Research Topic - Distributed Systems for Standing Subscriptions

This project started in January 2006, and is funded by the URC. The goal of this project is to study efficient techniques to disseminate data in a distributed publish/subscribe environment.

To date, we have achieved the following:

a) We have designed a publish/subscribe system to facilitate dissemination of streaming data where the subscription has a coherency constraint. Our proposed method is also adaptive to the runtime changes in the environment, and can adapt the dissemination tree dynamically at runtime to optimize performance.

b) We have also developed a novel architecture for scaleable distributed stream processing. Our system provides services to evaluate a large number of continuous queries for a large number of clients. The query layer is built over a distributed publish/subscribe architecture. We also proposed a solution to distribute queries in the query layer.

c) We have developed an efficient approach for filtering and disseminating fragmented XML documents in a publish/subscribe environment. The approach consists of novel query scheduling and optimization algorithms that exploit the semantics of fragmented data to minimize unnecessary and redundant processing overhead.

d) We have developed a new optimization technique that enables downstream routers to leverage the processing done by upstream routers to reduce response time.  This is achieved by dynamically piggybacking useful processing hints along with the XML data.

2) Some Publications

Y. Zhou, K. L. Tan, F. Yu. Leveraging Distributed Publish/Subscribe Systems for Scalable Stream Query Processing. VLDB 2006 Workshop on Business Intelligence for the Real-Time Enterprise, Seoul, Korea, 2006, 14 pages.

Y. Zhou, B. C. Ooi, K. L. Tan, F. Yu. Adaptive Reorganization of Coherency-Preserving Dissemination Tree for Streaming Data. International Conference on Data Engineering 2006 (ICDE'2006), Atlanta, USA, 2006, 10 pages.

C.Y. Chan, Y. Ni. Content-based Dissemination of Fragmented XML Data. International Conference on Distributed Computing Systems, Lisboa, Portugal, July 2006, 8 pages.

C.Y. Chan, Y. Ni, Piggyback Optimization of XML Data Dissemination. International Conference on Data Engineering, Istanbul, Turkey, April 2007, 2 pages (poster paper).

3) Names of the Faculty Members in the research area

Chee Yong Chan and Kian-Lee Tan