Trusted Cloud-Edge-End Resource Collaborative Scheduling Mechanism For IoT Services

With the development of the digital economy,the service model of the Internet of things(IoT)has gradually evolved from traditional data collection and business computing to providing ubiquitous IoT services through the collaboration of cloud,edge,and end resources.In providing the IoT service,the IoT heterogeneous resources formed by public or private cloud,edge and end are complex to guarantee the quality of the IoT service due to the distrust between multiple parties.Blockchain can solve the lack of trust between multiple parties through technologies such as consensus and smart contracts as distributed ledger technology.The integration of blockchain technology and IoT to realize the trusted sharing and collaborative scheduling of cloud-edge-end heterogeneous resources has become the development trend of IoT.With the integration of blockchain technology and IoT,multiple parties share resources in a peer-to-peer and mutual trust manner and provide IoT services through distributed collaborative scheduling.There are the following technical challenges:1)In terms of trusted sharing of heterogeneous resources,in the edge-end environment,computingintensive blockchain tasks such as consensus will compete with IoT applications for the resources of resource-constrained nodes,which affect the quality of the original IoT services,at the same time,in the distributed cloud-end environment,the centralized resource scheduling mode has the problem of malicious manipulation and opaque revenue distribution,which makes it challenging to guarantee the revenue of multiple parties in distributed services;2)In terms of heterogeneous resources collaborative scheduling,the peer-to-peer resource scheduling mechanism of the existing flat architecture is challenging to support trusted collaborative scheduling of resources under cloud-edge-end layered architecture.In addition,the traditional resource scheduling mechanisms ignore trust requirements and cannot be dynamically adjusted and need to be further optimized to meet the service quality requirements of large-scale distributed IoT.To solve the above technical challenges,we study the trusted cloudedge-end collaborative scheduling mechanism for IoT services from three aspects:edge-end,cloud-end,and cloud-edge-end.The specific innovation research contents are as follows:(1)Aiming at the problem that the competition of resources by blockchain tasks will affect the quality of IoT services in the IoT edge-end environment,we propose a distributed edge-end resource collaborative scheduling mechanism based on generalized Benders decomposition.We construct a distributed edge resource sharing collaboration scenario based on blockchain and form an optimization problem model of edge resource scheduling abstractly to maximize the overall revenue of resource nodes.Considering the partial constraints of the decision variables in this problem,a joint resource scheduling and task allocation algorithm is designed based on the generalized Benders decomposition.We iterate sub-problems to get the solution to the optimization problem.To improve the speed of resource scheduling decision-making,we design a hierarchical parallel resource scheduling and task allocation algorithm.Theoretical analysis and simulation experiments show that the proposed mechanism can effectively guarantee the execution performance of IoT applications while deploying the blockchain to achieve trusted collaborative sharing of edge resources.(2)Aiming at the problem that the existing centralized resource scheduling in the distributed cloud environment is challenging to guarantee the multi-party benefits of distributed services,we propose a distributed cloud alliance resource collaborative scheduling mechanism based on smart contracts.The mechanism can provide users with on-demand services while maximizing the resource providers’revenue.We first describe a blockchain-based distributed cloud alliance service scenario.Based on this scenario,the problem of matching multi-user requirements and distributed cloud services are abstracted into a dual-objective problem model with optimal overall trusted service quality and revenue.Then we propose a distributed cloud alliance service strategy based on smart contracts to solve the problem model.The service matching algorithm based on Kuhn-Munkres and the service matching algorithm based on ant colony optimization is designed under the guidance of Pareto optimality theory.The simulation results show that the proposed mechanism can realize the trusted on-demand matching of cloud services,enable cloud resource nodes to obtain better service benefits,and encourage nodes to share resources.(3)Aiming at the problem that the existing scheduling mechanism is challenging to support the trusted scheduling of resources with the layered cloud-edge-end architecture,we propose a trusted cloud-edge-end resources collaborative scheduling mechanism based on the Lagrangian multiplier method.The mechanism can realize hierarchical and differentiated blockchain deployment and ensure IoT service quality.We first build a blockchain-based scalable cloud-edge-end collaboration model and cloud-edge-end computing model.Based on these models,the collaborative scheduling of cloud,edge,and end resources is modeled as a problem model for minimizing system delay and energy consumption.Then a task allocation and resource scheduling strategy based on the Lagrange multiplier method is designed to obtain the optimal solution for this problem.The simulation results show that the proposed mechanism can effectively achieve credible and efficient collaborative scheduling of cloud-edge-end resources,optimize energy consumption and system delay,and ensure the quality of IoT service.(4)Aiming at the problem that the traditional resource scheduling mechanism ignores trust requirements and cannot be dynamically adjusted,we propose a cloud-edge-end resource collaborative scheduling mechanism based on deep reinforcement learning.The mechanism can provide fast,efficient,and reliable IoT services.We first propose a distributed intelligent resource collaborative architecture based on blockchain.Based on this architecture,the collaborative scheduling of cloud-edge-end resources is abstracted into a load balancing optimization model.Considering that the one-time decision of the problem model does not affect other decisions,we propose a dynamic resource collaborative scheduling strategy for load balancing.The dynamic resource collaborative scheduling algorithm based on deep reinforcement learning obtains the initial scheduling decision.The container scheduling algorithm for load balancing makes up for the initial scheduling decision and dynamically adjusts the status of IoT services.The simulation results show that the proposed mechanism can quickly update scheduling decisions,adjust network operation status,effectively maintain network load balance,and guarantee large-scale distributed IoT service quality.