An efficient framework for trust evaluation of secure service selection in fog computing based on QoS, reputation, and social criteria

Fog computing has been recognized as a new paradigm of distributed environment which is the result of cloud computing extension. Fog layer is located between cloud data center layer and Internet of Things device layer to provide data storage, processing through executing applications on the network edges. However, security is the main challenge of fog technologies from various aspects due to processing, managing and storing private information by fog servers. Therefore, adopting a trust management system in accordance with determined criteria by trustor is the main issue of fog computing technology to address the level of trustworthiness of trustee. The efficient metrics on trustworthiness value, including Quality of Service (QoS), social relationship, and past experience such as reputation, conflict with each other which is considered as a multi criteria decision making problem. Therefore, the determination of priority and contribution level of effective metrics on trustworthiness value is an important issue that must be addressed. In this paper, a complete set of effective criteria on secure service selection in a fog computing environment is identified. Moreover, an efficient multi-criteria decision making method with corporation of fuzzy and Best Worst Method are utilized to determine the contribution level of each metric on trust level considering uncertainty of metrics. The results indicate that QoS has the great impact on secure service selection with a degree of 0.470. Furthermore, central information of security including access control, integrity and confidentiality are ranked as the most effective metrics on secure service selection with degrees of 0.109, 0.099 and 0.088. Moreover, honesty from the social relationship criteria has high impact on secure service provision with a degree of 0.074. Also, the proposed framework outperforms convergence and accuracy of trust value with 30% and 5%, respectively compared to two-way trust management framework. Besides, the proposed framework outperforms the percentage of bad fog servers selected as service providers with 25% in comparison with two-way trust management systems.