In a recent technical report, authored by Ray Sweha, Azer Bestavros, and Ibrahim Matta, we have presented a novel approach to boost the performance of the popular Tor network (which uses onion routing) through the use of on-demand cloud resources. This work is currently under review and follow-up work is underway! [Read the technical report]
Results from our work on monetizing cloud workload flexibility (the CloudPack project) will be presented at the upcoming ACM/IFIP/USENIX Conference, to be held in Montreal, Canada, in December 2012. In this work we present a mechanism that provides incentives for cloud customers to express their flexibility, and argue that efficient cloud markets emerge when "elasticity" is exploited not only on the supply side of cloud markets, but also on the demand side. [Read the Paper]
The work by Bestavros and collaborators at the Eurecom Institute in France and at the University of Verona in Italy on the use of cloud resources for peer-assisted content distribution "on a budget", which was incorporated in the Cyclops system has been accepted for publication in COMNET. [Read the correponding technical report]
Professor Bestavros presented a number of recent results from the CloudCommons project in a colloquium talk entitled "Mechanism Designs for Efficient Cloud Marketplaces" in the Computer Science Department, at Worcester Polytechnic Institute, Worcester, MA, on January 13, 2012. The talk emphasized the most recent results on using Shapley Pricing as a mechanism to ensure efficient colocation of cloud workloads at scale.
Our work on "MorphoSys: Efficient Colocation of QoS-Constrained Workloads in the Cloud" was accepted for publication and presentation at CCGrid'12: The IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing, in Ottawa, Canada, May 2012. [Read the Paper]
Our work on "AngelCast: Cloud-based Peer-Assisted Live Streaming Using Optimized Multi-Tree Construction" has been accepted for publication and presentation at the ACM MMsys'12 Conference, in Chapel Hill, NC, February 2012 [Read the Paper].
Professor Bestavros presented a number of recent results from the CloudCommons project in a talk entitled: "In Clouds We Trust: Towards a Trustworthy Marketplace for Cloud Resources" at the DIMACS Workshop on Systems and Networking Advances in Cloud Computing, which was held at Rutgers University, in Piscataway, NJ, in December 2011.
Our work on "A Framework for the Evaluation and Management of Network Centrality" has been accepted for publication in SDM'12, in Anaheim, CA, April 2012. In this work we present a general framework for the efficient evaluation of arbitrary network centrality measures. [Read the Paper].
Professors Evimaria Terzi and Azer Bestavros have been selected to receive a Google Faculty Research Award in support of their research on “Generalized Centrality Measures with Applications to Information Networks.” This work seeks to develop new techniques for measuring the centrality of small groups within a larger network. Traditionally, different measures of centrality have been defined for individual nodes and links. Such measures enable applications to compare nodes and links to each other with respect to a particular interesting property of the network. Other applications may need to assess the centrality of groups within the network. For example, interesting questions might include ”What is the best group of nodes to carry out inspection, monitoring, sampling, aggregation, or de-duplication in a content flow network?” or ”What is the minimum group of nodes to intercept all possible communication between a specific set of sources and destinations?” To answer these and many similar questions, this project defines ways to express statements about group centrality for flow and access networks, and develop algorithmic techniques that can be implemented for use on large-scale networks.
The Google Research Awards program aims to identify and support world-class, full-time faculty pursuing research in areas of mutual interest. The intent of the awards program is to support academic research aimed at improving information access (defined broadly).
Our work on "The Filter-Placement Problem and its Application to Minimizing Information Multiplicity" has been accepted for publication in VLDB 2012, in Istanbul Turkey, August 2012. In this work we consider the problem of optimally placing de-duplication functionality in a distributed information (flow) network. [Read the Paper].
1-10 of 18