Link : https://www.zhihu.com/question/24723688/answer/160252724
In the computer world, reverse proxy
has a limit in the ability of a single server to process client (user) requests. When users' access requests flood in, it will cause the server to be overwhelmed. Multiple servers can be used. Sharing thousands of user requests together, these servers provide the same service, and for users, there is no difference at all.
The implementation of reverse proxy
1) requires a load balancing device to distribute user requests and distribute user requests to idle servers
2) The server returns its own service to the load balancing device
3) Load balancing returns the server's service to the user
The subtext of the above is: the user communicates directly with the load balancing device, which also means that when the user does the server domain name resolution, the IP obtained by the resolution is actually the IP of the load balancing, not the IP of the server . When removing a server, only the list of load-balanced servers needs to be modified without affecting existing services.
After talking about reverse proxy services, let's talk about proxy services commonly used by end users.
Proxy
1) The user wants the proxy server to help himself and communicate with the server to be accessed. In order to achieve this goal, the following work is required:
a) Destination IP of the user IP packet = proxy server IP
b) User packet port number = proxy server listening port number
c) The URL in the HTTP message should provide a link to the server
2) The proxy server can communicate directly with the server according to the link in c)
3) The server returns the web page
4) The proxy server packages the web page in 3) and returns it to the user.
Forward proxy proxy client, reverse proxy proxy server
