Generation of Distributed System Test-Beds from High-Level Software Architecture Descriptions

  • Authors:
  • John Grundy;Yuhong Cai;Anna Liu

  • Affiliations:
  • -;-;-

  • Venue:
  • Proceedings of the 16th IEEE international conference on Automated software engineering
  • Year:
  • 2001

Quantified Score

Hi-index 0.00

Visualization

Abstract

Most distributed system specifications have performancebenchmark requirements. However, determining the likelyperformance of complex distributed system architecturesduring development is very challenging. We describe asystem where software architects sketch an outline oftheir proposed system architecture at a high level ofabstraction, including indicating client requests, serverservices, and choosing particular kinds of middlewareand database technologies. A fully-workingimplementation of this system is then automaticallygenerated, allowing multiple clients and servers to be run.Performance tests are then automatically run for thisgenerated code and results are displayed back in theoriginal high-level architectural diagrams. Architectsmay change performance parameters and architecturecharacteristics, comparing multiple test run results todetermine the most suitable abstractions to refine todetailed designs for actual system implementation. Wedemonstrate the utility of this approach and the accuracyof our generated performance test-beds for validatingarchitectural choices during early system development.