Secure routing in peer-to-peer networks is challenging primarily because adversaries can easily participate in critical positions in the network. Assuming that the network is fully decentralized and thus no participant can be trusted, detecting and pinpointing malicious behavior is hard. We are interested in scalable protocols that preserve user privacy and provide performance guarantees (such as data integrity, limited bandwidth consumption and latency).
This research involves the creation of new consistency models and protocols that can be used efficiently given the constraints of a best-effort network and malicious peers. As part of our work, we are developing secure and cheat-proof protocols for peer-to-peer games. Games are particularly challenging because players have an incentive to cheat while protocols must meet the real-time message passing requirements dictated by the type of game itself. We are currently using traffic traces and simulations to validate our protocols.
Building a reputation system for a decentralized network without a trusted third party is a challenging problem. A trusted third party adds additional cost and security risk. Without one, collusion is a problem because reputation systems are based on feedback from peers, e.g. malicious peers can collude to give each other good ratings while giving negative ratings to other peers. This project deals with building a practical reputation system under some reasonable assumptions, e.g. the malicious nodes are in a minority.
Managing the high volume of alarms generated by large intrusion detection environments can be very challenging. A major problem faced by those who deploy current intrusion detection technology is the large number of false alarms generated by Intrusion Detection Sensors (IDSs), which can be well over 90 percent. In this project, techniques from data mining and web search (such as PageRank) are applied to separate the signal from noise.
The Domain Name System (DNS) is the hierarchical naming system used to locate hosts on the Internet. The most important task of a name server is to translate the human meaningful domain names to IP addresses. A DNS “Start of Authority” (SOA) is the DNS server responsible for a particular network. In this project, we are interested in the number of SOAs on the Internet and the paths messages take to the SOAs from various points in the network. Knowing these paths we can decide which routers are critical for the availability of large portions of the Internet and thus deserve special protection or additional redundancy.