分布式数据库同步
我试图从概念层面理解数据库在很大程度上分布时如何同步。例如铁路/航空公司预订系统。每个市场领导者是否都有自己的专有解决方案来处理这种情况?
它们是否都在本地存在,并且每笔交易都根据主副本进行检查吗?
我曾经使用过本地存在的故障转移数据库,即主从场景,其中一个在另一个失败时接管;但是需要实时事务能力的数据库如何同步呢?
I am trying to understand at a conceptual level , how are databases synchronized when they are largely distributed. eg a rail/airline reservation system. Do each of the market leaders have their own proprietary solution to handle this scenario?
Are they all locally present and each transaction is checked against a master copy?
I have worked with fail over databases which are locally present, i.e. master-slave scenario where on takes over when the other fails; but how are database which require real time transaction ability synchronized?
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(2)
维基百科有一篇关于航空公司预订系统的好文章。相当多的参考资料。
Wikipedia has a good article on the Airline Reservation System. Quite a few references.
许多分布式应用程序不需要始终保持 100% 一致。以下是 Amazon 的 Werner Vogels 撰写的关于该主题的文章:
http://www.allthingsdistributed.com/ 2008/12/最终_一致.html
Lots of distributed applications don't need to be 100% consistent all the time. Here's an article by Amazon's Werner Vogels on the subject:
http://www.allthingsdistributed.com/2008/12/eventually_consistent.html