Enhancing Software Reliability against Soft-Error using Minimum Redundancy on Critical Data
Saeid A. Keshtgar, Bahman Arasteh · International Journal of Computer Network and Information Security · 2017
Nowadays, software systems play remarkable roles in human life and software has become an indispensable aspect of modern society. Hence, regarding the high significance of software, establishing and maintaining software reliability is considered to be an essential issue so that error occurrence, failure and disaster can be prevented. Thus, the magnitude of errors in a program should be detected and identified and software reliability should be measured and investigated so as to prevent the spread of error. In line with this purpose, different methods have been proposed in the literature on software reliability; however, the majority of the proposed methods are inefficient and undesirable due to their high overhead, vulnerability, excessive redundancy and high data replication. The method introduced in this paper identifies vulnerable data of the program and uses class diagram and the proposed formula. Also, by applying minimum redundancy and duplication on 70% of the critical data of the program, the proposed method protects the program data. The evaluation of the operation of the propose method on program indicated that it can improve reliability, reduce efficiency overhead, redundancy and complexity.