An advanced reinforcement learning approach for energy-aware virtual machine consolidation in cloud data centers

摘要

Energy awareness presents an immense challenge for cloud computing infrastructure and the development of next generation data centers. Inefficient resource utilization is one of the greatest causes of energy consumption in data center operations. To address this problem we introduce an Advanced Reinforcement Learning Consolidation Agent (ARLCA) capable of optimizing the distribution of virtual machines across the data center for improved resource management. Determining efficient policies in dynamic environments can be a difficult task, however the proposed Reinforcement Learning (RL) approach learns optimal behaviour in the absence of complete knowledge due to its innate ability to reason under uncertainty. Using real workload data we evaluate our algorithm against a state-of-the-art heuristic, our model shows a significant improvement in energy consumption while also reducing the number of service violations.