Leveraging stored energy for handling power emergencies in aggressively provisioned datacenters

Abstract

Datacenters spend $10-25 per watt in provisioning their power infrastructure, regardless of the watts actually consumed. Since peak power needs arise rarely, provisioning power infrastructure for them can be expensive. One can, thus, aggressively under-provision infrastructure assuming that simultaneous peak draw across all equipment will happen rarely. The resulting non-zero probability of emergency events where power needs exceed provisioned capacity, however small, mandates graceful reaction mechanisms to cap the power draw instead of leaving it to disruptive circuit breakers/fuses. Existing strategies for power capping use temporal knobs local to a server that throttle the rate of execution (using power modes), and/or spatial knobs that redirect/migrate excess load to regions of the datacenter with more power headroom. We show these mechanisms to have performance degrading ramifications, and propose an entirely orthogonal solution that leverages existing UPS batteries to temporarily augment the utility supply during emergencies. We build an experimental prototype to demonstrate such power capping on a cluster of 8 servers, each with an individual battery, and implement several online heuristics in the context of different datacenter workloads to evaluate their effectiveness in handling power emergencies. We show that: (i) our battery-based solution can handle emergencies of short duration on its own, (ii) supplement existing reaction mechanisms to enhance their efficacy for longer emergencies, and (iii) battery even provide feasible options when other knobs do not suffice.

Sustainable Cloud Computing

“…We devise an online solution based on T-slot Lyapunov optimization to optimize both the electricity cost, and the peak power cost without significantly violating the carbon capping requirement of the data centers. Finally, while the existing work on data center peak power optimization rely on predictability of cloud parameters [47,50,51,52,61,75,116,121], we show that the prediction error of cloud parameters such as workload has a very harmful impact on the peak power optimization, and adopt stochastic programming to remove such an impact. The following sections give a detailed overview on the contributions.…”

“…There are also some recent works which explored the use of ESDs to reduce both the energy cost and the peak power cost within [47,50,51,52,61,75,116,121] and across [43] data centers. The idea is to store energy in UPS batteries during "valley" periods of lower demand, which can be drained during "peak" periods of higher demand.…”

Multi-tier energy buffering management for IDCs with heterogeneous energy storage devices

“…Typically charging rate is much less than discharging rate, e.g., charging rate is 5-10 times less than discharging for lead-acid batteries. As suggested by the previous studies [4], [8], the existing UPSes in data centers can be utilized for energy management of the tier one. However, the energy level at the battery should always be sufficient to guarantee the desired availability, denoted by B min .…”

“…Authors in [6] perform a trace-based simulation using Akamai CDN workload traces to investigate the energy cost saving that can be achieved by using batteries to shave the peak power draw from the grid. Govindan et al, propose to leverage existing UPSes to temporarily augment the utility supply during emergencies (i.e., peak power) [8]. Finally, [7] presents an energy buffering management policies for distributed per-server UPSes to smoothen power draw from grid.…”

“…Current data centers mainly use ESDs as power backup [4], [7]. These ESDs can also be used to store excess renewable energy and low-cost electricity during workload valleys, which can be used to meet high power demands [4]- [8]. However, energy buffering management should be aware of: (a) the intermittent nature of available energy from renewables, (b) variability in workload, (c) variability in the electricity pricing, and (d) heterogeneity in ESDs in terms of energy storage efficiency, discharge rate and storage capacity.…”