Things about caching in software design

​ Cache is the simplest and most complex concept , With it, you can make your application fly , I won't use it , Even the highest configuration , Your application is also difficult to achieve the best results .

What is caching

Cache is the direct access of data by hardware or software , Not from the original source （ database , Calculation, etc ） Access data , It will speed up the reading of data .

Memory caching is a caching technology we often use , By saving commonly used data into memory , We can avoid reading data from the database , This will greatly improve the loading speed of the application , Enhance user experience .

When to use caching

• When we need to request external resources , For example, the third party api, We want to cache the result of the request
• When the requested data changes , We can cache the returned results
• When an application needs to request the same resource frequently
• When applications need to run some complex computers
• When applying queries takes time

Cache benefits

• Improve application performance
• Reduce database overhead
• Increase concurrency

The type of cache

• Application cache
• Database cache
• DNS cache
• Client cache
• CDN cache
• API Gateway cache

Distributed cache

Distributed cache is managed and controlled by a server , There are multiple client nodes storing data , It can further improve the data reading rate . So when we want to read some data , Which node should I choose ？ If you search node by node , That's too inefficient . Therefore, it is necessary to determine the storage and reading nodes of data according to the consistent hash algorithm . With data D, Total number of nodes N Based on , The data is calculated by the consistent hash algorithm D Corresponding hash value （ Equivalent to the house number ）, According to the hash value, you can find the corresponding node . The advantage of consistent hash algorithm is that the number of nodes changes （ Reduce or increase ） There is no need to recalculate the hash value , Ensure that data can be stored or read correctly 、 Quickly find the corresponding node .

Distributed cache can read data with high performance 、 It can dynamically expand cache nodes 、 It can automatically find and switch fault nodes 、 Can automatically balance data partition , And it can provide users with graphical management interface , It's easy to deploy and maintain .

Cache policy

(1) Cache-Aside

Read the cache first , If the cache fails , Then read the database to get data , Then set the data into the cache .

(2) Read Through

Applications only fetch data from the cache , If you don't get , Then the cache will get data from the database , Then set the cache , Finally, the application continues to read data from the cache .

(3) Write Through

When the application saves data , They are stored in the cache first , The cache will save the data to the database , Applications do not deal directly with databases , Just save the data to the cache .

(44) Write Back or Write Behind

The application writes data to the cache , Then it will send a confirmation back to the application , Then the cache is written to the database .

(5) Write Around

Application data is written directly to the database , So as to read the data and get it from the cache .

summary

A great man once said , All the computer problems we can't solve at present , Can be solved by adding an intermediate layer , The cache is an intermediate layer , With this middle layer , Our application performance has been greatly improved , But caching is not everything , It also has its drawbacks , Only use it reasonably , We can give full play to its maximum value .