Web search for a planet:: The Google cluster architecture

被引：428

作者：

Barroso, LA

Dean, J

Hölzle, U

机构：

[1] Mountain View, CA 94043

来源：

IEEE MICRO | 2003年 / 23卷 / 02期

关键词：

D O I：

10.1109/MM.2003.1196112

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

AMENABLE TO EXTENSIVE PARALLELIZATION, GOOGLE'S WEB SEARCH APPLICATION LETS DIFFERENT QUERIES RUN ON DIFFERENT PROCESSORS AND, BY PARTITIONING THE OVERALL INDEX, ALSO LETS A SINGLE QUERY USE MULTIPLE PROCESSORS. TO HANDLE THIS WORKLOAD, GOOGLE'S ARCHITECTURE FEATURES CLUSTERS OF MORE THAN 15,000 COMMODITY-CLASS PCs WITH FAULT-TOLERANT SOFTWARE. THIS ARCHITECTURE ACHIEVES SUPERIOR PERFORMANCE AT A FRACTION OF THE COST OF A SYSTEM BUILT FROM FEWER, BUT MORE EXPENSIVE, HIGH-END SERVERS.

引用

页码：22 / 28

页数：7

共 5 条

[1] [Anonymous], 2002, INTEL TECHNOLOGY J
[2] [Anonymous], 1997, COMPUTER
[3] [Anonymous], 1998, Proceedings of the 7th international conference on World Wide Web (WWW), DOI [10.1016/S0169-7552(98)00110-X, DOI 10.1016/S0169-7552(98)00110-X]
[4] Memory system characterization of commercial workloads
Barroso, LA
Gharachorloo, K
Bugnion, E
[J]. 25TH ANNUAL INTERNATIONAL SYMPOSIUM ON COMPUTER ARCHITECTURE, PROCEEDINGS, 1998, : 3 - 14
[5] Barroso LA, 2000, PROCEEDING OF THE 27TH INTERNATIONAL SYMPOSIUM ON COMPUTER ARCHITECTURE, P282, DOI [10.1109/ISCA.2000.854398, 10.1145/342001.339696]

← 1 →