知识卡片
布隆过滤器先排除不存在
布隆过滤器用较小空间记忆大量键,查询说“不存在”可用于排除无需访问磁盘的请求;说“存在”则可能误报,仍需真实存储确认。因此它适合放在磁盘查询前减少无效I/O,不可把“可能存在”当数据已经查到,也不能代替存储读取。
相关:[[操作层级指导资源交换]]。
这张卡片看不懂?
这张卡片的反馈已满 10 条,暂时无法继续提交。
已有反馈 · 0 / 10 条
还没有人提交反馈。
在整书笔记中的位置
原书出处 · 1 处
- 理解性能的心智模型 · 查看原书依据
When queried, if the bloom filter says a key does not exist, it is always correct. When the bloom filter says yes, it does exist, it is only correct with a high probability. By fronting data in the disk with a bloom filter, we can reduce I/O by avoiding queries for things that are not in the store.
参考来源
- 位置:《Software Architecture and Decision-Making》第3章《理解性能的心智模型》(来源项目:book-sources;源文件:extracted/Software Architecture and Decision-Making Leveraging Leadership, Technology, and Product Management to Build Great Products… (Srinath Perera)/_epub-src/ch03.xhtml)
- 结论依据:原文的定义、比较或例子支持“布隆过滤器用较小空间记忆大量键,查询说“不存在”可用于排除无需访问磁盘的请求;说“存在”则可能误报,仍需真实存储确认”;正文据此说明用途和边界。
- 原始内容:When queried, if the bloom filter says a key does not exist, it is always correct. When the bloom filter says yes, it does exist, it is only correct with a high probability. By fronting data in the disk with a bloom filter, we can reduce I/O by avoiding queries for things that are not in the store.