Timewheel

The Timewheel group communication service consists of a membership and a reliable broadcast protocol that supports the construction of high performance fault-tolerant distributed systems. This service supports several different semantics simultaneously and provides good overall performance. The broadcast protocol supports three atomicity semantics---{\it weak}, {\it strong}, and {\it strict atomicity}---and three ordering semantics---{\it unordered}, {\it total order}, and {\it time order}. These result in nine group communication semantics and an application can dynamically choose the desired atomicity and order semantics when broadcasting an update. These protocols provide fast delivery and stability times, use a small number of messages to accomplish a broadcast or a membership change, distribute evenly the processing load among group members, and use efficient flow control techniques. The good overall performance is maintained in the absence of any failures and normal update arrival rates, in the presence of communication or process failures, and when updates arrive at a very fast or very slow rate.

Publications

S. Mishra, C. Fetzer and F. Cristian, The Timewheel Asynchronous Atomic Broadcast Protocol. Proceedings of the 1997 International Conference on Parallel and Distributed Processing Techniques and Applications, Las Vegas, NV (June 1997), 1239--1248. Abstract
S. Mishra and R. Yang, Thread-based vs Event-based Implementation of a Group Communication Service. Proceedings of the 12th International Parallel Processing Symposium & 9th Symposium on Parallel and Distributed Processing, Orlando, FL (April 1998). Abstract
S. Mishra, C. Fetzer, and F. Cristian, The Timewheel Group Membership Protocol. Proceedings of the 3rd IEEE Workshop on Fault-tolerant Parallel and Distributed Systems, Orlando, FL (April 1998). Abstract
S. Mishra, Constructing Applications Using the Timewheel Group Communication Service. Proceedings of the 1998 International Conference on Parallel and Distributed Processing Techniques and Applications, Las Vegas, NV (July 1998), 1591--1598. Abstract
S. Mishra and G. Pang, Design and Implementation of an Availability Management Service. Proceedings of the ICDCS Workshop on Middleware, Austin, TX (June 1999). Abstract
S. Mishra and R. Yang, Experience with Handling Concurrent Events in Group Communication Services. International Journal of Parallel and Distributed Systems & Networks, Volume 4, Number 2, 2001. Abstract
S. Mishra and G. Pang. Teams: An Availability Management Service for a Timed Asynchronous Distributed System. International Journal of Computers and Their Applications, 9(1), March 2002.
S. Mishra, C. Fetzer and F. Cristian. The Timewheel Group Communication System. IEEE Transactions on Computers, 51(8), August 2002.