一种分布式环境中海量XML数据的有效查询机制.pdfVIP

第 28 卷第 4期 重 庆 交通 大学 学 报 (自然 科学 版) V o.l 28 N o. 4 2009 年 8月 JOURNA L O F CH ONGQ I NG JIAOTONG UN IVER SITY ( NATURAL SC I ENCE) A ug. 2009 XML 1 2 罗 静 , 陈 宁 ( 1. , 400074 2. , 4 01331) : XM L , , XM L CPU , XM L , , , XM L CPU , XM L : : TP399 : A : 1674 0696 ( 2009) 04 0807 06 Q uery P ro cessing for L arge XM L D ata in D istributive E nvironm en t 1 2 LU O J ing , CH EN N ing ( 1. Schoo l of A pp lication T echno logy, Chongqing J iaotong U n iversity, Chongq ing 4 00074, Ch ina 2. Chongqing U n iversity of Sc ience T echno logy, Chongqing 4 01331, China) A bstrac t: A mi ing at the qu ery for large XM L da ta, an e ff icien t d istr ibutive qu ery pro cessing m ethod is proposed, w h ich can realize the que ry through p artition ing and d istr ibu ting XM L data to mu ltip le compu tation nodes. Ev ery p artitioned XM L datum is distributed to the com puta tion nodes to ba lance the bear ing cap ab ility o f CPU. It is very mi portan t to take account o f the qu ery w ork loads, w h ich are c losely related to the qu ery processing cost at each computation node in d istr ibutive env i ronm ent. O r in the case o f load unb alance am ong computation nodes, the p artitioned XM L data should be re located to b a l ance the CPU load. T he exp ermi ent resu lt ind icates tha t th is m ethod can mi prov e the perfo rm ance o f query processing o f larg e XM L data in distr ibutive env ironm ent obv ious ly. K ey w ord s: Ex ten sible M arkup L angu age ( XM L ) qu ery pa rtition relocation