SAFRank: Multi-Agent based Approach for Internet Services Selection

Abstract

In the era of modern world, organization are preferring to adopt smart solutions for their business tasks and managing huge and complex transactions. These solutions are provided through online application infrastructures of Internet of Things (IoT), cloud, fog, and edge computing. In the presence of numerous prospects, the selection benchmark for such offers becomes vibrant, especially, when there is no supportive platform available. Prevailing approaches provide services by evaluating the quality of service parameters, K-Nearest Neighbours (KNN) classifications, k-mean clustering, assigning scores, trustworthiness and fuzzy logic techniques on customer's feedback. However, these approaches classically depend on seeker’ feedback and do ‘not consider interrelationship between the services. Secondly, these techniques do not follow standards derived by well-known organizations like National Institute of Standards and Technology (NIST), International Organization for Standards (ISO), and IEEE. Feedback may be self-generated or biased and leading to inappropriate recommendation to end users. To resolve the issue, we propose multi agent based approach using service association factor that computes interrelationship values among services appearing together in a package as SAFRank and evaluates it on standards along with dynamically defined quality of service parameters. It assists seekers to select the best services on their preferences from pool of IoT and internet services. The technique is tested on leading cloud vendors and results show that it meets the desires of service seekers in all service models in an efficient manner.