Tuesday, July 27, 2010

Marking the successful Completion of Project epZilla

The project epZilla was initiated with the aims to create a scalable fault tolerant architecture for complex event processing systems. At the end of the allocated time period, we have looked back at the Distributed System Architecture which we came up after a successful research effort on both Distributed Systems and Complex Event Processing domains.

The weights of our research are nearly 70% on Distributed Systems domian and 30% on Complex Event Processing domain. As a result, we were able to design a sophisticated Distributed Systems Architecture for Complex Event Processing Systems. As this system is sophisticated for Complex Event Processing Systems, it poses some of the Complex Event Proceesing system features such as Stratification. This idea is introduced by several IBM researchers recently and therefore it is fairly new idea for Complex Evet Processing domain as well. Therefore we believe project epZilla might be the first practical implementation of the Stratification.

Since project epZilla is a research project, we implemented a prototype implementation of the Distributed System Architecture which we designed to validate the design and the performance measures. Therefore we have implemented all the components introduced in the design with several value added features such as dynamic load balancing and integrated them and drove lot of test runs with several physical nodes and measured the performance of the system. These performance information confirmed the correctness of our design and achievement of our design motivations such as high throughput while preserving Fault Tolerance and Scalability of the system. Therefore we have successfully demonstrated our research effort, Project epZilla to our Academic Staff of the Department of Computer Science & Engineering, University of Moratuwa, Sri Lanka using the prototype implementation of the design and explained why such design decisions were made and how those designs were practically implemented.

At the end, we have designed and developed a Scalable Fault Tolerant Distributed Architecture for Complex Event Processing Systems. We are proud and delighted to annouce that we are concluding this research project with a great success of 100% achievment of the project goals.

No comments:

Post a Comment