Distributed anti-replication design for large websites

1. Internet distributed injury?

        During the evolution of large websites, horizontal segmentation and vertical segmentation appear. A complete business line is divided into multiple business points, and the high availability, high scalability, high concurrency, and scalability of the website are artificially increased and improved. However, this has derived many problems that need to be dealt with, such as: consistency problems, idempotency problems, real-time, transactional, etc., a series of problems have emerged, and these things have become big pits. How to fill this pit, a series of technical solutions and middleware also come out, such as: message middleware, zk, anti-duplication table, global lock, redis, RPC, etc. Let me give you a design of anti-duplication table.

 

2. The problem solved by the anti-weight table!

        As a distributed system, our system may receive the same MQ message from multiple nodes, or it may be dispatched by distributed scheduling, or there may be problems such as multi-thread concurrency, etc., it may be repeated orders, repeated payments, and repeated processing. And so on, with the anti-duplication table, we can effectively deal with repeated things, serve the idempotency of the system, and make the final state result of the system consistent.

 

3. Design an anti-weight service

 A little busy, waiting to write!

Guess you like

Origin http://10.200.1.11:23101/article/api/json?id=326793496&siteId=291194637