Abstract
Open source software is widely deployed for both academic and commercial purposes. However, failures and attacks against open source software are often reported. In contrast to traditional software, failures of open source software are reported by users and fixed by developers. Because of the overheads and cost of updates, users do not update their software as soon as the latest patch is released. We are attracted by the optimal maintenance problem for open source software from the users’ perspective. In this paper, we propose periodic and aperiodic maintenance management models with non-homogeneous bug-discovery/correction processes in terms of total expense. Also, according to a dynamic-programming algorithm, we numerically derive the optimal maintenance schedule under the aperiodic policy. In numerical examples, we investigate the efficiency of the proposed policies by mathematical experiments. And, based on bug reports of Hadoop MapReduce, we predictively illustrate the optimal maintenance scheduling for a real open source software product.
Get full access to this article
View all access options for this article.
