On Expressiveness and Efficiency in Content-based Publish/Subscribe

The Faculty of Informatics is pleased to announce a seminar given by Patrick Eugster

DATE: Tuesday, July 10th, 2012
PLACE: USI Università della Svizzera italiana, room SI-008, Informatics building (Via G. Buffi 13)
TIME: 14.00

ABSTRACT:
Publish/subscribe is a generic communication model in which processes can act as data producers by publishing data and/or can act as consumers by subscribing to data of interest. This model underlies applications such as Twitter as well as many industrial distributed information systems, and has served as inspiration for the design of future networks. Content-based publish/subscribe (CPS), where consumers subscribe to data based on arbitrary attributes of that data, is an expressive variant of publish/subscribe with application to location-aware services for mobile clients or financial applications.
In this talk, we present advances on expressiveness and efficiency in CPS. In particular, we investigate the concept of parametric subscriptions - subscriptions with criteria that may change over time. We discuss feasible and desirable semantics for such subscriptions in asynchronous systems, and present the extension of a common decentralized routing algorithm for CPS with corresponding support. We evaluate two implementations of this support in two application scenarios, illustrating its benefits. Then we discuss practical extensions of our support and overview ongoing work on content-based matching algorithms motivated by our insights gathered. During the talk, we also highlight relationships between our work on CPS middleware and our work on distributed programming languages.

BIO:
Patrick Eugster is an assistant professor in the Department of Computer Science at Purdue University. He is interested in distributed systems and programming languages, and especially programming models and techniques for distributed applications. He holds M.S. and Ph.D. degrees from the Swiss Federal Institute of Technology in Lausanne (EPFL). Patrick has co-authored over 80 articles and has been awarded for his research by an NSF CAREER award (2007) and by a fellowship from the Alexander von Humboldt Foundation (2011). He is also a member of the 2011 DARPA Computer Science Study Group.

HOST: Prof. Fernando Pedone