Exploring Architecture-Based Reliability Analysis of Current Multi-layered Web Applications

Abstract

Web applications architecture evolved from simple web sites add-ons to complex n-layer applications. However, identifying components in this domain is usually a subjective task, as web applications typically comprise web pages, scripts, forms, applets, servlets or simply web objects. As a result of this subjectivity, a component-based life-cycle might reflect on inconsistencies not only on a clear definition of web components, but also on the process development itself. In addition, it is hard to identity which components are more critical according to specific tasks, such that developers could spend more time to improve their design. That quality certainly comprises reliability, availability and security, summing up as dependability attributes. The application of architecture-based reliability analysis techniques in various domains have contributed to solve those problems. However, very little has been done towards the assessment of current web applications in a real-life setting. In this work, we explore the feasibility to apply an architecture-based reliability analysis method in a real-life web application. Our preliminary results show the potential use of this method for the web applications domain, with a considerably accuracy.