Founded in 1966

Workload-Aware Content Overlays

Evaggelia Pitoura, University of Ioannina, Greece

Wednesday, March 29
12:15pm - SENSQ 5317

Refreshments at 12:00pm

Hosted by Panos Chrysanthis

Abstract

We consider large dynamic data sharing networks, where nodes connect in an ad-hoc manner to exchange data. Since it is not possible for each node to know all other nodes in the network, each node connects to (knows) a small number of other nodes, thus creating overlay networks on top of the physical ones. An issue central in this context is routing a query towards the nodes that have data relevant to the query. To assist routing, each node maintains information about the content of its overlay neighbors in the form of routing indexes. In this talk, we shall present routing indexes for both path queries on structural data and for value-attribute queries. These indexes are also used in the construction of the overlay so that a node selects its overlay neighbors based on the similarity of their content as captured by their routing indexes. Finally, motivated by the fact that the content of each node alone does not reflect the actual contribution of the node to the data sharing network, we present workload overlays that take also into account statistics about the queries.

Biography of Speaker

Evaggelia Pitoura received her B.Sc. from the Department of Computer Science and Engineering of the University of Patras, Greece in 1990 and her M.Sc. and Ph.D. in computer science from Purdue University in 1993 and 1995, respectively. Since September 1995, she is on the faculty of the Department of Computer Science of the University of Ioannina, Greece, where is currently an Associate Professor. Her main research interests are in distributed data management. Her publications include several articles in international journals and conferences and a book on mobile computing. She received the best paper award in the IEEE ICDE 1999. She serves regularly on the program committees of several conferences on distributed systems and data management.

You are using an older browser that does not support current Web standards. Although this site is viewable in all browsers, it will look much better in a browser that supports Web standards.