Wednesday, October 28, Volen 101, 2:00-3.00 pm
This talk will basically be an English version of my Ph.D defense. I will introduce rapidly how cooperation can be studied as a consistency problem. I will shortly describe why the consistency management should be brought by the ``system''. I will then develop the few principles that address the two critical issues for the consistency management in order to be a ``system'' mechanism : it must be developed independently of the application, so it can ``exist'' in the system. It must be flexible enough to address the correctness and the performance requirements of cooperative applications. I will describe the architecture I have implemented and some experiments done with it in order to validate these principles.
Host: Liuba Shrira