Petrifying Worm Cultures: Scalable Detection and Immunization in Untrusted Environments

We present and evaluate the design of a new and comprehensive solution for automated worm detection and immunization. The system engages a peer-to-peer network of untrusted machines on the Internet to detect new worms and facilitate rapid preventative response. We evaluate the efficacy and scalability of the proposed system through large-scale simulations and assessments of a functional real-world prototype. We find that the system enjoys scalability in terms of network coverage, fault- tolerance, security, and maintainability. It proves effective against new worms, and supports collaboration among among mutually mistrusting parties.