Fair rewarding in colocation data centers: Truthful mechanism for emergency demand response

Abstract

Reducing servers' power usage in data centers upon utility's request has been emerging as a valuable demand response resource for enhancing power grid's efficiency and reliability, especially during emergency events (e.g., extreme weather) that result in electricity production shortage and put the grid in jeopardy. Nonetheless, for demand response in multitenant colocation data centers, operators may have to leverage expensive and environmentally-unfriendly diesel generation, because individual tenants manage their own servers' power usage without coordination and are typically charged by data center operators based on fixed power contracts that provide no incentives for demand response. This paper focuses on emergency demand response (EDR) and proposes an auction-based incentive mechanism, called FairDR, that incentivizes and coordinates tenants' energy reduction through financial rewards for enabling cost-effective and low-carbon EDR in colocation data center. FairDR decides tenants' energy reduction online without knowing a priori the future energy reduction requirements. It is proved that FairDR ensures tenants' truthfulness in the auction process, attains a bounded overall cost saving compared to the offline optimum which knows all the demands, and guarantees fairness (i.e., similar rewards are offered if tenants reduce the same amount of energy) that is largely absent in the existing auction mechanisms. Finally, trace-driven simulations are performed to validate our analysis and demonstrate that FairDR outperforms the existing mechanisms by improving fairness and achieving a good cost saving that is comparable to the offline optimum.

A Truthful FPTAS Mechanism for Emergency Demand Response in Colocation Data Centers

“…Zhou et al [21] studied demand resource on geodistributed cloud through VCG-based mechanism, in which the utility of each agent depends highly on its interactive workload. Sun et al [22] considered fairness among the mechanism design and provided online mechanism with competitive ratio of 3.2 in expectation. Ahmed et al [23] proposed a contractbased mechanism, in which the colocation operator offers a set of contracts (i.e., a pair of energy reduction and rewards) to tenants and tenants can voluntarily select none or one of the contracts to accept, while the objective is to minimize the operator's cost, the sum of rewards plus the cost of BES.…”

Energy management for cloud computing environment.

“…Similarly, different works evaluate methodologies to enable the DC participation in DR programs. (WANG et al, 2013), (LE et al, 2016), (SUN et al, 2015), (LIU et al, 2014) and (…”

Auction-Based Mechanisms for Power Grid Balancing using Cloud Datacenters

“…They use VCG auction during each iteration and their objective function tries to minimize the total system cost (the incurred cost of the colocation tenants for deferring their workloads plus the cost or running on-site diesel generators if needed). Other proposals of Sun et al [120] and [121] have also focused on the same problem of determining optimal tenant rewards for colocations participating in DR programs. In [120], they developed a truthful online auction that focused on guaranteeing fair reward allocation among the participating colocation tenants (i.e.…”

“…Other proposals of Sun et al [120] and [121] have also focused on the same problem of determining optimal tenant rewards for colocations participating in DR programs. In [120], they developed a truthful online auction that focused on guaranteeing fair reward allocation among the participating colocation tenants (i.e. guarantees that the colocation tenants get the same amount of reward for each unit of reduction in energy demand).…”