A Two-Stage Framework for the Multi-User Multi-Data Center Job Scheduling and Resource Allocation

Jianpeng Lin,Delong Cui,Qirui Li,Zhiping Peng,Jieguang He

doi:10.1109/access.2020.3033557

Jianpeng Lin, Delong Cui + Show 3 more

Open Access

https://doi.org/10.1109/access.2020.3033557

Copy DOI

Journal: IEEE Access	Publication Date: Jan 1, 2020
Citations: 58	License type: CC BY 4.0

Affiliation: Guangdong University of Petrochemical Technology

Abstract

With the development of artificial intelligence and the Internet of things, the prospects of cloud computing applications have become broader, and the number of users and cloud data centers (CDCs) has exploded. It is a challenge to realize efficient job scheduling and resource allocation of multiple users and data centers. However, the traditional scheduling model based on heuristic algorithm has some limitations in the complex and changeable cloud environment. In addition, many existing single-agent models rarely consider the multi-objective global optimization of the system. Therefore, this paper proposes a two-stage job scheduling and resource allocation framework that adopts multiple intelligent schedulers to solve the cooperative scheduling problem between job scheduling and resource allocation. A heterogeneous distributed deep learning (HDDL) model is used in the job scheduling stage to schedule multiple jobs to multiple cloud data centers. The deep Q-network (DQN) model is a resource scheduler to deploy virtual machine to physical servers for execution. Extensive numerical results show that both HDDL-baesd job scheduler and DQN-based resource allocator outperform the benchmark algorithm in terms of optimizing energy consumption and job delay. Furthermore, the proposed framework not only can achieve a global near-optimum by achieving local optimization at each stage but also has good scalability and low computation delay.

Highlights

The rapid implementation of cloud computing has resulted in significant investment and fast growth
It is worth noting that the convergence value of the heterogeneous distributed deep learning (HDDL) curve is close to 1, which means that the closer the ratio is to 1, the better the performance of scheduling model
The results clearly show that the computation time of both HDDL and deep Q-network (DQN) is less than the classic heuristic algorithm multi-objective particle swarm optimization (MoPSO) and iterative algorithm FERPTS, respectively

Summary

Introduction

The rapid implementation of cloud computing has resulted in significant investment and fast growth. With the explosive growth of applications, users, and cloud service providers (CSPs), efficient job scheduling and resource allocation of multi-user multi-CSPs has become a major challenge [3]. An improved heuristic algorithm is the most common solution to the scheduling problems of cloud computing [5]. Alkayal et al [6] combined multi-objective optimization (MOO) and particle swarm optimization (PSO) to optimize resource allocation, aiming to schedule jobs to virtual machines (VMs) with minimal waiting time and maximum system throughput. Hu et al [7] devised a scientific workflow multi-objective scheduling algorithm for the reliability of workflow scheduling in a multi-cloud environment, with a goal to minimize the completion time and cost of workflow under reliability constraints. A number of hybrid algorithms [8, 9, 10] combine the excellent characteristics of multiple heuristic algorithms to achieve good performance

Methods

Results

Conclusion