Performance based Service Level Agreement in Cloud Computing
Afzal Badshah Khattak, Ateeqa Jalal, Tauseef U Rehman · 2015
Recent advances in collaborative computing are best envisioned in cloud computing model. In cloud computing the end user are provided different services (Both in term of hardware as well as software) and are required to pay for the time and amount of services utilized. Various vendors propose different scenarios for such utilization of services. Furthermore, advances in data ware / data mining techniques leads to generate data which are based on adhocism and all generally refined after several runs. The granularity required by such system require resources that are loosely defined in the initial phases but should be robust enough to guarantee scale ups. It is therefore imperative to have a clear cut SLA else migration of services will be disruptive with disastrous consequences for any business organization. For proper utilization of such SLA it should have adequate mechanism for measuring / monitoring the performance parameters. Unfortunately effective monitoring requires low level resource metric where as SLA conform to high end parameters. The situation is further complicated by the absence of frameworks that incorporate performance. More of the models are based on grid computing models which although are established but lack scalability and/or heterogeneity which are essential features of cloud computing.