Preserving Distributed Systems' Critical Properties: A Model-Driven Approach

  • Authors:

  • Arvind S. Krishna;Cemal Yilmaz;Atif M. Memon;Adam A. Porter;Douglas C. Schmidt;Aniruddha Gokhale;Balachandran Natarajan

  • Affiliations:

  • Vanderbilt University;University of Maryland;University of Maryland;University of Maryland;Vanderbilt University;Vanderbilt University;Vanderbilt University

  • Venue:
  • IEEE Software
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

As information users move increasingly toward distributed systems to supply basic information infrastructure, the ability of such systems to guarantee basic properties over time increases in importance. Guarantees for certain needs such as latency and throughput can be expressed as quality-of-service (QoS) requirements, whereas guarantees of other properties such as scalability, maintainability, and security can be expressed as persistent software attributes (PSAs). This article describes a model-based process for efficiently coordinating the use of distributed-computing resources to help identify viable configurations of systems and software for meeting stringent QoS and PSA requirements. This process, which the authors call Skoll DCQA, provides an efficient approach to navigating the very large space defined by all possible system and software settings in a modern system. They discuss the tools and procedures used and describe their use of Skoll DCQA with the large, rapidly evolving ACE+TAO middleware suite. Finally, the authors analyze their results and consider their implications for implementing QoS and PSA requirements in distributed systems.