ASP. NET CORE Study08

Get New knowledge :

cache

Relevant concepts ：

The type of cache ：

summary ：
in general , Private caching reduces the need for network bandwidth , At the same time, it will reduce the cache to API Request . Because the private cache is stored in the client browser , For requests , If the cache is still in a limited period , Then the request will not be sent even the network request, and the response will be directly obtained in the client browser , This reduces the number of network requests , It will also reduce API Number of requests .
Shared caching does not save caching to API Network bandwidth of , But it reduces requests to API Request . Because the shared cache is stored on the server side , The request must reach the server first , This will not reduce network bandwidth , But it will be reduced to API Request , Because subsequent requests are handled by the cache , Not every time API perform .

therefore , In general , Private cache and shared cache should be used together , But caching is more used for static resources , Such as images , Content pages, etc , And for those that change frequently API It's not appropriate , Such as adding or updating , The rest of the clients should get the latest resources , But if it's a cache , The remaining clients can only get the old resource information within the cache period .
This requires validating the model .

Cache usage

Expiration Model ：
The expiration model sets how long the response information can be kept “ fresh ” To keep whether the cache is expired , adopt Cache-Control Request headers to set whether the cache expires .

When the expired cache is working ,
The first request will be cached first , But at this time, the cache is empty , So the request goes to API,API Will return the response to the cache , At the same time, the response header will include Cache-Control Response head , Indicate cache lifetime, cache type, etc , The cache will copy the response into the cache , And return the response , Requests for subsequent entry will not enter as long as they are in the fresh period API Inside , Instead, the cache returns the response directly .
Validate the model ：
Used to verify whether the cached response data is kept up to date .
When the cached data will become the response of the client request , It first checks the source server or the intermediate cache with the latest data , See if the data it caches is still up to date .

In general ：

Cache-Control Commonly used instructions

Cache-Conrtol Respond to common instructions ：

Cache-Control Request common instructions ：

ETag Use

A third party NuGet package Marvin.Cache.Headers Provides a very complete and powerful Cache management function .
install Marvin.Cache.Headers package , stay startup Of configureservice Register for services , stay configure Method to use the cache middle price . Last , stay controller Go up or action On , Carry out property annotation .
The package provides complete functionality .

concurrency control

Concurrency issues , It can mean Two clients ,1 and 2, Customer 1 Get a resource first , Subsequent customers 2 Also obtained this resource , And customers 2 The update operation is performed and pushed to the server , At this point, the customer 1 Also update the resource , So customers 1 The change will bring the customer 2 Full coverage of , This is concurrency control .
Pessimistic Concurrency
Resources are locked for the previous customer , As long as the resource is locked , No one else can modify it , Only customers 1 You can modify , But this REST Can not be achieved under , because REST There are no state constraints .
Optimistic concurrency control
This means that the customer 1 You'll get one Token, And allow him to update resources , as long as Token Is reasonable and effective , So customers 1 You can always update the resource . stay REST Can be realized in , And this Token It's just a verifier , And it is required to be a strong verifier , So you can use ETag.

Optimistic concurrency control Example

figure ：