Stimulus-based Sandbox for Learning Resource Dependencies in Virtualized Distributed Applications

In this paper, we describe a profiling sandbox for cloud-based distributed applications where the dependencies on available computing resources can be methodically elucidated by dynamically applying a series of unitary perturbations on the underlying computing resources. Each such perturbation applied to an application node acts as a stimulus which propagates to performance meters of dependent nodes and reveals correlations and causal relations between the respective entities. The prime application of the sandbox is to learn the behavior of a distributed application under various resource insufficiency or fault conditions and to build dependency models for Quality of Service management and Root Cause Analysis.