Probability-based measure to calculate reliability of multi-layer web applications. (2017)

• Record Type: Journal Article
• Authors: Majeed, Hammad; Qamar, Zia Ul; Rasool, Raihan Ur
• Is Part Of: International journal of internet protocol technology. Volume 10:Number 4(2017)
• Page Start: 224
• Page End: 235
• Publication Date: 2017
• Subjects: reliability -- grid computing -- cloud computing -- probabilistic modelling -- multi-tier architecture -- Web 20 reliability -- web applications; File Transfer Protocol (Computer network protocol) -- Periodicals; Multicasting (Computer networks) -- Periodicals; 004.678
• Languages: English
• ISSNs: 1743-8209

Abstract

Distributed environments like, cloud computing and web services, comprise of multiple layers and interacting components. Each component needs to be reliable and dependable for a smooth functioning of the system. Calculating reliability of a system is a complex problem due to non-standardisation of its definition and non-availability of metrics for computing it. This paper proposes a probabilistic model for calculating reliability of multi-layer architecture-based web applications. It is based on hidden Markov model and takes user count and events generated by the users as input and calculate the failure probability of an event at browser and server levels. An object-oriented model of a system provides the foundations for state-chart diagrams for the browser and server, separately. These diagrams are then converted to Markov models, which are used to make a probabilistic model of the system. Finally, execution and the event generation rates are compared to calculate the reliability of a system. The results show that the proposed model accurately depicts the reliability of the system and identifies the scenarios with high probability of failure. This is particularly useful for the decision makers to develop reliable web applications.