Knowledge Mining of Web Service Usage

Website usage continues to increase, while the number of servers increases, their total energy consumption is growing. Meanwhile, server utilization is consistently very low. Servers can be combined to effectively increase the utilization rate and hence the energy consumption even while reducing maintenance costs. This requires, however, that the knowledge of usage of services is updated and the information is sufficient. Usage of services will change as a consequence of users' habits and as a result of internal changes of the service. Therefore, usage of the service has to be capable of being continuously analyzed and to be able to express the result in an easily understandable manner. In the literature, there are several ways to analyze usage of web service. So far, none of these has been found suitable for continuous and automated analysis, in a way that could produce results for programmatic use in the future. We present an automated method for modeling web service usage. The model can help increase knowledge of website usage. As a result, valid actions can be taken in capacity planning and to consequently reduce the energy consumption. At the same time, it will also ensure that service quality remains satisfactory in relation to real demand.