Lightweight security for Internet polls

Is it possible to implement practical Internet Polls that fulfill even the weakest security requirements? The technology available today would lead to a negative answer, because of the following practical constraints: standard, unmodified browsers are used, it is not economically possible to distribute certificates or even just user names and passwords, users connect from different workstations, possibly behind firewalls, proxies and address translation nodes. In this paper, we define an innovative notion of Internet Poll security, namely “Security against Massive Falsification”, and we present a method that we consider to be secure with respect to this definition. We discuss the security properties of the method with respect to existing techniques, and then propose a public challenge for testing the strength of our claim.