Performance and Energy-based Cost Prediction of Virtual Machines Live Migration in Clouds Moahammad Aldossary 1,2 and Karim Djemame 2 1 Prince Sattam Bin Abdulaziz University, K.S.A. 2 School of Computing, University of Leeds, Leeds, U.K. Keywords: Cloud Computing, Cost Prediction, Workload Prediction, Live Migration, Power Consumption. Abstract: Virtual Machines (VMs) live migration is one of the important approaches to improve resource utilisation and support energy efficiency in Clouds. However, VMs live migration leads to performance loss and additional costs due to increased migration time and energy overhead. This paper introduces a Performance and Energy- based Cost Prediction Framework to estimate the total cost of VMs live migration by considering the resource usage and power consumption, while maintaining the expected level of performance. A series of experiments conducted on a Cloud testbed show that this framework is capable of predicting the workload, power consumption and total cost for heterogeneous VMs before and after live migration, with the possibility of recovering the migration cost e.g. 28.48% for the predicted cost recovery of the VM. 1 INTRODUCTION With the increasing cost of electricity, cloud providers consider energy consumption as one of the biggest operational cost factors to be managed within their infrastructures. Most of the existing studies have focused on minimising the energy consumption and maximising the total resource usage, instead of improving the performance. Further, cloud providers such as Amazon 1 , have established their Service Level Agreements (SLAs) based on service availability without such an assurance of the performance. For instance, during service operation, when the number of VMs increases on the same Physical Machine (PM) stretching its capacity to its limits, resource competition may occur (e.g. once the workload exceeds the acceptable level of CPU such as 85% threshold) leading to VMs performance degradation which may affect the fulfilment of the SLAs and hence the cloud provider’s revenue. Hence to prevent such performance loss effects, it is necessary to have preventive actions such as re- allocating and migrating VMs. VMs live migration is an important mechanism to improve resource utilisation and achieve energy efficiency in Clouds. Live migration allows VMs to 1 https://aws.amazon.com/ec2/sla/ move from one PM to another without any interruption in the service. This mechanism plays an important role in load balancing among the PMs and reduce the overall energy consumption. However, VMs live migration is a resource-intensive operation which has an impact on the performance of the migrating VM as well as the services running on other VMs. Besides, there are additional costs in terms of migration time and energy overhead that need further consideration. Hence, understanding the impact of VM live migration is essential to design an effective consolidation strategy. Previous studies show that in most situations, live migration overhead is acceptable but cannot be ignored as stated in (Voorsluys et al., 2009; Liu et al., 2013). Consequently, predicting the future cost of cloud services can help the service providers offer suitable services that meet their customers’ requirements. Thus, a proactive framework has the advantage of taking preventive actions (e.g. re- allocating or auto-scaling VMs) at earlier stages to avoid service performance degradation. The effectiveness of such framework will depend on potential actuators/decisions to implement at service operation. The first step towards this is a Performance and Energy-based Cost Prediction Framework that 384 Aldossary, M. and Djemame, K. Performance and Energy-based Cost Prediction of Virtual Machines Live Migration in Clouds. DOI: 10.5220/0006682803840391 In Proceedings of the 8th International Conference on Cloud Computing and Services Science (CLOSER 2018), pages 384-391 ISBN: 978-989-758-295-0 Copyright c  2019 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved