Nart Villeneuve and Michelle Levesque are starting our afternoon in London with a detailed overview of techniques to detect and circumvent internet censorship. It’s a hugely useful talk – I’m enjoying learning some of the tricks Nart and Michelle use to figure out what’s being filtered.
Lots of what these guys learn about what’s going on with filtering techniques is through the block pages employed by ISPs that censor their users’ access to the web. A blockpage will sometimes tell what software is being used to block the internet, and (occasionally) why a page is being blocked. (Nart’s got a great gallery of these block pages, from nations, businesses and universities. Burma is my favorite.
Another critical tool for figuring out blocking technologies is to track the headers a web browser is receiving from a webserver. Nart and Michelle use LiveHTTPHeaders, a plugin for Mozilla that allows you to see the HTTP headers as a page downloads. Sometimes a block page (the sneaky ones) will give you a 404 error – page not found – suggesting that a page has been deleted or moved. Watching the headers, you can see that the page served is actually a 401 – forbidden. In other cases, providers redirect some sites to other pages – this will show a 301 or 302 redirect in the headers.
These guys also use a very nice kit of traceroute tools to see how IP blocking works. This includes not only conventional (ICMP) traceroute, but TCP traceroute, layer 4 traceroute and packet sniffers like Ethereal.
Nart and Michelle also add some useful suggestions that I’m planning on using in long-promised update to the anonymous blogging guide. They suggest that bloggers who know their sites are going to be blocked plan ahead, getting accounts with multiple ISPs and registering multiple domain names. They recommend switching domain names and IP addresses periodically to avoid blocking by government censors, and setting up a technical mirroring solution, so that when one domain is closed down, you can quickly move to another one. They also suggest licensing your content under creative commons, so people feel free to mirror it and to make sure third-party aggregators, like Bloglines, know about your site so they cache copies via RSS.
Finally, Nart shows off the new software Citizen’s Lab is developing (with a grant from my team at OSI) – Psiphon. Psiphon is a filtering circumvention tool that relies on you having a trusted contact in a nation where the net is not filtered. The trusted contact runs an easy install program and ends up with a window that looks much like an instant messaging program. Psiphon uses instant messaging protocols to tell friends in other nations what URL to access to use the proxy server you’ve just brought up, allowing them to surf the web through your connection. Very cool – can’t wait for a release of the software.
Just back from the opening event of the Global Voices conference – an informal dinner at a Lebanese restaurant near Marble Arch. (Thanks, Neha, for setting things up for us.) Amazing to be at a table with 40 bloggers from around the world, with conversations about the similarities between Malay, Chinese and Indonesian languages, the different forms of Islam in southeast Asia and the Middle East, what Jordanians think of Hungarians, wind power in China and Mongolian sumo. Can’t wait to see what the real meeting tomorrow brings.