Is node.js suitable for game backend development? _node.js

How are the website server and the game server connected?

1. There are many types of games. Let’s take a look at MMORPG first.

No matter how simple an RPG server is, it cannot avoid dealing with multi-person interactions. With hundreds of people in the same scene, each client needs to receive the operation information of everyone else.

Secondly, user operations are very frequent, and general servers tend to hold long connections. Moreover, these links interact frequently and there is no obvious persistent partitioning strategy, which limits the horizontal expansion of the server. The same scenario can often only be run on one physical machine.

Thirdly, client games usually don’t dare to put logical operations on the client. It is very common for users to crack it for you in minutes, change gold coins, and buy two pieces of equipment. Therefore, this map server has to verify the operations of all players in the map and calculate a series of business logic such as monster AI and drop rate.

We can see that there are obvious differences between traditional game servers and web servers, with unique business requirements such as long connections, multi-broadcast, complex business logic, and limited partitioning strategies.

2. Let’s take a look at the benefits of concurrency to game servers.

Concurrency is actually a program logic process, and it does not require multi-core physical support. The general meaning is to make multiple independent logic flows look like they are running at the same time. Operating system level concurrency is the multi-process multi-thread model. Let the OS handle clock interrupts, io blocking and other issues.

For the server, if the task spends most of its time on io, the concurrency mechanism can prevent the entire map service from being blocked by io access. When a task is blocked, allocate spare computing resources to other tasks. In this case, concurrency is beneficial to server efficiency and response time.

For programmers, independent logic flow means that they can complete their tasks in a reliable, simple, loosely coupled context.

Because the OS handler logic switching has to repeatedly fall into the kernel, some people think this is too slow, so they create some threads in user space and control multiple logic flows within the process. Because of the limitations of language description capabilities, it is too troublesome to write and use such things in C/C. As a result, coroutine syntax sugar in Erlang, Go, and Lua was born.

Node.js essentially controls multiple logical flows by itself, but this logical flow is distributed based on io status and priority. In the actual implementation, it tries to use non-blocking asynchronous io. When a single task calls io, I stop it, and when the io completion signal is sent, I restart it.

Note this, every time I run a task, I will not actively switch to other program flows until it is completed or an io call occurs. So if this task involves too much calculation, the entire map process will be blocked here.

Because node.js is asynchronous, you need to constantly write callbacks to monitor the io completion signal. The logical flow of a single task will be interrupted multiple times. When the task becomes quite complex, there is a so-called callbak hell, which will bring great trouble to debugging and development.

3. Because of the above reasons, I do not recommend using node.js in non-prototype MMORPG server development.

4. The recently emerged mobile game servers are quite suitable for node.js, because mobile games are limited to network problems. The server can only verify key data and cannot handle multi-person interactions. situation. The server side has been simplified to the point that it is no different from a web server, and the business logic is also simple. It only needs to process the data and then persist it.