Title: Artificial intelligence for energy and QoS-aware proactive dynamic virtual machines consolidation in cloud-edge data centers

Authors: Mbarek Marwan; Abdelkarim Ait Temghart; Mohamed Lazaar

Addresses: ENSIAS, Mohammed V University in Rabat, Madinat Al Irfane, Rabat, Postal Code 713, Morocco ' TIAD Laboratory, FST, Sultan Moulay Slimane University, Beni Mellal, Postal Code 523, Morocco ' ENSIAS, Mohammed V University in Rabat, Madinat Al Irfane, Rabat, Postal Code 713, Morocco

Abstract: By minimising the number of active servers, virtual machine consolidation (VMC) is a strategy for reducing electricity consumption while maintaining service level agreements (SLAs). However, future resource demands have not been considered by current VMC algorithms, which primarily concentrate on the requirements of all virtual machines (VMs) operating in a data center. Furthermore, the majority of the existing works ignore the security risks associated with VM placement. Therefore, we suggest using recurrent neural networks (RNNs) for capacity planning and multi-objective optimisation for SLA constraints. Initially, the trade-off between the competing objectives of power, performance, and security is evaluated using a multi-objective particle swarm optimisation (MOPSO) technique. Secondly, we provide a novel method for workload prediction based on gated recurrent units optimised by genetic algorithm (GA-GRU). Overall, the findings show that the suggested framework, which takes energy savings and quality of service (QoS) guarantees into account, leads to the optimal design of data centers.

Keywords: cloud edge computing; data centers; VM consolidation; workload prediction; gated recurrent units; genetic algorithm; MOPSO; multi-objective particle swarm optimisation; energy; security; QoS; quality of service.

DOI: 10.1504/IJHPSA.2024.145711

International Journal of High Performance Systems Architecture, 2024 Vol.12 No.1, pp.16 - 32

Received: 24 Apr 2024
Accepted: 27 Aug 2024
Published online: 16 Apr 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article