Performance-Aware Management of Cloud Resources: A Taxonomy and Future Directions

Dynamic nature of the cloud environment has made distributed resource management process a challenge for cloud service providers. The importance of maintaining the quality of service in accordance with customer expectations as well as the highly dynamic nature of cloud-hosted applications add new levels of complexity to the process. Advances to the big data learning approaches have shifted conventional static capacity planning solutions to complex performance-aware resource management methods. It is shown that the process of decision making for resource adjustment is closely related to the behaviour of the system including the utilization of resources and application components. Therefore, a continuous monitoring of system attributes and performance metrics provide the raw data for the analysis of problems affecting the performance of the application. Data analytic methods such as statistical and machine learning approaches offer the required concepts, models and tools to dig into the data, find general rules, patterns and characteristics that define the functionality of the system. Obtained knowledge form the data analysis process helps to find out about the changes in the workloads, faulty components or problems that can cause system performance to degrade. A timely reaction to performance degradations can avoid violations of the service level agreements by performing proper corrective actions including auto-scaling or other resource adjustment solutions. In this paper, we investigate the main requirements and limitations in cloud resource management including a study of the approaches in workload and anomaly analysis in the context of the performance management in the cloud. A taxonomy of the works on this problem is presented which identifies the main approaches in existing researches from data analysis side to resource adjustment techniques.