Impact analysis for distributed event-based systems

  • Authors:

  • Daniel Popescu;Joshua Garcia;Kevin Bierhoff;Nenad Medvidovic

  • Affiliations:

  • University of Southern California, Los Angeles, CA;University of Southern California, Los Angeles, CA;Two Sigma Investments, New York, NY;University of Southern California, Los Angeles, CA

  • Venue:
  • Proceedings of the 6th ACM International Conference on Distributed Event-Based Systems
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Distributed event-based (DEB) systems contain highly-decoupled components that interact by exchanging messages via implicit invocation, thus allowing flexible system composition and adaptation. At the same time, these inherently desirable properties render a DEB system more difficult to understand and evolve since, in the absence of explicit dependency information, an engineer has to assume that any component in the system may potentially interact with, and thus depend on, any other component. Software analysis techniques that have been used successfully in traditional, explicit invocation-based systems are of little use in this domain. In order to aid the understandability of, and assess the impact of changes in, DEB systems, we propose Helios, a technique that combines component-level (1) control-flow and (2) state-based dependency analysis with system-level (3) structural analysis to produce a complete and accurate message dependence graph for a system. We have applied Helios to applications constructed on top of four different message-oriented middleware platforms. We summarize the results of several such applications. We demonstrate that Helios enables effective impact analysis and quantify its improvements over existing alternatives.