The Research On Publish/subscribe Algorithm Of Aviation Logistics Data Based On XML

With the rapid development of the aviation logistics industry,airlines,agents,airport cargo terminals,and customs departments need to exchange XML messages frequently.Existing publish/subscribe algorithms are no longer able to meet the growing demand of air logistics subscribers.Among them,the use of multi-core processors for parallel query processing of XML data has become an effective method.The existing parallel query algorithms are mainly divided into two categories: parallel query algorithm based on XML data partitioning and parallel query algorithm based on query partitioning.However,the parallel query algorithm based on XML data partitioning is not suitable for small XML messages in aviation logistics.The parallel query algorithm based on query partitioning needs to decompose the twig pattern query into multiple linear subqueries when processing the twig pattern query.These subqueries There is a correlation between them.Later,these subqueries need to be linked,so the performance of the algorithm is degraded,and thus there are limitations.This paper proposes a publish/subscribe algorithm for air logistics for ordered small XML messages.The algorithm consists of adaptive dynamic task allocation sub-algorithm and fast FiST sub-algorithm.The main idea is to firstly dispatch the Sequencing Twig Patterns query to each CPU core by using the adaptive dynamic task allocation sub-algorithm,and then use the fast FiST sub-algorithm to query the XML message.First,the adaptive dynamic task allocation sub-algorithm improves the multi-core CPU utilization by grouping the label sequences.The algorithm consists of parallel merging sorting algorithm and task group allocation algorithm.The main idea of the algorithm is to use the parallel merging sorting algorithm to sort the time-consuming array of tasks,and then use the task group allocation algorithm to sort the time-consuming arrays according to the ordered tasks.The tag sequences are grouped so that the tasks of each group of tag sequences are time-consumingly balanced,and finally the grouped tag sequence is scheduled to each CPU core.Experiments show that the adaptive dynamic task allocation sub-algorithm can effectively group the label sequences.Secondly,the fast FiST sub-algorithm filters the Sequencing Twig Patterns queries on each CPU core.The main idea of the fast FiST sub-algorithm is to split the query sequence according to the branch node($),share the same sub-sequence,and perform one-time matching on the shared sub-sequence to avoid continuous matching of the query sequence.Experiments show that the fast FiST sub-algorithm can effectively improve the query efficiency of Sequencing Twig Patterns.Finally,the experimental results show that the proposed publish/subscribe algorithm for aviation logistics can effectively improve the query efficiency of Sequencing Twig Patterns.