Dynamic Modelling of Tactics Impact on the Stability of Self-aware Cloud Architectures

Maria Salama, Ahmed Shawish, Rami Bahsoon

Research output: Chapter in Book/Report/Conference proceedingConference contribution

192 Downloads (Pure)

Abstract

Given the elasticity, on-demand nature, and runtime dynamics of the cloud, a stable self-adaptive architecture should keep the fulfilment of Quality of Service objectives stable, while performing stable adaptations that converge towards these objectives. The dynamic management and selection of architectural tactics, as adaptation mechanisms, shall be in the heart of the adaptation process, as being essential for effective and stable adaptations. This calls for measuring the impact of tactics on the stability of inter-related quality attributes during run-time. In this paper, we introduce a Markovian-based analytical model for dynamically assessing the impact of tactics on the stability behaviour of self-adaptive cloud architectures. The model also employs self-awareness capabilities for betterinforming the selection of optimal tactics configurations leading to stability. Experimental evaluations have shown the accuracy and efficiency of the model in measuring and predicting the impact of tactics on stabilising the Quality of Service provision and the adaptation process.
Original languageEnglish
Title of host publication2016 IEEE 9th International Conference on Cloud Computing
PublisherIEEE Computer Society Press
Pages871-875
Number of pages5
ISBN (Print)978-1-5090-2619-7
DOIs
Publication statusPublished - Aug 2016
EventThe 9th IEEE International Conference on Cloud Computing - San Francisco, United States
Duration: 27 Jun 20162 Jul 2016

Conference

ConferenceThe 9th IEEE International Conference on Cloud Computing
Country/TerritoryUnited States
CitySan Francisco
Period27/06/162/07/16

Fingerprint

Dive into the research topics of 'Dynamic Modelling of Tactics Impact on the Stability of Self-aware Cloud Architectures'. Together they form a unique fingerprint.

Cite this