HBase在Facebook Message存储的使用经验总结

HBase数据存储状况 1、2PB+ of online data in HBase (6PB+ with replication; excludes backups),存储了message data, metadata, search index 等信息。 2、每天大概有8B+Messages,增长到每月大概产生250TB的数据。 3、Traffic to HBase ▪ 75+ Billion R+W ops/day  ▪ At peak: 1.5M ops/sec   ▪ ~ 55% Read vs. 45% Write ops 。 Facebook选择HBase的原因 ▪ High write throughput ▪ Good random read performance ▪ Horizontal scalability ▪ Automatic Failover ▪ Strong consistency ▪ Benefits of HDFS Fault tolerant, scalable, checksums, MapReduce internal dev & ops expertise Facebook的HBase架构 Multiple clusters/cells for messaging ▪ 20 servers/rack; 5 or more racks per cluster Controllers (master/Zookee