Performance impact of imitation in Multi-Objective Security Service Provisioning

This paper presents the results of a case study conducted by author in order to provide evidence in support of the notion that the use of imitation (recall) has a positive impact on dynamic selection of cryptographic algorithms in Multi-Objective Security Service Provisioning (MOSSP) applications. The case study included the use of a full-scale proof-of-concept prototype designed and developed by author using Microsoft. NET. The prototype was equipped with Runtime Optimal Cryptographic Algorithm Selector (ROCAS) engine that chose provider ciphers by either exploration using a hybrid Genetic Algorithm (GA) or imitation (recall) via the history of past decisions. In the case of hybrid GA, ROCAS used Weighted Metric Method to find optimal providers (servers) ciphers with security, speed, and energy attributes for a diverse array of security services demanded by requesters (clients) with often competing security, speed, and energy objectives. The results of hybrid GA's optimal cipher selections were written to DBMS to be used in future recall. As in biological life, the results of the case study show that the use of imitation (recall), i.e. bypassing the hybrid GA's search and selection process, has a positive impact on performance of MOSSP applications.