There is a lot of discussion about the performance of languages. I did a lot of research. Then, I worked at a giant bank with a hundred developers that serves 4 million operations every day. Let me tell you what I realized:
In one meeting, our manager asked about performance advice as usual. We were a mobile dev team but we were also maintaining a backend between the app and the core backend.
A developer threw an idea: Let’s move to GoLang.
We were using Java; an OOP-based, heavy language. Eager to RAM and thinks a lot to run the code. Lots of instructions in and out to CPUs.
Let's forget the learning curve and assume everyone in the company knows GoLang.
Since GoLang is optimized with every byte, it sounds like a good idea, right?
Does the backend project become lighter? Yes.
Can we cut some server costs? Yes.
Can we run the backend on our computer more easily? Yes.
Does GoLang seduce developers? Yes.
Should we spend a second to move to GoLang? No.
The real problem is not memory or CPU limitations. You can solve these problems with money, just another server and done.
If you can solve the problem with money, it's not worth spending developer time on it. Developers are always more expensive.
The real problem is that performance and scalability are always keeping, sharing, and managing the data on the disk. You must keep all the statuses the same, and consistent.
Reading data is fast. Some cache, a good CDN and we are good. Writing data is the bait. While you wring data, everyone is waiting. They may or may need the new data, does not matter, they must wait.
The real problem is never the computational performance of the language for the backend. It's always a database.
