Reengineering Cost Estimation using Scrum Agile Methodology

Abstract

Estimating the budget for developing software is one of the prime tasks for software stakeholders. Good estimation increases the customer faith and goodwill for the development company. Many estimation techniques exist for estimating the cost of the software. Estimating reengineering projects are equally important. Researchers estimated cost of Reengineering using conventional algorithmic estimation methods. They also used classical software development approaches to perform reengineering. Conventional estimation methods are suitable in an environment where requirements are predefined and fixed. Practically, these methods can not fit in today’s software development environment. We need more realistic approach to estimate. Since a decade, we have witnessed a change in the Software development approaches. Now software development process is more people centric and realistic for their stakeholders. This change in process is due to Agile. Agile methodology has gained the interest of both customers as well as developers. The main objective of this research is to estimate the cost of reengineering with consensus based estimation technique of Scrum development methodology. Agile Reengineering model is also proposed for estimation and performing reengineering. Thus the research is aimed to provide a model, which not only helps in performing the reengineering estimations but also guides how to perform reengineering. Scrum approach with sprint iteration of three weeks is used to perform reengineering. Chidamber and Kemerer (CK) metric is applied to determine the complexity metrics for various classes of the software. Reengineering is performed to make the project more maintainable by reducing the CK metric complexity. Various tools used in this work include CK java Metric tool (CKJM) ver-1.9 for calculation of CK metrics suit, IBM Rational Rose ver7.5 for Unified Modeling, Rapid Minor studio ver7.1 for determining the reengineering requirements of the software.