In a production environment, it is better to have a more powerful server or multiple servers with relatively low performance.

if server downtime is excluded, there are other servers available at the same time. In terms of performance, is it better to use a machine with more memory, more processes, and stronger CPU, or is it better to use multiple servers with relatively low performance?

for example, which method has stronger processing power and faster response in the case of server processing performance, especially in the case of high concurrency, between a server with 16 GB memory with 8 processes and two servers with 4 processes with 8 GB memory?

Golang python java php node.js

Apr.08,2022

currently there are multiple servers, and there are some restrictions on a single server even without downtime, such as the number of ports.
therefore, multiple servers are recommended. In a high concurrency environment, if one fails, the load balancer will distribute the traffic to other servers to ensure that the service runs.
one server hangs up

depends on the application, but there are usually multiple servers, because no matter how single server is, there are IO restrictions and so on.
and multiple machines can avoid the single point problem of single machine failure.

still need to be analyzed on a case-by-case basis. Even if you put aside a single point of failure, there are still many factors to consider, such as data sharing. Session persistence? Wait

if the degree of high concurrency is general, and the number of users is not particularly large, using a machine is certainly good, convenient and easy to operate.

if it is really large concurrency, do you want to use a supercomputer?

Previous: I started debugging with the weex-devtool command, and the page that appeared

Next: How to find the intersection and mark two multi-dimensional arrays