Large-scale Incremental Processing Using Distributed Transactions and Notifications

[hawkingrei]

https://storage.googleapis.com/pub-tools-public-publication-data/pdf/36726.pdf

[hawkingrei]

这篇paper主要介绍了google Percolator在分布式事务和增量通知功能。其建立在Bigtable之上，来为google网页搜索提供索引。

[hawkingrei]

Percolator事务实现

Percolator 使用两阶段提交的方式来实现

1、prewrite阶段

• 获取start timestamp
• 检查是否有冲突发生
• if the transaction sees another write record after its start timestamp, it aborts; this is the write-write conflict that snapshot isolation guards against.
• 如果发现没有冲突，则在lock和data中写入timestamp

2、commit阶段

• 从timestamp ocrale 获取 start timestamp
• the client releases its lock and make its write visible to readers by replacing the lock with a write record. The write record indicates to read- ers that committed data exists in this cell; it contains a pointer to the start timestamp where readers can find the actual data. Once the primary’s write is visible (line 58), the transaction must commit since it has made a write visible to readers.