Answer the question
In order to leave comments, you need to log in
Should I put a large site on node.js? Is not it too early?
Good day, friends!
Just the other day there were negotiations with large customers who want a high-load project. The site will be launched and promoted under the wing of the site with a traffic of more than 180 thousand people / day. Thought, decided to put on node.js, although on the other hand there is an alternative in the form of python. As far as I know, the node has not been run in yet, but the technology is awesome and clearly outperforms both php (stupid comparison) and python. It seems to be obvious that, in theory, the node is the clear future of server development, but, to be honest, we have not yet had such a project where you could really see all the advantages and disadvantages of the node. Therefore, the question arises: how meaningful is it now to put large projects on a node and is it worth waiting for something unpleasant? Thanks in advance!
Answer the question
In order to leave comments, you need to log in
Take python if you have no experience in node.js. Or erlang, since you still want to try something new for yourself.
At one time (2011, node.js 0.3.x) we suffered with callback-hell in the code, memory leaks and other hemorrhoids like "there is a library A, it seems to do what we need, let's fasten it", and after 2 months the author of either says "I lost interest, I will not cut further." And I had to sit down and fix other people's bugs ourselves.
Now, most likely the situation has changed for the better (I don’t follow), but what kind of highload do you want to get from a single-threaded application?) The node is good in a certain niche - as a full duplex data bus between clients and the server, if necessary, for example, a chat a- To file vk.com, or update currency quotes. Working with the database is convenient up to a point, calculations, filtering, data mapping are slow, any blocking call (data processing) - and 100500 clients are smoking bamboo ...
node is the clear future of server-side development
I worked on a large project with an audience of several million users per day, using nodejs. The flight is normal, the only negative is the frames. We need a strong nodejs programmer on the team.
On the same express (django for the node), very loaded sites, such as MySpace, work. Full list: expressjs.com/applications.html.
I would never write a project on a node as an auxiliary tool - perhaps.
It's not even a question of what is fast/slow, but that JS is the ugliest and most inferior server-side programming language imaginable.
Excuse me, do you have web sockets there or just sockets? Any nginx + php-fpm (python + wsgi something there, I just don’t know it) can receive a request and give a page. What will you benefit from Node.js ?
Can you implement business logic better? I doubt something. Therefore, choose from what you know better, and not from "it will pull better or not"
If you need a standard ORM, migrations, ACL, MVC, then definitely do not take node.js, take any platform that claims to be enterprise, Django, RoR, Symfony. If you need a complete enterprise, take Java Spring. Node.js is reliable only if you know how to program perfectly, understand all server interactions perfectly and will not use third-party libraries as much as possible like migrations or ORM, etc., you still need to write a lot on it yourself, or be ready to add any of the used libs, in case error detection in production. Personally, I would recommend the node for comet applications and as a layer between ZeroMQ / RabbitMQ and external services, the same SES, but for the regular web, use something more traditional and established.
Try playframework .
- It is more convenient to write asynchronous code on it than in a node;
- It grins well;
- Compiled to JVM bytecode;
- Static typing is better for large projects;
Of the minuses - compilation time.
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question