Optimal Load Balancing In Three Level Cloud Computing Using Osmotic Hybrid And Firefly Algorithm

• Published in: 2020 International Conference Engineering and Telecommunication (En&T) pp. 1 - 5
• Main Authors: Ojha, Sanjeev Kumar, Rai, Himanshu, Nazarov, Alexey
• Format: Conference Proceeding
• Language: English
• Published: IEEE 25.11.2020
• Subjects: Analytical models; Cloud computing; Computational modeling; Genetic Algorithm; Load Balancing; Load management; Software as a service; Virtual machining; Virtualization; Web services
• DOI: 10.1109/EnT50437.2020.9431250

Cloud Computing is an emerging paradigm of computing which facilitates computing as a service. It enables to use the computing facilities "on-demand", for example, Software as a Service (SaaS), Platform as a Service (PaaS), Infrastructure as a Service (IaaS) etc. This model is best suitable for modern application deployment in which the software /service is required to be deployed in a fast, efficient and cost effective manner. The underlying technology behind cloud computing is virtualization, which enables sharing of the computing resources across multiple users across the globe, connected with each other through internet. Service offering is provided by Cloud Service Provider Companies, some of which are big market players like Amazon (Amazon Web Services; AWS), Google (Google App Engine; GAE) and Microsoft (Microsoft Azure). Load Balancing in Cloud Environment is a central issue of research. This is critical as it leads to efficient utilization of cloud resources, thereby resulting into low cost per user through optimal utilization of resources. The contribution of this paper is two-fold. It extends the approach of using Osmotic Bio inspired algorithm at all the three level of task scheduling, viz, Physical Machine, Virtual Machine and at Task Level. Also, this paper enhances the Osmotic Algorithm with Firefly algorithm which has already been proved significant in load balancing in distributed environments. Also, in this paper, benchmark techniques of load balancing are discussed at depth, both for their effectiveness and limitations. Moreover, techniques are presented for load balancing among virtual machines using Opportunistic Load Balancing (OLB) and LBMM (Load Balance Min-min) scheduling approaches. The Simulation is performed over CloudSim and the results derived are compared to those of the analytical model to prove the validity of the approach.