Visualizing the underlying trends of component latencies affecting Service Operation Performance

Abstract

This paper presents a technology agnostic method for extracting the underlying distinct patterns of variations in the overall Performance of a Service Operation for changes to different application components supporting the Service Operation in a computer based Service Provider-Consumer contract. This short paper advocates that visualizing these patterns would help in early projection of the operation's performance due to modification of the application components/processing catering to the operation, without the need of repetitive performance and load testing of the whole service. Lookup datasets against different component configurations are created to associate the variability of component processing impedances to the service operation's Performance and best fit regression types are applied to enable trend extrapolation and interpolation.