Skip to main navigation Skip to search Skip to main content

Prediction model for information transmission performance of publish/subscribe systems with heterogeneous servers

  • Harbin Institute of Technology

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

A novel prediction model for information trans-mission performance is presented in this paper, which is suitable for publish/subscribe systems with heterogeneous servers. Firstly, a uniform business logic model is developed by analyzing the principle of several typical types of publish/subscribe distributed systems. In addition, a mathematical model named MMPP(2)/G/1 cascaded queuing model is proposed by introducing heterogeneous server weighting factors into classical queuing model and Markov modulated Poisson process, which can eliminate iteration errors of the queuing model. The proposed prediction model is fully verified by carrying out a series of experiments in an Internet test platform. The results of the experiments illustrate that the proposed prediction model can achieve a higher prediction accuracy than the traditional model.

Original languageEnglish
Title of host publicationProceedings - 8th International Conference on Instrumentation and Measurement, Computer, Communication and Control, IMCCC 2018
EditorsJun-Bao Li
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages426-430
Number of pages5
ISBN (Electronic)9781538682463
DOIs
StatePublished - Jul 2018
Event8th International Conference on Instrumentation and Measurement, Computer, Communication and Control, IMCCC 2018 - Harbin, Heilongjiang, China
Duration: 19 Jul 201821 Jul 2018

Publication series

NameProceedings - 8th International Conference on Instrumentation and Measurement, Computer, Communication and Control, IMCCC 2018

Conference

Conference8th International Conference on Instrumentation and Measurement, Computer, Communication and Control, IMCCC 2018
Country/TerritoryChina
CityHarbin, Heilongjiang
Period19/07/1821/07/18

Keywords

  • Distributed system information transmission
  • Performance prediction
  • Publish/subscribe
  • Queuing model

Fingerprint

Dive into the research topics of 'Prediction model for information transmission performance of publish/subscribe systems with heterogeneous servers'. Together they form a unique fingerprint.

Cite this