Log based Automated SMI Parameter Identification and Resource Recommendations in Cloud

Objectives: Resource provisioning is the major requirement in cloud provisioning. The major objective is to provide effective resource provisioning can improve the utility of a cloud service at reduced costs. Methods/Analysis: This paper presents an effective method to identify the quality parameters for effective provisioning of cloud resources. User log files are used to identify the quality parameters. It is assumed that the user migrates from a web service, cluster based service or another cloud based service. The log files from these architectures are used to map the SMI parameters and the quality values are obtained by analyzing them. Findings: Experiments were conducted on an access log data with 4.4 million entries and 3 million independent users. The required QoS and provided QoS were plotted and it was observed that most of the points are situated either on the diagonal or in the top left. This exhibits the efficiency of our approach to appropriately identify the user requirements and provide appropriate allocations. The ratio between the time taken for the entire process to complete and the data size was also analyzed for identifying the scalability of the system. It could be observed that as the size of the data increases, the time taken also increases. Hence the time taken is observed to be linear. Applications/Improvement: Identification of quality parameters were never performed with such granularity. Hence the results obtained exhibits effective quality assignments appropriate to user’s needs.