Epidemiological Protocols

Meng-Jang Lin Keith Marzullo Aleta Ricciardi

Goals


Epidemiology is the study of how a disease spreads through a population. An epidemiological protocol is designed with the paradigm of a disease spreading through a population. Epidemiological protocols are attractive because they are relatively simple yet are robust againt common failures such as message loss and processor crashes. They also seem to scale very well with the size of the network. Examples of epidemiological protocols include gossip protocols that are used to disseminate information in a wide-area network and viruses and worms that are used to covertly disseminate programs throughout a network.

An epidemic can be characterized by many different parameters, but our research focuses on one: how an individual can contact another thereby transmitting the disease (in our case, information). We are studying how this parameter affects the distribution of the number of infected individuals as a function of time. For the case of viruses and worms, we are interested in slowing the infection down, while for gossip protocols we are interested in speeding it up.


Papers

Meng-Jang Lin, Keith Marzullo, and Aleta Ricciardi. A New Model for Availability in the Face of Self-Propagating Attacks. In Proceedings of the 1998 New Security Paradigms Workshop,Charlottesville, VA, November 1998, pp. 134-137. Full version available as UCSD Technical Report TR CS99-0610, January 1999.

Meng-Jang Lin and Keith Marzullo. Directional Gossip: Gossip in a Wide-Area Network. UCSD Technical Report TR CS99-0622. A shorter version can be found in the Proceedings of the Third European Dependable Computing Conference, Prague, Czech Republic, September 1999 (Springer-Verlag Lecture Notes in Computer Science 1667), pp. 364-379.

Meng-Jang Lin, Keith Marzullo, and Stefano Masini. Gossip versus Deterministic Flooding: Low Message Overhead and High Reliability for Broadcasting on Small Networks. UCSD Technical Report TR CS99-0637.

 
 
 
 
 

Last updated by Keith Marzullo on 18 November 1999