Optimal Power Cost Management Using Stored Energy in Data Centers (1103.3099v2)

Abstract: Since the electricity bill of a data center constitutes a significant portion of its overall operational costs, reducing this has become important. We investigate cost reduction opportunities that arise by the use of uninterrupted power supply (UPS) units as energy storage devices. This represents a deviation from the usual use of these devices as mere transitional fail-over mechanisms between utility and captive sources such as diesel generators. We consider the problem of opportunistically using these devices to reduce the time average electric utility bill in a data center. Using the technique of Lyapunov optimization, we develop an online control algorithm that can optimally exploit these devices to minimize the time average cost. This algorithm operates without any knowledge of the statistics of the workload or electricity cost processes, making it attractive in the presence of workload and pricing uncertainties. An interesting feature of our algorithm is that its deviation from optimality reduces as the storage capacity is increased. Our work opens up a new area in data center power management.

• The paper presents a novel algorithm that uses stored energy from UPS units to dynamically lower electricity costs in data centers.
• The paper applies Lyapunov optimization for adaptive energy management without requiring prior statistical data on workloads or pricing.
• The paper demonstrates that increased storage capacity significantly narrows the cost gap, transforming UPS systems into strategic economic assets.

Optimal Power Cost Management Using Stored Energy in Data Centers

The paper, "Optimal Power Cost Management Using Stored Energy in Data Centers," explores strategies for leveraging energy storage to reduce electricity costs in data centers, which consider electric utility bills as a substantial operational expense. This research suggests a paradigm shift in using Uninterruptible Power Supply (UPS) units traditionally viewed as failover systems to potential economic assets for cost optimization.

Core Contributions

The authors introduce a method to exploit stored energy utilizing UPS units, which systematically deviates from customary use. Central to this research is the application of Lyapunov optimization to devise an online control algorithm aiming to minimize the time-average electricity bill. The proposed solution operates efficiently without prior statistical knowledge of workload or energy pricing variabilities.

Key features of the algorithm include:

• Storage-Based Optimization: The algorithm leverages energy storage devices to manage electrical loads more effectively, particularly by charging when prices are low and discharging during peak tariffs.
• Adaptive Control: Avoiding traditional dynamic programming, the solution offers a prominent advantage by tackling the "curse of dimensionality" through simplified adaptive controls.
• Scalability with Storage: The methodology exhibits improved performance, nearing optimality as storage capacity increases, mitigating deficiencies present in dynamic programming approaches.

The paper further identifies the merit in the storage capacity to refine the algorithm's efficacy, presenting an approach that is intuitively simpler and dynamically more practical.

Numerical Results and Claims

The results underlined substantial numerical gains—showing improved cost efficiency as energy management via UPS units becomes more prevalent. Notably, the gap between the devised algorithm's results and an optimal cost scenario diminishes with augmented storage capacity, indicating strong performance metrics.

Implications and Future Directions

Practically, adopting such a strategy can lead to significant cost reductions for data centers, which can dedicate vast storage capacities already installed primarily for reliability, to also contribute to cost management. Theoretically, this work opens new avenues in power management within data centers, emphasizing the importance and potential of stored energy in real-time decision processes.

Going forward, questions concerning the integration with various utility tariffs and participation in demand response programs could be further contemplated. Additionally, expansion into mixed-energy environments involving multiple utilities or renewable sources offers promising research prospects.

In conclusion, this paper presents a novel approach to energy cost management, utilizing stored energy smartly within current data center infrastructures and potentially shifting operational perspectives to view energy storage not merely as a backup but as a strategic asset for cost efficiency.