryption, they will, causing that huge number of users to slowly upgrade. Given the uptake and popularity seen in peer to peer software, filtering will not remain effective for any reasonable length of time before users begin to utilize encrypted channels and networks. Traffic levels will quickly return to the point that they were at before any of this ever happened. During this time the carriers will be paying untold volumes of money as their success rate quickly drops from day one.

So, does filtering make economic sense knowing that it is a short term patch? Does filtering make economic sense knowing that they are pouring money into a lost cause? In the long-term, is this really cheaper than building up their infrastructure?

Furthermore, this is all predictable. If the providers were to upgrade their networks tomorrow to support the offered speed, they would not have any worries; whether one downloaded content all day or not, he would not exceed his contracted speed cap. However, to fight off competition, they continue to increase their promised bandwidth. Charter is now offering 10Mb/s in markets threatened by fiber optic providers and Comcast is pushing their PowerBoost technology which offers quick bursts up to 12Mb/s during the beginning of a download or upload. Pairing these increases with record sized subscriber bases and little infrastructure improvement is bound to result in a collapse of the network if something is not done.

Unfortunately, high speed data services are an oligopoly in this country. Cable rates have increased greatly over the past decade in America. The trend far outpaces other industries as well as inflation. Compounding the poor situation is that money is not going into the network infrastructure. Instead, resources are spent deploying capabilities to more homes and pushing other services such as high-definition TV, video on demand, and home phone service. The providers want to push more traffic than ever over the network without improving it.

Future-proofing is also becoming an issue. The major players would like to develop a network for long-term use. For example the phone companies continue to use technology and systems developed and built nearly a hundred years ago. Likewise much of the technology created and deployed by the cable providers is anywhere from 20-50 years old. Effectively the network carriers are trying to fight off the inevitable; in high-technology anything beyond about five years in the future is unclear. It simply cannot be assumed that what one has now is going to last half a century. And, if increasing rates is not enough to build up the infrastructure to meet consumer demand, then markets will find new alternatives and the traditional providers will fall. Content filtering will not stop this, only pleasing the consumers will. Perhaps a research and development investment into content filtering will buy a short reprieve until encrypted peer to peer networks become the norm, but that only returns to the question: “Is this the answer?”

Ironically, the providers also stand to harm themselves. Clear-text traffic may be cached, saving a copy to serve to future users who request the same content. This is common with other protocols like web traffic. Forcing traffic to become encrypted makes it impossible to cache meaning each user is served a brand new copy. Consider World of Warcraft’s ten million strong user base, which downloads massive game patches via BitTorrent on a regular basis. The potential bandwidth savings from caching this data is massive. Encrypting it only ensures the exact same thing, possibly several hundred MBs, is passed along in its entirety ten million times.

Companies also risk the safe harbor protection they fought so hard to establish in the late 1990s. This stated due to their inability to police their networks they couldn’t’ be held responsible for infringing materials on them as long as they responded promptly to legal notices. Implementing a filtering solution opens up legal avenues of attack for what infringing content remains.

Some final food for thought revolves around the negative impact on society of forcing all this traffic to become encrypted. Suddenly malware can pass through perimeter defenses. After all, that firewall scanning for viruses can’t detect what’s inside the encrypted traffic. Also, other traffic which more people would agree should be tracked down, child pornography being the classic example, would be shielded along with the run of the mill movies and music. Does the internet provider industry really want to force these people to upgrade their defenses and become that much harder to track as well?