Cloud Computing Model Research Based On The Search Box/Resource Pool

In this paper facing e-commerce search engine for the application background. On the basis ofanalyses the centralized current engine system, sums up the advantages and disadvantages of theexisting system, in view of the defects and the insufficiency, Based on the search box／resource poolof cloud computing framework, using object-oriented methodology as the instruction, the key focuson the analysis of the cloud search engine design and implementation. This paper is focus on researchof cloud computing model to search engine, Furthermore, cloud computing model based on searchbox/resource pool is established. The main research efforts and achievements are spread out asfollows: the improved of traditional search engines function module, the searcher, index apparatus,inquires the machine parts are analyzed in detail, using distributed cloud computing technology with afull redesign. At the same time, the concept of the application of Map/Reduce programming model,the data calculation task is packaged into the Map function, and the data merge task is packaged intothe reduce function. The search engine has been improved to deploy on consisting of cheap machinesin a distributed environment. And improving significantly the search engine system precision rate,recall rate and the speed of response.First of all, this article is based on the search box／resource pool of cloud computing framework,a preliminary study has been done on the cloud computing definition and characteristics, search box,search engines, resource pool.etc. Then the key technology about cloud computing and typical cloudcomputing platform, parallel workflow and parallel programming model and the resource pool anddistributed file system also be discussed. As well as analysis the usage of the distributed file systemand Map/Reduce programming model in data processing of the specific application, at the same time,discussed the search engine related inverted document index mechanism, the Chinese wordsegmentation principle and the scope of application of the related tools as well, the preliminarydiscussion on the deserialize improvement of traditional search engine could also take place.Secondly, according to the search engine business functions of each subsystem. this paperdescribes the design of a distributed search engine. And the function design has been given to thesearch engine’s core three parts: searcher, index apparatus and finder as well as the various parts of theimplementation process, BPEL business process description language used as formalizationdescription, use Map/Reduce model to improve the three functional unit, so that they have the abilityof distributed processing. Thirdly, according to the function design and distributed search engine, index apparatus, finderparts of internal implementation process, combining BPEL business process description, given therelationship of the class realization diagram, the improvement of stand-alone operation mode of nutchsearch engine parts by cloud distributed, the implementation of the code programming, and discussionof the difficulties in the implementation process, and the methods to resolve as well.Finally, to form a development environment, implement of an electronic commerce cloud searchengine including the typical distributed search engine parts and experimental of grab data comparisonand analysis. Finally, the simulating experiments are conducted for establishing the feasibility andsuperiority of the proposed approaches.This paper proposed based on search engine／resource pool of cloud computing model hascertain theoretical significance and engineering practical value.