"Illuminating the shadows of the Internet" Opportunistic network and web measurement Home •• Project home page Overview •• Brief overview and news FAQ •• Frequently-asked questions Pubs •• Publications and people Stats •• Preliminary results Maps Proxies •• Data coverage and proxy maps Contribute •• Join our community Frequently-asked questions Can I help? Yes, we'd love your help! Please see our contribute page for instructions on how you can assist in our measurement efforts. If I add such a link, are all requests measured by your servers? Yes, unless you employ some more advanced server-side tools to only probabilistically include such a link, e.g., via php scripts. How does this work? Please see our overview page for detailed information on how we perform our opportunistic measurements. What is your current view of the Web? Thanks to integration into CoralCDN web proxies, we receive data from 10,000s of vantages per day. Check back in the near future for near-real-time information as to usage statistics and locality information about from where requests originate. Are all CoralCDN requests measured? No: we are rather conservative as to when requests to CoralCDN result in redirects back to our measurement servers. We try our best to ensure that such redirects can be safely handled by browsers, for example, they are not part of <img src or <embed src requests and do not correspond to RSS or other XML feeds. Measurements are done probabilistically, on average one in every couple thousand requests is used for measurement. All measured IP addresses are cached for an hour to ensure they aren't measured more than once within that time period. What information do you log? All requests to our measurement servers are logged, which includes a client's IP address and port number (both local and public), its DNS resolver's IP address, and full HTTP headers including referer, via, and user-agent. What is your privacy policy? While we use logs to measure Internet topology, network environments, network deployments, and other peculiarities, any published information will either be in aggregate or anonymized form. We will not release personally-identifying information to any third-parties under normal operating conditions. Your team numbers don't seem to add up to the total numbers? Information gathered from CoralCDN, which accounts for the majority of our measurement traffic, is not counted amongst the team data. That just wouldn't be fair to the other teams! The team CoralCDN only corresponds to traffic originating from the website www.coralcdn.org, not from all measurements gathered through the actual content distribution network (i.e., any third-party website loaded via something.nyud.net). What's with the dog? Roo is the unofficial mascot of the illuminati project (at least until we find a more suitable one). She's a little psychotic and has to wear a muzzle when on walks. Oh...and she yodels when she's happy (kind of sounds like 'A-rooo-roooo-roo'): hence the name. Beware of the roo.