ETD

• autonomic computing
• self-adaptiveness
• spark

The steadily increasing demand for resources of current-day data analytics applications needs to be counterbalanced by performance optimizations. This work aims to improve Spark-based applications with self-adaptive capabilities. For this, two components were designed and developed: one monitor and one actuator. These components will be part of a real-time automated control system that monitors and adjusts performance. The components interact with the TMA framework, a framework developed by researchers of the University of Coimbra. The BULMA entity matching service, a city traffic management application, will be used to demonstrate the system.