Node.js 是否适合构建大型文件共享 Web 应用程序？

Question

I'm a beginner programmer and I am taking an online course in Node.js and the instructor said not to use Node.js to build CPU-intensive apps. I don't know what is considered CPU-intensive. I am planning on building a large file sharing web app, but I don't know if Node.js is the right tool to get the job done.

Answer 1

Node.js has an event-driven architecture, meaning it has a built-in event loop and a asynchronous I/O API.

CPU-intensive tasks are those which use a lot of CPU computation in order to complete, such as finding prime numbers, sorting strings, et cetera.

When a CPU-intensive task takes place in Node.js it will block the event loop until its completion. During this time the program will be unable to process new incoming messages.

Conversely, I/O operations doesn't need that much of CPU usage.

Since generally speaking web apps and file sharing involves more I/O operations than CPU-intensive tasks (such as reading a file from disk and sending its contents through a network socket), Node.js can be well suited to create a large file sharing web app.

Answer 2

Is Node.js appropriate for building a large file sharing web app?

Yes.

Modern node.js has multiple ways to handle CPU-intensive tasks these days so that advice is perhaps a bit dated. The advice comes from the fact that it runs your main Javascript in a single thread and relies on asynchronous I/O to allow the single thread to very efficiently handle lots of requests using asynchronous I/O.

But, if some of your request handlers are using a lot of CPU cycles, then it bogs down the event loop.

But, modern node.js has the ability to:

1. Cluster it using the built-in cluster module so you have multiple node.js processes taking requests in a load balanced fashion that will involve many CPUs.

2. Use Worker Threads to start up individual threads where you can run CPU-intensive tasks and not block the main event loop.

3. Create some sort of work queue where you can use either additional processes or worker threads to CPU crunch on queued up work. I could imagine an image processing application doing it like this.

In your case, a file-sharing application probably doesn't even involve much CPU intensive stuff as most likely the main logic about who can access what is not CPU intensive and the act of serving shared resources is entirely I/O bound (something node.js is very efficient at). If, once you got into your development, you found a CPU-intensive bottleneck somewhere, you could very easily work around it with one of the above approaches.

And, there are many ways to horizontally scale node.js with clustering on a single server or multiple server clusters. If your app is I/O bound because of all the shared content, your scaling is probably mostly about network bandwidth and disk I/O scaling, not about the actual app logic or the CPU.