Abstract:
Complex Event Processing (CEP) is one of the emerging areas in computer science and it is being heavily used
in real time systems during the recent times. CEP systems should
be scalable and fault tolerant. The research project epZilla is to
build a scalable, highly available and fault-tolerant distributed
architecture for CEP systems. Software Transactional Memory,
Leader Election algorithms, dynamic load balancing, dynamic
service discovery,
Stratification are the concepts which are used to build the
proposed architecture. The results of the pilot run are promising
and show that the proposed architecture is scalable and fault
tolerant.