The Research On Publish/Subscribe Mechanism Of Aviation Logistics Data Packet Based On ILink

Air cargo involves a large number of participants in the transportation.Each participant needs frequent data exchange between each other to know the transportation state.At present,China has no unified aviation logistics data exchange platform,and even many companies use paper files to transfer data,but which has a low efficiency and a poor reliability.Therefore,it's extremely important to send efficiently and reliably the large amounts of air cargo data to participants which requires,and to provide one-stop,industry-class applications and services for aviation logistics participants and regulators.According to the business process of aviation logistics and the data packet exchanged between participants,an iLink-based aviation logistics data packet distribution/subscription system model is established.Then,a parallel matching algorithm based on B+ tree is proposed,the idea of which is to use the storage structure of B+ tree,to divide the subscription conditions according to the attribute value,to establish the multi-level index subscription attribute tree,to improve B+ tree query method for data packet matching,and to achieve rapid matching by combining with the idea of counting data.The experimental results show that the packet matching time of this algorithm is shorter,the subscription attribute tree is generated quicker,and the packet matching efficiency is higher compared to other algorithms in the large-scale distributed publish/subscribe system.In order to improve the service quality and throughput of the system,the prioritization method of data packet based on weighting coefficients is proposed.The method determines the data packet judgment matrix according to the ratio of two packets in the single factor,and calculates the data packet priority under the comprehensive elements by the judgment matrix.So the principle of "first-to-first-forwarding" is avoided,and the packet with higher priority is forwarded firstly.Meanwhile the application of data packet access to Kafka is designed,which defines Kafka's consumers,producers and the relationship between the two.Finally,the feasibility of the algorithm is verified by experiments.