Modeling web usage profiles of cloud services for utility cost analysis

Early proponents of public cloud computing have come to identify cost savings a key factor for adoption. However, the adoption and hosting of a web application in the cloud does not provide any such guarantees. This is in part due to the utility pricing model that dictates the cost of public cloud resources. In this work we seek to model and simulate data usage for a web application for the purpose of utility cost analysis. Although much research has been performed in the area of web usage mining, previously proposed models are unable to accurately model web usage profiles for a specific web application. In this paper, we present a simulation model and corresponding algorithm to model web usage based on empirical observations. The validation of the proposed model shows that the simulated output conforms to that of what was observed and is within acceptable tolerance limits.