M.A. Thesis:

A new hybrid technique for improving load balancing in cloud computing

Abstract: With the advancement of technology and the emergence of new tools and needs, to make better use of resources and facilities, and according to the Internet and the need for users to access their information and resources anywhere and anytime, a new concept called computing Cloud was introduced. In fact, cloud computing is a large data center consisting of a large number of physical servers, which are connected to each other and may not be in one place. Cloud computing is a computer technology to provide customer service on demand. The main goal of the cloud is the optimal use of distributed resources to achieve high power and efficiency. Because the cloud offers a variety of services, it needs to monitor service quality (QoS) to evaluate the services provided to meet user demands and maintain service level agreement (SLA). Currently, load balancing in the cloud (LBC) is one of the main challenges. In this paper, we present a new method to provide a dynamic and efficient load balancing model in the cloud environment to provide the best waiting time calculation rate, resource load calculation and finally select the best virtual machines to determine and evaluate service quality criteria. , Gained. According to previous studies on WRR and DWRR algorithms, which are the main methods of this research, an innovative idea is considered in this research, which is the use of artificial neural network of the neural network of adaptive oscillation theory. This research is based on the two components of waiting time calculation and resource load calculation, which uses the neural network of adaptive oscillation theory (ART), which can provide a near-optimal structure for load scheduling and balance. Based on the simulation performed in MATLAB environment, the simulation results indicate an improvement of the proposed approach compared to the basic reference of the research.

Keywords: Keywords : SLA; QoS; LBC; WRR; DWRR; ART