nginx load balancing

Question

1. When nginx does load balancing, does it mean that the number of concurrency supported by nginx used for forwarding by the front end is the maximum number of visits supported by the architecture??

2. If I have multiple server nodes and use nginx for load balancing, and nginx is on the same intranet as each application server, will all server nodes not use external network traffic, only the nginx server will use it? External network traffic.

My understanding is that nginx acts as a transfer station for all requests. When a request comes, it hands it to a specific server through the intranet. The server transfers the processing result to nginx, and nginx returns it to the client. ,

In this case, does all the network pressure fall on the nginx server? If all my services are deployed on Alibaba Cloud, do I have to use all the bandwidth on the nginx server?

3,

When nginx is used as a forward proxy, can it do the load?

In other words, nginx forward proxy can distribute requests to different machines like a reverse proxy, instead of one-to-one and can only go to one server

Answer 1

1. It can be roughly understood. However, nginx is generally not the bottleneck.

2. The server’s web services are all routed through nginx, and other traffic will still take the default route. Bandwidth pressure is on nginx, while computing and database pressure is still on the backend.

3. Don’t use nginx as a forward proxy.

Answer 2

The third question is discussed:

Can nginx, like lvs, serve as the client request entry on the front end, distribute the request to the back-end server, and the back-end server directly returns the response information to the client?

I learned that now nginx generally acts as a reverse proxy to distribute requests to the backend, and the backend responds to nginx, and nginx returns to the client.

Can nginx forward proxy only forward to one backend server? If it can be distributed to multiple backends, wouldn’t it be 2 times more efficient than a reverse proxy?