Method and system for updating caching server in website

Abstract

The invention provides a method and a system for updating a caching server in a website. The method comprises the following steps: a main database receives data updated by a user, and the caching server receives a plurality of concurrent active update requests sent by the user to the updated data and caches the data updated by the user in the plurality of concurrent active update requests; the main database synchronizes the received data updated by the user to a slave database; the slave database writes the data updated by the user and synchronized from the main database, and then the data updated by the user and cached by the caching server under the active update requests are invalid. According to the method, due to reasonable policy decision and the adopted manner of combined active updating caching and passive updating caching, the hit rate of caching is increased ingeniously, the response time is reduced, and meanwhile, the data processing problem in case of concurrent data is also solved by use of a corresponding policy; as a result, the data in the cache and the database can be finally consistent as much as possible.

Description

technical field [0001] The invention relates to Internet technology, in particular to a method and system for updating a cache server under a website architecture where a database and a cache structure coexist, so as to solve data consistency and concurrency. Background technique [0002] With the rapid growth of the amount of information and the rapid growth of the number of Internet users, how to reduce user response time and how to solve the problem of data concurrency are accompanied. The existing technology has provided us with several ways to solve the problem, using the technology of database master-slave separation and caching technology. When we apply these technologies, there will be problems at the same time. These problems mainly focus on data concurrency and data consistency. What this patent describes is also a relevant strategy to solve this type of problem. [0003] Firstly, memcache (hereinafter referred to as mc) and database separation technology in the...

AI Technical Summary

Problems solved by technology

[0008] The inventor found that the problem of using this method is: when using master-slave database separation + cache to perform data processing at the same time, there is a certain delay in the master database synchronizing data to each slave database (which cannot be completely avoided). When synchronizing to the slave database, when the user performs a query at this time, since the data of mc comes from the slave database, there is no data that the user wants in mc, which will bring delays in the slave database due to the delay of master-slave synchronization. Inconsistency with the data in mc

In this way, all the requests that should hit the mc will fall on the main database, causing the performance of the database to further decline, and then creating a vicious circle

[0009] (2) mc's active update process and the problems it brings

[0011] The inventor found that the problem caused by using this method is: concurrent processing of data, when multiple requests are required to update data at the same time, since the requests are completely concurrent at this time, the user's data may be overwritten or directly updated at this time. In the case of loss, the direct result is that even if the user updates the data, it may not be displayed on the website, and this loss will be permanently lost if the user does not repeat the operation, which will cause losses to the user

As a result, it is impossible to write all the data into the mc correctly, resulting in data loss

Images