POEM+: Pricing Longer for Mobile Blockchain Computation Offloading with Edge Computing

Blockchain as an effective security solution has been applied in many mobile devices. But due to the limits of storage and computational capacities, it is hard for mobile devices to run blockchain applications locally. To solve this challenge, blockchain applications are offloaded to edge servers with mobile edge computing (MEC). However, most existing auction mechanisms on mobile blockchain fail to utilize parallel execution and the long term performance has not been well addressed. This paper investigates the mobile blockchain computation task offloading problem to improve total utility of auction participants. We propose an auction mechanism named POEM+ by solving an NP-hard multiple-choice multi-dimensional knapsack problem. POEM+ can offload tasks of one mobile user to edge servers on heterogeneous edge serves. We theoretically prove that POEM+ satisfies economic properties, such as individual rationality, budget balance, truthfulness and computation efficiency. Both simulation and real world experiments demonstrate that POEM+ achieves 130.6% higher utility than the existing heterogeneous tasks double auction WBD and 138.77% higher allocation efficiency averagely. Besides, the long-term performance of POEM+ can be guaranteed.