Diagnosing Performance Changes by Comparing System Behaviours

Abstract

The causes of performance changes in a distributed system often elude even its developers. This paper develops a new technique for gaining insight into such changes: comparing system behaviours from two executions (e.g., of two system versions or time periods). Building on end-to-end request flow tracing within and across components, algorithms are described for identifying and ranking changes in the flow and/or timing of request processing. The implementation of these algorithms in a tool called Spectroscope is described and evaluated. Five case studies are presented of using Spectroscope to diagnose performance changes in a distributed storage system caused by code changes and configuration modifications, demonstrating the value and efficacy of comparing system behaviours.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 2010
Accession Number
ADA528600

Entities

People

  • Alice X. Zheng
  • Elie Krevat
  • Gregory R. Ganger
  • Lianghong Xu
  • Michael Stroucken
  • Raja R. Sambasivan
  • Spencer Whitman
  • William Wang

Organizations

  • Carnegie Mellon University

Tags

Communities of Interest

  • C4I

DTIC Thesaurus Topics

  • Algorithms
  • Astronomical Bodies
  • Case Studies
  • Clustering
  • Computations
  • Data Science
  • Data Storage Systems
  • Databases
  • Debugging
  • Estimators
  • Frequency
  • Information Science
  • Instrumentation
  • Machine Learning
  • Software Development
  • Statistical Tests
  • Workload

Fields of Study

  • Computer science
  • Engineering

Readers

  • Fluid Mechanics and Fluid Dynamics.
  • Software Engineering.
  • Systems Analysis and Design