← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 2313.e.5.f.11.f

渐进替换积累运行经验

理解等级:高级

整套重写往往一面追赶旧系统持续新增功能,一面缺少真实生产磨合,因而长期无法“准备好”。可先给存储等易变部分建立抽象,再逐个组件改进并逐步暴露流量,使改造可交付、可比较、可恢复。替换策略仍应由迁移风险与验证成本决定,不把所有重写一概判为错误。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • Chapter 16. Common Failure Patterns · 查看原书依据

      Rather than replacing the entire distributed system, focusing on improving individual microservices can make real progress tractable and significantly reduce the costs of running two systems in production at once. Similarly, it makes sense to invest in abstractions that separate the core business logic from code that may

    参考来源

    - 位置:《Designing Distributed Systems, 2nd Edition》第16章《Chapter 16. Common Failure Patterns》(来源项目 book-sources;源文件:extracted/Designing Distributed Systems, 2nd Edition/_epub-src/OEBPS/ch16.html) - 结论依据:原文功能追赶和生产就绪债务解释second system风险,并给出逐组件改进。 - 原始内容:Rather than replacing the entire distributed system, focusing on improving individual microservices can make real progress tractable and significantly reduce the costs of running two systems in production at once. Similarly, it makes sense to invest in abstractions that separate the core business logic from code that may