Effective User Requirements Identification and SMI Parameter based WorkLoad Recommendation for Academic Clouds based on Usage logs

摘要

A cloud computing environment is a simplified and centralized platform that can be utilized for performing computations at low cost. The major advantage of this system is automatic allocation of resources when required. This makes the applications deployed on cloud infinitely scalable. However the major downside of this property is that resource upscaling is generally more costly than requirement specification in the early stages. Hence the initial requirement specification phase plays a major role in determining the cost efficiency of the cloud instances corresponding to the user. This paper presents a technique that can be used to identify cloud requirements from the past usage logs of the user. This approach extent the cloud requirements layer by accepting the usage logs and automatically identifying the user's requirements. The major area of concentration in this approach is on the academic industry. Since academic industry operates on predefined timings, identifying requirements from the logs will clearly depict the requirements. The log based requirements are then mapped with the workload details to identify the appropriate workload requirements.