My research spans the areas of security and privacy, overlay and P2P
networks and in general
large-scale networks and distributed systems.
"The
crucial
variable in the process of turning knowledge into value is
creativity"
–
John
Kao
Research interests
-
Security and Privacy: Protection of personal privacy in today's Internet needs to extend beyond email and web browsing to contemporary applications such as Voice-over IP,
streaming media and web-services. Anonymous systems seem to be the right answer in protecting users' identity over Internet. However, those systems have to face several issues, among which tolerate nodes join and leaving these systems and improve their quality-of-service, that is usually very poor.
- Strongly Dynamic Anonymous Systems. Deployed infrastructure-based solutions, such as Tor, offer static end-to-end anonymous paths, which need to be rebuilt any time a router on the path leaves the system. The dynamic nature of peer-to-peer networks poses significant challenges to anonymous communication. Nodes can enter or leave the network at any time. This churn breaks existing anonymous paths and forces them to be rebuilt by communication endpoints. Each path rebuild not only disrupts communication and consumes resources, but also exposes the flow to passive logging attacks. Selection of routers in the anonymous path construction is either performed randomly, or relies on self-described resource availability from each router, which makes the system vulnerable to low-resource attacks.
- Low latency Anonymous Systems:
Selection
of routers in the anonymous path construction is a critical operation
in order to protect the endpoints of the anonymous communication from
passive logging attacks. However, traditional anonymous systems perform
this operation in a secure manner, via random router selection
strategies, they sensibly effect the path performances. The design a
path construction algorithm for anonymous routing networks that
provides both performance and anonymity guarantees, requires a deep
knowledge of all the factors that play a role in optimizing both
performance and anonymity.
-
Social network: Access to realistic, complex graph datasets is critical to research in a variety of network domains, ranging from community detection in the study of social networks, to searches for efficient routing paths in communication networks. Social networks, such as Facebook, Flicker and MySpace to cite a few, are the right candidate to address those research exigences. However, it is difficult to obtain complex graph datasets, therefore a synthetic graphs generated by precise social network models are the practical solution to achieve those goals. We are exploring different methodologies to compare social network models and we are identifying a practical and efficient way to reuse complex and simple models to accurately reproduce social networks.
