Energy-aware Load Balancing in Heterogeneous Environments
Usman Iqbal Ahmed, Jerry Chun‐Wei Lin, Gautam Srivastava · The 5th International Conference on Future Networks & Distributed Systems · 2021
In this paper, we propose an energy saving strategy for heterogeneous clusters. A load balancer is then proposed by using code features and a resource-aware processor selection based on machine learning. The proposed technique identifies energy efficient kernel applicants (from the job pool). It then selects a pair of kernel candidates among all candidates that leads to a reduction in energy consumption and execution time. The proposed kernel technique has achieved good performance compared to the past SOTA models.