Efficient Task Allocation in Data Center Environments Based on Information Theory

This paper introduces a task allocation technique to optimize resource allocation in data center environments with multiple hosts. Our approach, grounded in Information Theory principles, aims to enhance the efficiency and performance of task allocation within a single data center. Efficient task allocation is crucial for maximizing resource utilization and minimizing processing time. We can capture causal relationships (Complexity and Entropy) patterns by leveraging Information Theory measures. Our method intelligently allocates tasks to hosts by considering their capabilities and current workloads. Through extensive simulations and performance evaluations, we demonstrate the effectiveness of our approach compared to traditional task allocation techniques. We assess our method across three scenarios: low-load, high-load, and real-world. Our comparison encompasses First Fit, Round Robin, Random, and Ranked Heuristic techniques, analyzing host utilization, queue management, and allocation efficiency. Despite utilizing 10% more servers than the baseline, our solution outperforms others, demonstrating a peak improvement of 48%. The results underscore the advantages of employing information theory measures for data center task allocation, allowing more adaptive and dynamic resource management strategies.