Performance Evaluation and Cost Optimal Service Quality Prediction of Cloud Computing Systems in Unreliable Environments

Cloud computing, is a hot spot in current research and a new information service model. It provides reliable and cheap computing resources to users. The cloud system is built by the Service-Level-Agreement (SLA). Cloud providers must meet SLA performance commitment to their clients. Thus, researchers are concerned about the performance issues of cloud infrastructure. Based on the in-depth analysis of the essential attributes of cloud computing systems, this study introduces a method of service cloud performance analysis by using stochastic processes, queuing networks, time series prediction and other methods and methods. Fine-grained modeling of cloud computing systems in low-reliability environments focuses on describing the probabilistic characteristics of cloud system state transitions and transitions under different task complexity and fault frequency, and establishes a performance model for optimal capacity decision problems. The implementation cost is minimized and the optimization problem is solved numerically by simulated annealing.