Accelerated Analytical Parametric Method for Reliability Evaluation of Component-based Software in the Time of Component Replacement

In component-based software systems, there are many cases in which a number of components may be replaced with other ones in line with promoting the system. Furthermore, there is a need to adding new components for adding new features or services to the existing system. In such situations, it is advisable to predict impacts from changes to the resulting system prior to making any changes. Reliability is an important factor for verifying the performance of a software system so it can be used for predicting the effects of changes on the system. Also by reliability estimation it can be found that which composition of components is more suitable for a specific system among the existing options. If we want to apply common methods of reliability evaluation to decide about the best selection, inevitably we have to repeat all computations several times. In modeling by Markov processes this means repetition of solving the differential equation system which is difficult and time consuming. Here we introduce a method which using analytical modeling while avoiding repetitive computations, is extendable so that it can be used in the time of replacing components and helps system designers to select a suitable combination of alternative elements.

Keywords:

Component-based software (CBS) , Component replacement , Reliability evaluation , Analytical modeling , Markov processes