Research And Implementation On Distributed Resource Discovery Mechanism Under Restricted Conditions

Resource discovery is an important research topic in distributed systems,because the computations and services provided by each entity must be based on the premise that it can directly or indirectly discover and use the required resources.With the rapid development of edge computing and Io T,resource discovery faces new constraints and challenges in restricted scenarios.This thesis focuses on considering the impact of restricted environments on resource discovery technology,and proposes a three-tier architecture consisting of cloud-edge-device.By setting up an edge layer with resource catalogs deployed between cloud and terminal devices to solve the problem of limited Organization of nodes and communication latency issues with the cloud.Under this architecture,a resource discovery scheme adapted to constrained networks is proposed.The main work of the thesis is as follows:1.The construction of resource discovery problem model.According to the characteristics and limitations of the restricted environment,this thesis builds a model for resource discovery,and proposes a cloud-edge-device three-tier architecture suitable for edge computing and the Internet of Things.By deploying edge nodes at the network edge between the cloud layer and the terminal layer,the traditional cloud computing architecture is extended,and the edge nodes are used to build clusters with the communication range as the natural boundary.The architecture has key advantages such as real-time processing and fast scalability.2.The design of resource discovery mechanism.A resource discovery mechanism that can adapt to different network environments is designed.Different network quality is considered as poor network status and good network status,and different discovery strategies are selected by automatically detecting the network environment.The three-tier centralized resource strategy is mainly based on the publish and subscribe mechanism,and a priority-based two-level publish-subscribe algorithm is proposed.Nodes can also directly obtain global resource information through cloud nodes.The edge-edge heuristic resource discovery strategy realizes resource discovery without cloud through the cooperation of nodes in the edge layer,and uses particle swarm optimization(PSO)and distributed learning automata(DLA)to feedback the results of different query paths,to improve the accuracy and speed of the query.3.The design and implementation of resource discovery system.For the designed resource discovery mechanism,a distributed resource discovery model is implemented,which provides a solution to the problem of distributed resource discovery in restricted environments.Taking the edge layer as the core,it interacts with the terminal layer and the cloud layer,and realizes functions such as node management,resource management,resource discovery,resource subscription,and visual display.Finally,the effectiveness and feasibility of the proposed scheme are verified through simulation and testing of the performance and functions of the realized resource discovery model.