Research On Subscribe/Distribute Algorithm For The Mass Data Of Air Logistics

Aviation Logistics is a rapid, safe and reliable mode of transport. In the process that the air cargo is delivered to the consignee, many participants are involved, like the consignor, the freight forwarder, the airport and airlines and others. Besides, there are a variety of cargo information packets exchanged between them, such as the delivery entrusted letter, cargo manifests and the like. How can the cargo information packets be sent to relevant interested participants is very important.In this paper, select the subscribe/distribute system as the model of the distribution of aviation logistics data, which has the advantages of spatial decoupling, time decoupling, and flow decoupling. Firstly, process the air logistics data which is in XML format, using a DOM-based method. And extract characteristic properties to provide base for subscription match. Then, Comb the type of aviation logistics data packets, analyze the relationships of data subscribe/distribute between different participants, propose a method for optimizing the storage of subscription conditions, establish the subscribe/distribute model for aviation logistics data. Secondly, for aviation logistics data distribution, a subscription matching algorithm is proposed based on Bloom filter, test results show that under the situation of increasing the number of subscriptions condition, the proposed algorithm has an excellent efficiency comparing with counting and matching tree method. Finally, improve the standard Bloom filter, reduce the running time of the mapping of multiple hash functions, and do the experiment to compare with matching tree method and standard Bloom filter method, analyze the influences of the number of subscriptions, the number of event attributes and the number of constraints. The results show that, comparing in these three areas, the improved Bloom filter algorithm has a better efficiency than the other two algorithms, especially in the case of large number of event attributes, the improved Bloom filter algorithm's time efficiency is even more pronounced.