INCENTIVE-VACATION QUEUEING FOR EDGE CROWD COMPUTING

Abstract

Edge computing is designed to bring services closer to end users, significantly improving latency and scalability. However, beyond the conventional edge lies an untapped resource the extreme edge. Extreme edge computing (XEC) leverages computational power at the user’s immediate environment, offering new opportunities for efficient resource utilization. Edge crowd computing (ECC) takes advantage of this by creating a sharing economy model where users can rent out idle resources on their devices in exchange for compensation. This paper explores an orchestrated ECC model that incentivizes users to provide their devices' computational power. We introduce the incentive-vacation queueing (IVQ) model, which links performance to incentive payments using a vacation queueing mechanism, factoring in device multitenancy and the dynamic nature of server availability tied to incentive levels. This framework allows for the analysis of any sharing economy system modeled by IVQ, providing insights into how incentives and server "vacation" periods influence performance. We investigate two types of IVQ functions that can be tailored to either benefit the orchestrator or the device owners and propose a performance metric to evaluate these trade-offs. Our analytical expressions account for the stochastic availability of user devices, influenced by changing incentives, and offer a comprehensive approach for evaluating system performance in edge crowd computing environments.