知识卡片
按实际瓶颈选择优化资源
CPU密集且少阻塞时,适当线程数即可;内存密集时要看缓存一致性、对象分配和GC;I/O加入后,等待可能主导CPU利用率,可用预取、批量或混合线程事件模型。计算器也可能受输入搬运限制,查询服务可能用CPU合并追加数据来节省I/O。业务名称不等于资源瓶颈,缓存和重算是在内存、计算与I/O之间交换成本。
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
- 服务端架构 · 查看原书依据
The main goal of the application and the real bottlenecks of the implementation can be different. Update and Lookup Services Most applications that look up data run as servers. Among examples are relational databases, NoSQL databases such as Cassandra and MongoDB, and both time series and graph databases.
参考来源
- 位置:《Software Architecture and Decision-Making》第11章《Server Architectures》(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/ch11.xhtml)
- 结论依据:摘录说明CPU密集且少阻塞时,适当线程数即可;内存密集时要看缓存一致性、对象分配和GC;I/O加入后,等待可能主导CPU利用率,可用预取、批量或混合线程事件模型。其余条件结合本章相邻例子与取舍说明,数字示例不作当前性能承诺。
- 原始内容:The main goal of the application and the real bottlenecks of the implementation can be different. Update and Lookup Services Most applications that look up data run as servers. Among examples are relational databases, NoSQL databases such as Cassandra and MongoDB, and both time series and graph databases.