Energy-Aware Query Scheduling in Cloud-Based Relational Database Systems
Abstract
Energy consumption in cloud-based relational database management systems (RDBMS) has emerged as a critical challenge facing modern data centers, with operational energy costs accounting for up to 40% of total infrastructure expenditure. Traditional query scheduling strategies, including First-Come-First-Served (FCFS) and Shortest-Job-First (SJF), prioritize performance optimization while neglecting explicit energy consumption considerations. This study proposes an Energy-Driven Adaptive Scheduling (EDAS) algorithm that dynamically prioritizes queries based on estimated CPU utilization, disk I/O costs, and historical energy profiles without requiring modifications to the underlying database engine. Experimental evaluation was conducted on a cloud-based MySQL 8.0 system deployed on Amazon Web Services (AWS) EC2 instances using light (50 queries), medium (150 queries), and heavy (300 queries) workloads derived from Sakila and TPC-H benchmarks. Results demonstrate that energy-aware scheduling exhibits workload-dependent performance characteristics: SJF achieves optimal energy efficiency under light and medium workloads with 14.2% and 16.0% savings respectively, while EDAS achieves measurable energy savings of 10.4%, 12.3%, and 20.5% under heavy workloads compared to FCFS, SJF, and baseline scheduling approaches. EDAS demonstrates greater resilience under CPU throttling conditions, maintaining 15.8% energy reduction when processor frequency drops from 2.5 GHz to 1.8 GHz. The energy-delay product (EDP) improves by 18.7% under heavy workloads, indicating superior energy-performance trade-offs. This study establishes the importance of workload-aware query scheduling for improving cloud database energy efficiency and provides practical guidelines for implementing energy-conscious scheduling in production RDBMS environments. The proposed approach reduces operational costs by approximately $1,247 annually per mid-sized database instance while maintaining ACID compliance and query performance guarantees. Keywords: energy-aware computing, query scheduling, cloud database, green computing, energy efficiency, adaptive scheduling, TPC-H benchmark, energy-delay product