RESEARCH THEME 02 / DOCTORAL WORK
Moving from stored batches to processing events as they arrive, with a stream model for cloud execution.
Monitoring and alerting depend on a continuous supply of new information. Curracurrong Cloud applies stream processing to cloud-hosted clusters, allowing computation to operate alongside the systems generating the data.
The work describes processing pipelines through an algebraic-style representation and supports automatic placement across compute resources. It builds on the Curracurrong approach developed for wireless sensor networks.
IEEE 30th International Conference on Data Engineering Workshops (ICDEW), CloudDB 2014, pp. 207–214
A system for continuous data processing within cloud-hosted clusters. It describes processing pipelines algebraically and automatically places computation across available resources, allowing data sources and processing to share a cluster.