Optimizing Age of Information and Security of the Next-Generation Internet of Everything Systems

Abstract

We analyze the information exchange and interactions among the major components, i.e., people, things, data, and processes, of the Internet of Everything (IoE) system, where raw data generated by people and things must be processed to obtain relevant higher level information that can be utilized by IoE processes for decision making and actions. Accordingly, the value of information obtained in the IoE system depends on the Age of Information (AoI)—time elapsed from the moment when raw data is generated to the moment when the data is processed and delivered to the processes. To reduce the AoI, the system is realized in the multiaccess edge computing network, where data can be processed by the edge devices (EDs) in proximity to people, things, and processes. The system security and resilience are further enhanced through coded distributed computing when each data input of EDs is encoded with a specific encoding function so that the final result of data processing by EDs can be recovered even if some processing outputs of EDs are erroneous or delayed. We then define a stochastic optimization problem where the AoI, security, and resilience are optimized jointly to maximize the expected long-term system payoff—difference between the value of information and data processing costs. Since this problem is hard to solve directly due to hidden information about the correctness of processing outputs returned by EDs, we develop a machine learning (ML) framework to obtain the problem solution.