--// 这一章 ,还是有很多干货儿:
-- Network state: TCP ==connection-oriented(用在FTP的2端连接) ,和 UDP的nun-connection-oriented (用在 例如 对DNS-->IP的单项翻译,即便翻译成功了,不需要收到服务器端的response 等等)
除了文章中 大量的术语 ,定义的干货(我以前 都不理解哦,小白,太懒了 ==惭愧)
还有 对 “为什么 今天有这些技术变化 ,(背后是 为什么今天 的用户行为,和 用户 场景 有了这些变化的分析)
-- // 以前 web上的一个application 是 事件驱动的(类似 鼠标 click),(软件上 stateful ,就支持一个客户点击,假设的也是 客户点击后,客户的场景 是在那里 傻等很长时间,然后 这个 另一端的服务器 也是 stateful 痴痴 但也 迟迟的 做出反应)
然后 服务器对 application的性能的支持,就是主要 依靠 硬件 (专门的服务器的硬件 ,相应 200 个鼠标 click event 构成的事件)
就像我 在 *.conf 中 ,设置的 thread =200 ,一样; 关键 靠 硬件上的强大和支持
-- // 但现在的用户 ,是移动的使用 网路,而且是 不需要 state,用户也等不了 你家服务器的缓慢和 犹豫
今天的客户,不等你家的服务器,
要的是 : ”所要就所得的.... " ,我 要,现在立刻就要 ; 而且 只是 在使用的这段时间 保持对服务器application的需要,用之前可以不安装,用之中不傻傻等待,用之后也可以无情的用后即删除!
用的过程中,你家后台的服务器 既要做到 实现客户的application需求,同时也具备你自己家的服务器可以边提供服务边升级的超人/分身的本领 (而且 你家服务器 提供的多个功能,还要 各不干扰 ,各自升级/完善自己)
-- //包括 在 服务器 可以一边 提供application /一边 自己升级/ 一边 有多个各自为政的多个功能的前提下, 为什么 要发展 JSon 的理由也就 呼之欲出了== 各家各玩自己的,微服务/API/各行其道,数据就作为一种 公共的媒介资源,就必须的跨界(不分 html/ python/ utc )的自由流通了, 就tmd 和 bitcorn 一样,是 大势所驱.
As organizations decouple applications from databases, options become available to better manage data and availability. Migrating data from a relational database into a NoSQL database to decrease customer query times is a good move. Product queries are made faster by searching denormalized data in a NoSQL database; when the interaction moves to purchase, the relational database can be used to record the transaction.
从速度上对比: 查询的事 用MongoDB,更快==因为主要是用“读”功能; 更新和写入的事 (最典型的 就是 trasaction 事务性的记录,一笔都不能差错,但是又不想log那样,需要立刻/反复的调用和查看的事 ,就多用 Mysql DB.
Stateful and stateless programming can be defined as software code that main-
tains a state or data element, or sees each interaction without previous con-
text. Stateful programming is the dominate sibling because any time a variable is
set ( i =1 ), a data element is captured (for example, capturing customer order
information on an e-commerce site), a variable in a loop is incremented, or
an array is used, a state is present. In and of itself, stateful is not a problem
because state is needed for many transactional interactions. DevOps does
not mandate that all code must become stateless, but there are times when
stateless brings opportunity.
Stateless programming is growing as containerization, microservices, and
DevOps are growing because code segments need to execute in parallel with-
out dependency on another code segment. Because code can execute without
dependency, it becomes easier to update the code without impacting other
pieces of the system. If each container is autonomous, then creation, change,
or the destruction of it or surrounding containers has zero effect on the rest.
公司 (用服务器的公司) 为什么 要被督促着 ,做 :
simplification and application segmentation ?
ASW:
maller, Agile-developed, and uncomplicated applications provide reason-
able opportunities for live implementations , stemming from the initial
database design consideration for continuous availability. Application seg-
mentation—API calls, web services, containerization, and microservices—
sanctions application updates to be done in real time
Reaching the same point for making database changes “live” requires
rethinking database availability, the database model, and (most importantly)
the database. Whether graph or document, a database may be a better fit
than a relational database, so determining the best database for the job is
important for availability.
---------------------------------------------------------------