Reliability Management Framework and Recommender System for Hyper-converged Infrastructured Data Centers

Cloud computing is a miracle of the 21st century that grasps its significance in almost every field of the digital world. In this era, most of the organizations look towards cloud service providers to acquire resources such as processing power, data storage, networks, servers, and etc. After the evolution of clouds, the paradigm shift towards the development and maintenance of data centers. The traditional infrastructure of data centers allows the user to select resources from different vendors. On the other hand, in converged data centers resources are merged and sold as a single unit. The recent trend is to build the data centers on the hyper-converged technique in which storage, computational and networking resources are merged together as a single unit with a layer of virtualization. In hyper-converged infrastructure (HCI) data centers reliability management is one of the challenging tasks because of modular dependency and complex layout. This paper presents, a hierarchical framework to measure the reliability of the HCI data centers. The proposed framework measures the reliability of networking, storage, and computational resources on the basis of their availability, outage, and downtime with the help of Neural Networks. In the case of an unreliable situation, the framework recommends the areas that need to be fixed. Moreover, the recommender system is developed based on the proposed framework for the data center to measure reliability. The experimental results show that the proposed framework gives 94% accuracy. The development of such a system enhances the reliability of the data center.