Future Generation Computer Systems 87 (2018) 35–42 Contents lists available at ScienceDirect Future Generation Computer Systems journal homepage: www.elsevier.com/locate/fgcs Adaptive selection of dynamic VM consolidation algorithm using neural network for cloud resource management Joseph Nathanael Witanto a , Hyotaek Lim b, *, Mohammed Atiquzzaman c a Department of Ubiquitous IT, Dongseo University, 617-716 Busan, South Korea b Division of Computer Engineering, Dongseo University, 617-716 Busan, South Korea c School of Computer Science, University of Oklahoma, Norman, OK 73019, United States highlights • Studies on multiple dynamic VM consolidation algorithms on cloud environment. • Discusses about trade-off between energy and SLA violation due to VM migration. • Proposes adaptive system which chooses best algorithm based on provider’s priority. • Evaluates performance of proposed system and individual VM migration algorithms. article info Article history: Received 16 October 2017 Received in revised form 28 February 2018 Accepted 23 April 2018 Available online 3 May 2018 Keywords: Cloud computing Infrastructure as a service Resource management Dynamic consolidation Virtual machine migration Neural network abstract Cloud resource management becomes more important with the increasing usage of cloud resources. With various cloud options available, cloud provider may have different priority in managing the resource through resource scheduling and provisioning. Dynamic VM (Virtual Machine) consolidation algorithm is one of the techniques which can be used to reduce energy consumption through VM migration. Higher VM migration may lead to lower energy consumption and higher SLA violation. Although previous research has successfully decreased energy consumption and SLA violation, cloud providers may need to manage trade-offs between energy and SLA violation through availability of priority in the system. This paper proposes neural network-based adaptive selection of VM consolidation algorithms which adaptively chooses appropriate algorithm according to cloud provider’s goal priority and environment parameters. Dataset generation and performance evaluation using simulations on real-world PlanetLab VMs workload trace showed that adaptive selector produced better average performance score than independent methods on various evaluation priority. © 2018 Elsevier B.V. All rights reserved. 1. Introduction Cloud computing is the technology to enable provisioning of resources (hardware and software) over the Internet. The pay-as- you-go pricing gives opportunity for cloud users to eliminate up- front cost [1]. Cloud resource management is important because it affects performance, functionality, and cost of a cloud system [2]. The integration of Internet of Things and cloud computing (Cloud of Things) for developing smart applications [3] will increase the number of cloud computing usage and the importance of cloud resource management. There are two types of resources to be managed, physical re- sources (CPU, memory, storage, workstation, network elements, * Corresponding author. E-mail addresses: josephwitanto@gmail.com (J.N. Witanto), htlim@dongseo.ac.kr (H. Lim), atiq@ou.edu (M. Atiquzzaman). sensors) and logical resources (Operating System, energy, network bandwidth, information security, protocols, APIs, and network de- lays) [4]. Since cloud involves large number of shared resources affected by external events, cloud resource management requires complex policies for optimization. For each cloud delivery model (Infrastructure as a Service, Platform as a Service, and Software as a Service), the resource management strategy also differ [2]. Important issues in cloud resource management include re- source provisioning and resource scheduling [4]. Resource provi- sioning is the allocation of a cloud provider’s resources to a cus- tomer. Inefficiency of resource provisioning leads to either over- provisioning or underprovisioning problem [4]. Since in the IaaS (Infrastructure as a Service) model services are deployed as Virtual Machine (VM), the resource provisioning problem is reduced to how to place the VM [5]. https://doi.org/10.1016/j.future.2018.04.075 0167-739X/© 2018 Elsevier B.V. All rights reserved.