How to design a high-throughput architecture with high concurrency

High concurrency and high throughput are contradictory. No locks are needed without locks. When pessimistic locks or optimistic locks are used, locks are only required after the query results are processed. At this time, transactions can often be supported. If granularity is required Smaller and can be used without transactions. But once it is locked, it is necessary to consider whether other queries are allowed to execute.

Guess you like

Origin http://43.154.161.224:23101/article/api/json?id=326309372&siteId=291194637