Abstract
Hadoop is an open-source MapReduce implementation widely adopted in industry and academia. However, achieving effective trade-offs between the cost and execution time of Hadoop applications is challenging due to the numerous Hadoop parameters that need to be configured properly. Running Hadoop applications on public clouds compounds this challenge because cloud parameters such as the number and types of virtual machines used to run the application also need to be decided. We will develop an approach for the multi-objective optimisation of Hadoop applications running on public clouds, enabling users to make the best of Hadoop processing when deployed in a cloud environment. The approach will provide those responsible for the configuration of a Hadoop application with a set of Pareto-optimal configurations, allowing them to run the application optimally within their time and/ or budgetary constraints.