Addressing Node Failures Using Node Management & Cluster Management Techniques in Cloud Computing

Abstract

Cloud computing has been the rave of users globally over the past few years. This has mainly been because of the many different services available in the cloud that has made it the technology of choice for many users. Whether knowingly or unknowingly once a user connects to the Internet and accesses a service be it an application or storage they are using a cloud service. Availability, however, has been a thorn in the flesh for most Cloud Service Providers (CSPs). One of the sources of outages at cloud infrastructure level has been node failures. In this context the nodes are the servers found within the datacenter (DC) that house the virtual machines (VMs), which in turn execute the tasks required by users. The objective of this study was to determine whether node management and cluster management are proactive effective availability Mechanisms (AMs) in countering node failures. To investigate this we simulated node failures in the DC using a random number generator. We then countered these failures using a node management technique and then secondly using cluster management, using CloudSim simulation toolkit. The availability of the infrastructure was then measured using two parameters namely service availability and execution availability. The findings showed that both techniques are effective techniques in countering node failures at infrastructure level; further effective management of the cluster results in effective management of the node at infrastructure level; the converse is also true.