Fuel cells are a promising power source for future data centers, offering high energy efficiency, low greenhouse gas emissions, and high reliability. However, due to mechanical limitations related to fuel delivery, fuel cells are slow to adjust to sudden increases in data center power demands, which can result in temporary power shortfalls. The problem of dynamic power requirements was addressed in a paper titled, "SizeCap: Coordinating Energy Storage Sizing and Power Capping for Fuel Cell Powered Data Centers" presented by Yang Li, Di Wang, Saugata Ghose, Jie Liu, Sriram Govindan, Sean James, Eric Peterson, John Siegler, Rachata Ausavarungnirun, and Onur Mutlu, with Carnegie Mellon University and Microsoft Research, during the 22nd IEEE Symposium on High-Performance Computer Architecture.
The paper proposes an energy storage device (ESD) sizing framework for fuel cell powered data centers which coordinates ESD sizing with power capping to allow data centers to employ a cost-effective solution to power shortfalls. According to the paper, it costs about $20,000 to increase capacity of the supercapacitors used to store energy by 1kWh. “With an ESD-only solution that uses worst-case sizing, the total cost of ownership (TCO) increases significantly to cover these infrequent occurrences.â€
To mitigate the impact of power shortfalls, prior work has proposed to either perform power capping by throttling the servers, or by leveraging ESDs that can temporarily provide enough power to make up for the shortfall while the fuel cells ramp up power generation. Both approaches have disadvantages: power capping conservatively limits server performance and can lead to service level agreement (SLA) violations, while ESD-only solutions must significantly overprovision energy storage capacity to tolerate the shortfalls caused by worst-case (i.e., largest) power surges, which greatly increases the TCO.
The paper proposes SizeCap, the first ESD sizing framework for fuel cell powered data centers, which coordinates ESD sizing with power capping to allow data centers to employ a cost-effective solution to power shortfalls. SizeCap sizes the ESD just large enough to cover the majority of power surges, but not the worst-case surges that occur infrequently, to greatly reduce TCO. It then uses the smaller capacity ESD in conjunction with power capping to cover the power shortfalls caused by worst-case power surges.
The proposal is to use medium-capacity ESDs to absorb the small spikes, together with some small ESDs to help absorb the big spikes. During those big spikes, the small ESDs get some help from power capping, with the framework helping to determine the degree of power capping that allows the data center to continue meeting its Service Level Agreements. SizeCap is a framework for determining what the most optimal sizes of those ESDs are and what the most optimal power capping policy should be.
As part of this new flexible framework, multiple power capping policies are proposed with different degrees of awareness of fuel cell and workload behavior, and evaluate their impact on workload performance and ESD size. Using traces from production data center systems, it was demonstrated that SizeCap significantly reduces the ESD size (by 85% for a workload with infrequent yet large power surges, and by 50% for a workload with frequent power surges) without violating any SLAs.
