TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

The evolution of a no-lock, no-wait write consolidation algorithm in WiredTiger

2 点作者 metheus将近 8 年前
Sue LoVerso of the WiredTiger team wrote an article in two parts about updating the lock-free algorithm the WiredTiger write-ahead log uses to consolidate writes. The original algorithm, based on the research paper &quot;Scalability of write-ahead logging on multicore and multisocket hardware&quot; published in the Very Large Databases Journal, was a very nice peice of engineering, but was inappropriate for the loads WT would see as a storage engine embedded in MongoDB. To handle a large increase in the number of threads per available core, it had to be updated to be entirely wait-free.<p>Part 1: https:&#x2F;&#x2F;engineering.mongodb.com&#x2F;post&#x2F;breaking-the-wiredtiger-logjam-the-write-ahead-log-1-2<p>Part 2: https:&#x2F;&#x2F;engineering.mongodb.com&#x2F;post&#x2F;breaking-the-wiredtiger-logjam-the-wait-free-solution-2-2

暂无评论

暂无评论