International Journal of Computing and Digital Systems ISSN (2210-142X) Int. J. Com. Dig. Sys. 9, No.3 (May-2020) E-mail: muhammadmateen@cqu.edu.cn, nasrullah_ce@yahoo.com, hayat.uestc@yahoo.com, tooba.tehreem28@yahoo.com, azeem.akbar@ymail.com http://journals.uob.edu.bh A Self-Adaptive Resource Provisioning Approach using Fuzzy Logic for Cloud-Based Applications Muhammad Mateen 1 , Nasrullah 1 , Shaukat Hayat 2 , Tooba Tehreem 3 , and Muhammad Azeem Akbar 1 1 School of Big Data & Software Engineering, Chongqing University, Chongqing, China 2 School of Information and Software Engineering, University of Electronic Science and Technology of China, Chengdu, China 3 School of Computer Science, National University of Computer and Emerging Sciences, Islamabad, Pakistan Received 18 Sep.2019, Revised 1 Jan. 2020, Accepted 1 Mar. 2020, Published 1 May 2020 Abstract: The resource adaptation on demand is an important factor in the field of cloud computing. During runtime, autonomic resource provisioning is not an easy task to choose the accurate amount of resources for service-based cloud applications. For this reason, it is required to guess the future demands for self-adaptive resources to deal with the irregular requests based on runtime workload changes of service-based cloud applications. In this paper, an efficient approach to increase the utilization of resources is proposed that is based on self-adaptive computing with fuzzy logic. Additionally, the proposed fuzzy logic approach enhanced the performance of planning phase for better decision making. Based on fuzzy logic, cloud applications having self-learning provisioning resources outperformed the hybrid resource provisioning approach. To calculate the quality of the proposed technique, real-world ClarkNet and NASA workload traces are used. The results of experiments show that the proposed technique has decreased the entire cost and has boosted the utilization of resources as compared to the other contemporary techniques. Keywords: Autonomic computing, Cloud computing, Fuzzy logic, Resource adaptation 1. INTRODUCTION In the field of information technology, cloud computing has become a source of knowledge that is useful to access the information, media, resources, and services. Cloud computing is a modern style of computation through which users, consumers, and customers can get the benefits of virtualized resource access. To handle the uncertainty and dynamics of resource access, the term “self-adaptively” is used in field of cloud computing. Here, the autonomic system is considered as a system that is able to adjust the performance according to the prediction of the uncertain state and its environment. Nowadays, self-adaptively has become a hot topic for challenging research in the cloud- based applications [1]. Self-adaptive can be explained with the help of its properties including, self-configuring, self-healing, self- optimizing and self-protecting. Self-configuration means installing, integrating and composing of software entities according to the change of uncertain environment. Self- healing is a process of diagnosing, discovering, and responding to interruptions. It also performs some actions to avoid the failure. Self-optimizing can also be considered as self-adjusting or self-tuning, is the ability of resource allocation and manage performance to full-fill the requirements of users. Self-protection enables the system more secure and effective [2]. In cloud computing, self-adaptive resource utilization is not an easy task to select an optimal amount of resources for service cloud. For that purpose, it’s necessary to predict the future demands on the basis of past and current utilization of resources. In this context, to guess the future demands is a great achievement. The purpose of resource prediction is to minimize the cost and to maximize the utilization of demanded resources. Zia Ullah et al. [3] introduced a real-time system of resource utilization on the basis of previous demands predictions. For this purpose, researchers followed Gaussian distribution which includes Autoregressive Integrated http://dx.doi.org/10.12785/ijcds/090301