Datacentre Event Analysis for Knowledge Discovery in Large-Scale Cloud Environments

Abstract

Cloud Computing paradigm is increasingly being adopted for various business needs, and improving the quality of Cloud services has naturally become essential for the providers to stay competitive. The characteristics of Cloud-based workloads are still not perfectly clear, and the extensive level of heterogeneity among both the Cloud workloads and the server resources imposes various levels of intrinsic and extrinsic complexities in achieving efficient datacentre management. To this end, this chapter conducts extensive analysis on a real-world datacentre execution trace, with the motivation of exhibiting the inherent knowledge among the workload behaviours and server usage patterns in a large-scale datacentre environment for the scope of decision making in Cloud datacentre management. The analysis presented in this chapter is conducted by subjecting the workload behaviours with various periodical effects. The sever usage patterns are analysed by dwelling into their usage frequencies for scheduling the incoming jobs. The analysis presented in this paper is believed to provide sufficient knowledge to the Cloud providers for achieving efficient datacentre management, and further the inferences obtained in this analysis find applications in workload behaviour modelling and prediction analytics in Cloud datacentres.