Research On Active Data Processing In Mobile Internet

With the development of Mobile Internet,cloud computing and Mobile communication technology,the huge amount of data on Internet has increased with geometric level.The age of big data is coming.Therefore the scale and frame of distributed system has also changed a lot.The data on Internet is mostly expressed by Semi-structured XML,which has been the standard of message exchange and storage under the background of big data.As the basic of personalized intelligent service platform,the distributed computing requires both the great scalability and the high degree ability to adapt dynamic,heterogeneous,distributed Internet resource.The characteristics(multi to multi,dynamic plug and play,loosely coupled,asynchronous communication)of publish/subscribe systems are widely used in processing big data.In the leap-forward development of mobile Internet,the processing of complex structures real-time active data streaming is a hotspot.The traditional publish / subscribe systems lack for the capability of understanding and analysis of semantic level of subscribe requests,which lead to the matching results may contain a lot of irrelevant information and some uncertain factors in querying cannot be accurately judged,thus there are also some problems in expression,efficiency and scalability.For these reasons,the matching algorithm based on unranked tree automata in publish / subscribe systems is proposed in this paper.Firstly,active data is preprocessed and XML document similarity calculation method is proposed.On the base of the feature of data that the mobile terminal and proxy note receives as well as the semantic similarity under Internet background,the paper analyses and summarizes the problem in semantic similarity computation.The paper also proposes a new semantic similarity computing method,which is competent in computing the similarity between any XML document,classifying the data flow rapidly in the preprocessing stage of the publish / subscribe system matching,improving efficiency and accuracy of system.Secondly,this paper puts forward suffix tree automata algorithm on the base of semantic similarity calculation as preprocessing.The algorithm introduces the tree automata technology and suffix ideology.It accomplishes the inquiry of the complex structure that contains descendant axis and uncertain elements via bottom-up processing method of unranked tree automata.The method helps to reduce the intermediate state of the same transferring and avoid repeated calculation in the state set.Finally,compare the algorithm proposed in this paper with the classic publish subscribe system matching algorithm and obtain simulation results.Simulation results show: in the aspect of expressing ability,efficiency,the competence of processing uncertain factors and scalability,the algorithm proposed in this paper has more advantage than other classic ones.