Water-Constrained Geographic Load Balancing in Data Centers

Abstract

Spreading across many parts of the world and presently hard striking California, extended droughts could even potentially threaten reliable electricity production and local water supplies, both of which are critical for data center operation. While numerous efforts have been dedicated to reducing data centers' energy consumption, the enormity of data centers' water footprints is largely neglected and, if still left unchecked, may handicap service availability during droughts. In this paper, we propose a water-aware workload management algorithm, called WATCH (WATer-constrained workload sCHeduling in data centers), which caps data centers' long-term water consumption by exploiting spatio-temporal diversities of water efficiency and dynamically dispatching workloads among distributed data centers. We demonstrate the effectiveness of WATCH both analytically and empirically using simulations: based on only online information, WATCH can result in a provably-low operational cost while successfully capping water consumption under a desired level. Our results also show that WATCH can cut water consumption by 20% while only incurring a negligible cost increase even compared to state-of-the-art cost-minimizing but water-oblivious solution. Sensitivity studies are conducted to validate WATCH under various settings.

Computation Peer Offloading for Energy-Constrained Mobile Edge Computing in Small-Cell Networks

“…The closest work to our paper is [11], which aims to minimize the long-term operational cost of data centers subject to a long-term water consumption constraint. However, the long-term constraint is imposed on the entire system whereas in our paper each SBS has an individual energy budget constraint.…”

Enhanced Evolutionary Computing Assisted Robust SLA-Centric Load Balancing System for Mega Cloud Data Centers

“…To alleviate such issues a hybrid model has been developed that considers both inter-service (task) relation as well as variational information to perform dynamic thresholding. We have derived a hybrid model where both InterQuartile Range (IQR) [17] and modified Local Regression (LRR) approaches have been applied to estimate dynamic CPU threshold. IQR which is a statistical dispersion method signifies the in between the first and the third quartile (1).…”

“…As already stated the classical Loess method [17] based LRR often gives inferior performance thus to alleviate this issue, in this paper the classical Least-Square (LR) model has been augmented to the bisquare. The LR enhancement has been performed in iterative manner so as to calculate the initial fitting where we have estimated tricube weights using a Tricube Weight Function (TWF).…”

“…To solve such issues approaches like overload detection and distributed computing approaches have been explored in a number of researches. To solve this problem numerous meta-heuristic approaches like Particle Swarm Optimization (PSO) [11][12][13][14][15], Genetic Algorithm (GA) [16][17][18][19], Ant Colony System (ACS) [20][21][22], Cuckoo Search Optimization (CSO) [24], Honey Bee Foraging (HBF) [7,10,23,25], etc., have been proposed. Undeniably, these approaches have made effort to augment VM placement policy however lacks dealing with local minima and convergence, which is predominant in major meta-heuristic approaches.…”

Lifetime-Aware Cloud Data Centers: Models and Performance Evaluation

“…Additionally, in [11] authors take into account capacity provisioning in cloud DCs with the goal of reducing the brown energy-i.e., energy generated from non-renewable sources. Finally, in [12] authors consider the consumption of water for dispatching jobs in geographical DCs. All of these works do not consider the impact of the proposed policies on the failure rate of the servers in each DC.…”