Component-based tools and techniques have received wide popularity among researchers and software practitioners. Component-Based Software Development (CBSD) not only enhances the productivity, but also reduces the cost of software development. The growing size of component-based systems brings greater challenges towards its maintenance. In a large and complex Component-Based System (CBS), the development cost is much lower than its maintenance cost, which is a major concern.Maintenance of such software consumes a lot of time and resources. Change effects need to be considered for planning the maintenance phase. Understanding dependencies among components is an essential requirement to carry out the change impact analysis.
Therefore, it becomes necessary to measure the change impact on the whole system. The paper concludes that the higher the dependency of a component, the higher the change impact on the system. In support of the applicability of the hypothesis, a case study is also presented. Finally, the paper presents an algorithm to compute the change impact on the system.
There is a continuing interest in using the components to build large and complex software. Component-Based System (CBS) can be viewed as a set of interrelated components. Component-Based Software Development (CBSD) approach achieves high quality and productivity (Vieira et al., 2001), but maintaining a large scale CBS is a great challenge for its developers. Software maintenance cost for such systems is very high as compared to their development cost. Therefore, controlling the maintenance phase is an important concern for its researchers. Dependency is an important factor that directly affects the maintainability. To fulfill the changing requirements of the user, modifications in the software are required.
Modification may be in the form of code change or relations change among the components. In the CBS, components are dependent on each other due to which modification in a component also requires modification in other components. A component is said to be dependent on another component if it provides/uses some service(s) to/from other components. In the Component Dependency Graph (CDG), it is denoted by an edge between the two components. Changes in the component which provides the services are reflected back into the component which uses the services. In this way the changes are rippled in the system through dependencies between the components. A higher change impact on the system increases the efforts as well as the cost required for maintenance. By measuring the change impact in advance, change cost and efforts can be estimated, which may help to control them in a better manner (Abdi et al., 2006).
|
