I raised the topic of Internet censorship in a recent posting, Wikileaks and the Streisand Effect and I’ve since been corresponding about it with others interested in the topic.

One of the fundamental problems of the Internet is that freedom of speech means different things in different countries. For example, The Anarchist’s Cookbook is banned in the UK, because it contains instructions on how to build bombs and do damage using chemicals and components that can be easily obtained. In the US the book isn’t banned because of the First Amendment. “It is my right as an American citizen to tell people how to make bombs.” This is a fairly mild disagreement when compared to how the attitude of Islamic countries to blasphemy contrasts with that the US and Europe.

One of my correspondents maintained that “Traditional top down regulation and censorship systems will not work in a bottom up system such as the web”. This is a view that’s worthy of some analysis.

The test case “par excellence” is China. China illustrates that while the traditional censorship approach runs into problems of practicality – banning access to everything is not going to work – banning access to most things is probably enough, if done in the right way. I don’t believe that the Chinese government currently thinks that its censorship system is failing – and this has a lot to do with how you define success.

The world seems to have forgotten that, as far as the instruments of power are concerned, the current Chinese government is a direct descendant of Mao’s Stalinist regime. Many political writers have concluded that because it at first sanctioned capitalism and then strongly promoted it, the Chinese government suddenly became “almost like a democracy”. It didn’t. Two quite different political assertions have been confused; “capitalism is far more efficient than a planned economy” with “democracy is far better than despotism”. China acknowledges only the former.

Culturally China has no tradition of democracy and it had no compunction about crushing democracy in Tiananmen Square in 1989. In the 1990s it paid big money to launch the Golden Shield project – with the goal of achieving effective Internet censorship for its citizens.

The Golden Shield Mechanisms

The Golden Shield works in the following way: Right now there are only three pipes that connect China to the Internet. Two go through Japan and one through Hong Kong. China controls the routing through these pipes. It indulges in:

1. Direct IP blocking (not very efficient because you block whole sites including some unintended targets)
2. DNS poisoning (which you can get around simply by knowing the IP address rather than the URL)
3. URL filtering i.e. random blocking of URLs that contain specific words (again you block unintended targets)
4. Packet filtering i.e. terminate communications if specific packets contain banned words (again you block unintended targets).
5. Web feed and blog blocking. Kill all RSS feeds and blogs because they cannot be trusted (again you block unintended targets).

The effectiveness of the Golden Shield comes from the fact that it is well automated. But there is a downside; links going outside China have an unusually high latency, because some of the blocking techniques slow down all traffic. China doesn’t care, because it’s simply another disincentive to access sites outside China. They’ve even created a censorship free area for the duration of the Beijing Olympics!

It’s been suggested that because the number of possible connections within the Internet rises exponentially as the Internet grows, China’s great firewall will eventually fail to keep pace. But it’s not the number of connections in total that will cause a technical problem, it’s the number of attempts to access connections outside China. This number probably rises linearly rather than exponentially (i.e. every new Chinese user adds to this “attempted connections” figure linearly) so the needed compute power to keep the Golden Shield in place may never hit a brick wall. It is however possible that growth could be worse than linear because of other network factors – such as the number of truly useful links between China and elsewhere growing exponentially. It’s moot.

Another aspect of the Golden Shield is that, every now and then, China arrests some dissident or other and punishes him or her severely for “activities against the state”, which can be as trivial as getting past the Golden Shield.

The Golden Shield “Algorithm”

China’s censorship “algorithm” is quite effective, I believe, because it leans on several dependable psychological traits:

• Most human beings are impatient
• Most human beings are lazy.
• Most human beings are apolitical or not activists.
• Most human beings have a fear of authority
• Most human beings are technophobes

It isn’t necessary to stop access to banned sites, it’s only necessary to make such access burdensome to achieve and let these human tendencies kick in. China does want to stop dissidents particularly, it wants to prevent a popular uprising of the Tiananmen variety.

The cyberspace antidote to China’s censorship is the Streisand effect (virtual civil disobedience) and thus far China has kept this under control. I’m not sure that it can continue to do so. However, depressingly, I can also see how it might improve its current level of effectiveness.

1. It’s feasible for China to cache the whole Internet onto servers in China and effectively prevent any external access by China residents. (As time passes this become increasingly feasible because of leaps and bounds in networks speeds and storage technology)
2. Using semantic technology, such as is offered by Cogito, the Golden Shield could be far more effective in determining the meaning of web pages i.e. what is and is not “seditious”
3. China could because of 1. and 2. above focus more strongly on the dissidents and isolate them more effectively.
4. Use China’s financial muscle to support the operation. (For a while China redirected all search traffic away from Google’s chinese site to China’s Baidu.com. A relatively unreported event that was equivalent to an act of commercial war).

The Music Business: A Parallel

It’s interesting that after a number of years of floundering to understand the Internet at all, the music business is finally making commercial sense (see Nokia, Ominfone & Apple: All you can eat?). The grand irony of the music business is that the young people of today will not pay for music, but they will pay for ring tones. That apparent paradox is the direct result of the music business’ flirtation with DRM and its attempt to implement it, when it was impractical.

The music business wanted to censor your access to and participation in file sharing services. DRM on its own never worked and was never going to.

The solution that’s developing is not so different, in spirit, to the Chinese one – the music industry will provide the consumer with an “all-you-can-eat” service. The price will be hidden in the cost of a device (music player/phone) which you shed regularly on a 18 month-2 year basis. It then becomes easier and cheaper to “apparently” not buy music than to actually not buy music.

I have little doubt that this strategy will work well on impatient, lazy, apolitical, fearful technophobes. Stealing music will involve far too much effort by comparison. In time the music theft mechanisms will fade away through lack of use. Look out your window and see.