| With the development of Network multimedia technology, Video Search Engine isbecoming a popular application. More and more people search the video they want byusing Video Search Engine. Retrieval function is an important part of Video SearchEngine. It consists of Multi-user concurrent request processing, show proper result tothe user; establish efficient caching and indexing mechanism, and analysis User log.This article analysis the main technical of Search Engine, and complete the design andrealization of the retrieval function of a small Video Search Engine.The design and implementation of this retrieval function includes concurrentrequest processing, index and cache, User log and feedback help.The retrieval functionof Video Search Engine in this article is based on text retrieval. In this way, the Engineretrieves the title of each video by using the search text of use. This is the mainapplication of Video Retrieval currently.Multi-user concurrent request processing isprocessed by Tomcat Cluster system. Using the combination of Apache Server and afew Tomcat instances, more computers can be added to the cluster. In this way, theperformance of concurrent can be improved. The Web application in Tomcat instanceuse the MVC framework, making it easy to realize and extend.This article design andimplement the caching and indexing mechanism in the search server, and the way ofestablishing, updating and searching on this basis. Cache uses the Hash Table to supportefficient retrieval, and update its words with the content of log. Index uses the InvertedIndex structure in most Search Engine application. The words is put in memory, and thedocument numbers of the web page which include these words are put in files on thehard drive, accessed through the file pointer.Through functional and performance tests, the retrieval function of Video SearchEngine designed in this article has reliability. It can support the index of large number ofvideo and concurrent multi-user query. |
PDF Full Text Request