Parametrics Sensitivity Analysis of Systems: Studying cloud computing impacts

Abstract

Cloud computing paradigm supports reducing costs of acquisition and maintenance of computer systems and enables the balanced management of resources according to the demand. Hierarchical and composite analytical models are suitable for describing performance and dependability of cloud computing systems concisely as well as supports dealing with the massive number of components of such systems. Analytical modeling helps to plan and manage hardware, network, and software infrastructures, either by comparing alternative configurations before implementing a system or by allowing the prediction of effects in system availability and performance after changes in its components. Reliability Block Diagrams, Fault Trees, queuing networks, Markov chains, and stochastic Petri nets are among the formal models commonly adopt in this context. Systems modeling usually handle many different parameters, for both performance and dependability studies. Each parameter can have a distinct impact on availability, reliability and performance measures, hence is crucial to know the Òorder of importanceÓ of the model parameters so that you can decide the appropriate level of attention given to each one. Parametric sensitivity analysis is a method to determine the order of influence of the parameters on the results of a model. The computation of the order of importance for parameters in a hierarchical model is not trivial, because there are multiple components in models of different levels. Considering the challenges just presented, this proposal aims at developing methods for accurate identification of performance, reliability and availability bottlenecks, especially for designers and administrators of complex systems, such as virtualized data centers and Infrastructure as a Service (IaaS) cloud computing environments. This proposition suggests methods for evaluation and detection of bottlenecks of cloud computing systems and a methodology for sensitivity analysis. The methodology is based on hierarchical modeling and parametric sensitivity analysis. This piece of research should introduce methods to build unified sensitivity rankings when distinct modeling formalisms are combined.

Document Details

Document Type
DoD Grant Award
Publication Date
Mar 04, 2019
Source ID
W911NF1810413

Entities

People

  • Paulo Romero Maciel

Organizations

  • Army Contracting Command
  • Federal University of Pernambuco
  • United States Army

Tags

Fields of Study

  • Computer science
  • Engineering

Readers

  • Computational Modeling and Simulation
  • Distributed Systems and Data Platform Development
  • Software Engineering.