Final Workload, Load Propagation, and Application Models

Summary
Advancedworkload and application models that a fine granularity capture the composition and statistical properties of request-based workloads for data centre applications, and provide stable predictive capabilities including, e.g., proactive auto-scaling techniques, burst and anomaly detection techniques, and provisioning of simulation models that allow (re-)construction of artificial traces that retain the statistical properties of the original traces. Load propagation models that build on the preliminary load models and provide more advanced application elasticity and remediation capabilities such as probabilistic estimations for alternative load paths and component health-monitoring based recovery mechanisms that can be used in deployment optimisationand self-tuning of applications.