Design And Implementation Of High Concurrency Knowledge Subsystem Based On Optical Fiber Time Service Operation Control System

The rapid development of the Internet of Things technology brings about the emergence of a large number of Internet applications,and a large number of data streams are generated with a large number of physical devices.If the data stream processing is not timely enough,a lot of valuable information will be lost.This thesis models the equipment in the optical fiber timing operation and control system,and uses ontology language to describe its class,instance and relationship information.At the same time,the event is modeled,and the relationship between event attributes and knowledge attributes is established,and the flow of events is processed by means of logical reasoning,and the hidden complex events are discovered.The main work and contributions of this paper include the following four aspects:(1)Ontology description language modeling and knowledge management.Through analyzing the resource environment in the timing system,using ontology modeling tools to describe the information such as the class,instance,and attribute relationships in the system,and manage it in the form of ontology model files.The knowledge in the file is stored in the form of graphs to provide convenient query services.At the same time,it provides directory services to facilitate the hierarchical management of knowledge.(2)Management of events and goals.Model events in the timing system,including atomic events and complex events.Atomic events bind real events with knowledge,which facilitates the conversion of publish and subscribe message streams into event objects that reflect changes in the state of knowledge.A complex event is a collection of atomic events and includes conditional constraints on the group of atomic events.Conditional constraints represent the conditions that need to be met for the complex events to be discovered.(3)The realization of complex event engine.The complex event engine is responsible for extracting complex event definitions and subscribing to the topics contained in them when starting complex event monitoring,and introducing event messages from the timing system.And maintain the event queue and related time window or length window.When the window is satisfied,the target in the complex event definition and the event in the event queue are converted into logical expressions and submitted to the inference part for execution.(4)Design and implementation of parallel reasoning module.Due to the limited reasoning efficiency of a single machine,it is easy to become a bottleneck of the system.The parallel reasoning module decomposes the logical expressions converted from the complex event engine,and then performs reasoning calculations in parallel.The module divides the logical expression into different clause sets,which are distributed in different computing nodes for execution.At the same time,the satisfiability of shared variables between different nodes is verified,and the efficiency of inference calculation is improved.Finally,each module of the system was tested in detail,and the feasibility and efficiency improvement were verified.