Bernard

Bernard Wong

Assistant Professor

Davis Center Room 3514
David R. Cheriton School of Computer Science
University of Waterloo
200 University Ave. West
Waterloo, ON N2L 3G1, Canada

E-mail: bernard AT uwaterloo DOT ca
Phone: (519) 888-4567 x31301

 

My research interests span distributed systems and networking, with particular emphasis on problems involving decentralized services, self-organizing networks, and distributed storage systems.

Prospective students: I am currently seeking motivated graduate students who enjoy building large systems. More information about the School's graduate program requirements can be found here.

Teaching

Winter 2012 CS 454/654 Distributed Systems

Fall 2011 CS 854 Advanced Topics in Computer Systems: Cloud Computing & Management

This is a graduate course on current research topics in cloud computing and management. It is offered as a "paper-chase" course: we read and discuss important papers on different subtopics in cloud computing and management each week. Students must also individually complete a final project that explores a new idea in one of the sub-topics covered in the course. The goal is that by the end of the course, the final project can serve as a starting point for a workshop paper submission.

Projects

Microfuge Performance Isolation in Cloud-based Storage Systems.

Cloud computing has enabled significant improvements to datacenter utilization through the consolidation of services from different businesses onto the same machines. The tradeoff to improved utilization, however, is reduced isolation between services, which leads to inconsistent and unpredictable performance. The Microfuge projects aims to tackle the problem of providing performance isolation in the underlying, shared storage systems that often dictates the performance characteristics of cloud-based services.

HyperDex A Distributed, Searchable Key-Value Store for Cloud Computing.

HyperDex is a distributed key-value store that provides high performance and strong consistency for get/put operations using the primary object attribute (key), and provides for efficient search operations on any combination of secondary object attributes. This is possible through the use of Hyperspace Hashing, which maps objects with multiple attributes into points in a multidimensional hyperspace. Search queries on secondary object attributes can therefore be mapped to small, hyperspace regions representing the set of feasible locations for the matching objects.

Cubit Approximate Matching for Peer-to-Peer Overlays.
Octant A Comprehensive Framework for Geolocalization on the Internet.
Meridian A Lightweight Approach to Network Positioning.
ClosestNode.com A DNS Redirection Service for Closest Node Selection.

Publications

Patents

Professional Activities

Last modified: May 15, 2012