Measurement and Prediction of SaaS Reliability in the Cloud
Mohammed O. Alannsary, Jeff Tian · 2016
Reliability is one of the most important attributes of quality for SaaS (Software as a Service). We develop a method of measuring and predicting the reliability of SaaS through analyzing web server logs. The observed failures in conjunction with the workload are analyzed for reliability assessment and prediction. The workload is measured by counting the number of system invocations and the time taken for each invocation. Operational reliability is measured using IDRM (input domain reliability modeling). Potential reliability change or growth is estimated through SRGM (software reliability growth modeling). We applied our proposed method to a CMS (content management system) installed on a Cloud configured using Amazon web services to demonstrate its applicability and effectiveness.