Secure software design evaluation and decision making model for ubiquitous computing: A two-stage ANN-Fuzzy AHP approach

Ubiquitous computing systems are deployed everywhere to process and service complex applications in real time. It has been noticed that most ubiquitous computing practitioners are trying to develop a highly secure design while maintaining usability. Unfortunately, the highly secure design of software becomes worthless because the usability of software is very low. When improving the security of software design, underlying security design threats (SDTs) and secure design practices (SDPs) play an important role in improving the design security of ubiquitous computing projects. Therefore, the objective of this study is to develop an SSD evaluation and decision-making model using artificial neural network (ANN) and fuzzy analytical hierarchy process (FAHP), for analyzing the SSD threats and its practices to improve ubiquitous computing. Initially, we identified 11 major SDTs and 50 SDPs to mitigate these threats by conducting an empirical study with 50 ubiquitous computing security experts. Moreover, 13 ubiquitous computing security expert panel reviews were conducted to analyze the interrelationships among the SDTs and SDPs using ANN and FAHP approaches. The ANN outcome presents that the foremost significant independent threat in predicting SSD intention is "SDT2: failure to adhere to accepted security design principles" (100%) importance. Furthermore, the FAHP was used to prioritize the list of SDPs. The results and analysis of this study provide a ranked-based decision-making model, which assists the ubiquitous computing practitioners in considering the most weighted SSD practice ("SDP: Implement Sandboxing and respect for User Privacy" having global weight: 0.05088 and global rank: 1) on priority. The study findings will assist real-world industry experts and academic researchers to improve and develop new techniques for the improvements of ubiquitous computing projects.