Efficient Ultra-Elastic Resource Provisioning through Hyper-Converged Cloud Infrastructure using Hybrid Machine Learning Techniques
Bharanidharan.G1, S.Jayalakshmi2
1Bharanidharan.G, Research scholar, Dept. of Computer Science VISTAS, Pallavaram Chennai, Tamilnadu, India.
2Dr.S.Jayalakshmi, Professor, Department Of Computer Applications VISTAS, Pallavaram Chennai, Tamilnadu, India.
Manuscript received on March 15, 2020. | Revised Manuscript received on March 24, 2020. | Manuscript published on March 30, 2020. | PP: 4367-4374 | Volume-8 Issue-6, March 2020. | Retrieval Number: F9753038620/2020©BEIESP | DOI: 10.35940/ijrte.F9753.038620
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Ultra-flexibility is future asset provisioning method so as to deftly meet the clients’ prerequisite in powerful way. Be that as it may, more components are required for execution improvement, for example, CPU and the capacity. It is trying to decide a reasonable edge to effectively scale the assets up or down. In this paper, we propose an efficient resource provisioning using hybrid machine learning techniques (ERP-HML) that emphasis on mutually advance the vitality utilization of servers and system. Here, the proposed asset provisioning is utilized for ultra-versatile cloud benefits in hyper-joined cloud framework. In a Hyper-converged Infrastructure the resources such as CPU, storage and Network will be virtualized and software-defined as pools to meet the current demand. The principal commitment is to present an artificial plant optimization algorithm to improve the administration inertness and lessening over-provisioning of flexible cloud administrations. The subsequent commitment is to delineate a deep Q neural network (DQNN) for anticipating the server’s preparing load. At that point, an improved hunting search (IHS) calculation is use to register the quantity of assets that must be provisioned dependent on the anticipated burden. The principle target of proposed ERP-HML strategy is precisely foresee the handling heap of a conveyed server and gauge the proper number of assets that must be provisioned to decrease vitality utilization. At last, the presentation of the proposed ERP-HML strategy is contrast and the current condition of-craftsmanship strategies as far as energy consumption, infrastructure costs and QoS.
Keywords: Virtual Machines; Hybrid Machine Learning; Artificial Plant Optimization; Deep Q Neural Network; Improved Hunting Search
Scope of the Article: Mechanical Design.