|
In recent years, applications reacting on or predicting a happening
of interest have received increasing attention. One example are
business applications such as supply chain management,
manufacturing, or e-business where information about the condition
and location of business events are of central interest. So-called
CEP (complex event processing) systems emerged in order to detect
situations resulting from events and to notify business processes
about them. These systems are mostly deployed in a centralized
fashion receiving events from multiple sources. Especially the
recent advance in sensor technology that now enables transmission of
condition and location information about goods, persons, and
materials in real-time makes the number of event sources grow
rapidly making a central processing more challenging. The
cooperation with other business partners extends an event-processing
system to multiple domains, each of it having its own
characteristics like used CEP engine or business rules. In this
context, a heterogeneous correlation technology is necessary in
order to ensure interoperability as well as efficient utilization of
network capability. This diploma thesis is part of the Distributed
Heterogeneous Event Processing (DHEP) project involving IBM
Böblingen lab and the University of Stuttgart. The objective of the
project is to develop a novel generic framework that allows for
reuse of existing possible heterogeneous correlation technology
distributed over multiple nodes. As part of this project, this
thesis addresses three key issues: make use of a novel correlation
language that enables interoperability, use a couple of different
communication mechanisms in order to support different requirements,
and address traffic issues with an intelligent routing mechanism
based on correlation rules.
|