Extra Seminar Computer Science - Dr. S. Schulte

Title: Elastic Data Stream Processing for the Internet of Things

Abstract:
In volatile Internet of Things (IoT) scenarios, the data volume to be processed and stored changes permanently. Hence, there is a necessity to adapt computational resources to be able to handle varying data volume. Despite this, existing data processing engines only provide limited support regarding runtime reconfigurations, distributed deployments, and the utilization of elastic computational resources.

Within this talk, elastic computing will be discussed as a solution to satisfy volatile demands for computational resources in the IoT. Rapid elasticity is one of the basic capabilities of cloud computing, and allows to flexibly adapt computational resources to dynamic changes in the environment, requirements, and data volume. A special focus will be on elastic stream processing for the IoT. We will present an approach for the realization, configuration, and optimization of stream processing topologies using containerized software components. The goal of the proposed optimization is cost-efficiency under changing data volumes, while respecting given Service Level Agreements.