Distributed heterogeneous event processing

Abstract

Although a significant amount of research has investigated the benefits of distributed CEP in terms of scalability and extensibility, there is an ongoing reluctance in deploying distributed CEP in an industrial context. In this paper we present the DHEP system developed together with the IBM R 1 laboratory in Böblingen. It addresses some of the key problems in increasing the acceptance of distributed CEP, for example supporting interoperability between heterogeneous event processing systems. We present the concepts behind the DHEP system and show how those concepts help to achieve scalable and extensible event processing in an industrial context. Moreover, we verify in an evaluation study that the additional cost imposed by the DHEP system is moderate and 'affordable' for the benefits provided.

VISIRI - Distributed Complex Event Processing System for Handling Large Number of Queries

“…In order to calculate the cost of a given query, we have developed a cost model that gives a numeric value for a query based on the empirical studies done by Schilling et al [6] and Marcelo et al [5].…”

“…Furthermore, a cost value is assigned to the number of attributes in the input stream definitions and the output stream definitions. This is because the literature [6] suggests that when the number of attributes in the event streams for a particular query increases the resource requirement for that query increases. Apart from that, the number of input streams and the output streams count is also added to the cost value.…”

“…Load on a node is calculated by considering the cost of the CEP queries. The cost of queries is calculated by the cost model, which is based on the empirical studies done by Mendes et al [5] and Schilling et al [6]. With this query distribution algorithm, VISIRI balances the resource utilization of the CEP nodes, so that no node gets overloaded (given that there is no sudden change in the incoming event streams) and adversely affects the throughput of the overall system.…”

“…Schilling et al [6] have studied about the cost of executing different query operators in their empirical study. According to their study, filter rules have the lowest latency and logical rules have much higher latency.…”

An Approach for Dynamic Personal Monitoring based on Mobile Complex Event Processing

“…The DHEP [9] project develops an industry system of distributed CEP, which can support interoperability between heterogeneous event processing systems. In [10] the authors develop a platform called DiCEPE (Distributed Complex Event Processing) that focuses on the integration of CEP engines in distributed systems, and which is capable of using various communication protocols.…”