Computational task offloading based on edge computing can deal with the performance bottleneck of traditional cloud-based systems for Internet of things (IoT). To further optimize computing efficiency and resource allocation, collaborative offloading has been put forward to enable the offloading from edge devices to IoT terminal devices. However, there still lack incentive mechanisms to encourage participants to take over tasks from others. To counter this situation, this paper proposes a distributed computational resource trading strategy addressing multiple preferences of IoT users. Unlike most existing works, the objective of our trading strategy comprehensively considers different satisfaction degrees with task delay, energy consumption, price, and user reputation of both requesters and collaborators. The system design uses blockchain to enhance the decentralization, security, and automation. Compared with the trading method based on classical double auction matching mechanism, our trading strategy has more tasks offloaded and executed, and the trading results are friendlier to collaborators with good reputation.
翻译:以边缘计算为基础的卸载计算任务可以处理传统以云为基础的物联网系统(IoT)的性能瓶颈问题。为了进一步优化计算效率和资源分配,开展了协作卸载工作,以便能够从边设备卸载到IoT终端设备。然而,仍然缺乏激励机制鼓励参与者从其他人手中接过任务。为了应对这种情况,本文件提出了针对IoT用户多重偏好的分布式计算资源交易战略。与大多数现有工作不同,我们贸易战略的目标全面审视了任务延迟、能源消耗、价格和用户声誉方面的不同满意度。系统设计利用块链加强分散、安全和自动化。与基于传统的双拍卖匹配机制的贸易方法相比,我们的贸易战略有更多的任务被卸载和执行,而贸易结果对声誉良好的合作者更为友好。