This website provides the source-code and datasets used the experimental evaluation of TPStream. TPStream is an event processing operator for detecting complex temporal relationships among situatios. Situations describe a set of circumstances lasting for periods of time and are derived from raw event streams by predicate evaluation and aggregation. TPStream was published at the 21st International Conference on Extending Database Technology (EDBT). The full research paper can be found here.
TPStream is entirely written in the JAVA programming language and embedded into JEPC a middleware for uniform event processing. The source-download contains two Maven projects which can be easily imported into modern IDEs. The module jepc/tpstream-evaluation contains the source code of our experiments. Inside the scripts subfolder, there are ready-to-use scripts for re-running all experiments of the original paper. However, before running the scripts, run mvn clean install for both projects and make sure to adjust the M2_REPO variable inside classpath.sh to point to your local Maven repository.
We extended TPStream with a parallel implementation. This includes processing on a single machine using multiple threads as well as in a distributed environment backed by Apache Kafka. The extensions were published in the Journal for Distributed and Parallel Databases on ??-??-????. The full article can be found here.
The source-code for the extended version can be downloaded with the following links. To re-run the experiments, please also download the Linear Road data set from above.