Performance Evaluation and Social Optimization of an Energy-Saving Virtual Machine Allocation Scheme Within a Cloud Environment

Abstract

Achieving greener cloud computing is non-negligible for the open-source cloud platform. In this paper, we propose a novel virtual machine allocation scheme with a sleep-delay and establish a corresponding mathematical model. Taking into account the number of tasks and the state of the physical machine, we construct a two-dimensional Markov chain and derive the average latency of tasks and the energy-saving degree of the system in the steady state. Moreover, we provide numerical experiments to show the effectiveness of the proposed scheme. Furthermore, we study the Nash equilibrium behavior and the socially optimal behavior of tasks and carry out an improved adaptive genetic algorithm to obtain the socially optimal arrival rate of tasks. Finally,

Allocation strategy for time‐sensitive tasks in mobile edge computing: An observable perspective

“…From the numerical results shown in Section 6, we note that when the pure threshold strategy is adopted, too many tasks will be allocated to the MEC system, this will no doubt reduce the social benefit per second. In order to achieve the social optimal threshold strategy in practical applications, we present a charging policy by imposing an appropriate toll 40,41 on tasks for sojourning in the MEC system.…”

Performance analysis and system optimization of an energy-saving mechanism in cloud computing with correlated traffic

“…In the second phase, all the requests are firstly prioritized by a Q-learning based scheduler on each server according to task laxity and task lifetime, then a continuously updating policy is used to assign tasks to VMs. In [30], an energy-saving VM allocation scheme with synchronous multi-sleep and sleep-delay was proposed to address the trade-off between providing higher quality of experience to users and reducing energy consumption. A multi-server queueing model was established to quantify the effects of changes to different sets of parameters, such as the sleep parameter and the sleep-delay parameter.…”