RPRA: Reputation-based prioritization and resource allocation leveraging predictive analytics and vehicular fog computing

Abstract

Predictive analytics and vehicle fog computing (VFC) have emerged as important methodologies for enabling seamless and instantaneous access to storage and information in real-time and facilitating the execution of resource-hungry smart applications. VFC-enabled IoT devices that use predictive analytics efficiently can process data in a timely manner. However, besides time sensitivity, better security and privacy solutions for several such smart systems, such as smart healthcare systems, are essential but have rarely been developed so far. With the nature of healthcare information being personal-specific and trust being of paramount significance, our research aims to fill a gap by proposing reputation-based prioritization and resource allocation (RPRA). It is a VFC-enabled framework for time-sensitive, computationally resource-deficient smart healthcare IoT devices that uses latency and energy as key constraints. RPRA is a subjective logic-based heuristic approach that uses context-dependent, indirect, and perceived reputation-based trust. Our method minimizes trust bias and trust-based assaults and uses static weight allocation and a specialized algorithm that guarantees utility maximization. We employed three distinct key metrics, i.e., individual perception, real-time perception, and marginal perception, to identify rogue nodes with hostile intent and remove them from active participation. In a dynamic VFC-enabled healthcare system, our approach outperforms in terms of utility-based reputation values with the highest range on the scale of [0,70] and the least algorithmic complexity.