Answer the question
In order to leave comments, you need to log in
Which CMS is suitable for high-load portals?
Good time of day habrosociety.
Recently I have been trying to figure out which CMS are most suitable for high-load portals
with the ability to go to the clouds, scaling and using nosql databases.
Thank you.
Answer the question
In order to leave comments, you need to log in
you yourself will later regret choosing a CMS, I would take the framework. And for nosql, you can look for modules ... or write it yourself.
And what is the task? Social network, news site or some kind of data processing service? In theory, a good CMS will allow you to hold out for as long as necessary until the load increases with various optimizations along the way. In this regard, Drupal's Pressflow fork looks good. If we take a framework, then it is initially necessary to understand which one is for what, to have developers and to lay on the highload (which is not yet?).
As for the clouds - do you have such a floating highload that the clouds will be more efficient than several rented servers?
In general, it looks more like a set of marketing cliches - highload, clouds, NoSQL. What else is trending right now? It should be displayed well on mobile devices... Although this is at least some benefit to the user, the user is neither cold nor hot from the method of storing information in the database.
Perhaps all this is suitable for your purposes, but then you should tell us why you need MapReduce over Joins, how you will launch new instances in the clouds under load, what caching strategies can be applied on your site, etc. .
I support the opinion about the rejection of ready-made CMS.
A ready-made CMS is, in essence, a universal tool aimed at mass application. Therefore, “everything” that may be required is laid there. And you, in a specific task, from this “everything”, as a rule, even 2/3 will not be needed. But this “everything” will devour resources in excess of the norm.
Correspondingly, as mentioned above: it’s more correct to take a framework (in my opinion, yii, if we are talking about php), as a database, in my opinion, mysql \ pgsql + sharding is quite suitable (since highload is planned).
And it’s not a fact that it will be easier than finishing / reshaping the finished CMS, and the result will be a full-fledged product in accordance with your requirements.
Any more or less popular CMS can be configured / sharpened for heavy loads. If there are more or less strong specialists in the team.
99% of programmers can only slander decisions and talk about how eggs interfere with their dancing.
If you hire such guys, then they will break firewood on pure php both on laravel and zend. And then they will persuade you to switch to NodeJS. If you give in, then it turns out that NodeJS is also garbage. And in order not to be given, they will crap it and find the reason, but not in their knowledge and their stupidity.
Of the confirmed and No. 1 in the WordPress world: wp.com - more than 10 million hosts per day, techcrunch.com - also something like this, w3bsit3-dns.com & lifechaker.ru about 1 million hosts, etc.
A little more expensive and worse than Drupal, but if there are specialists it will pull out.
Bitrix is even worse, but even it can be sharpened and taught.
Joomla - it's quite sad there. I agree it's not worth it.
Frameworks can be taken if there are at least a couple of competent specialists. But in Russia they are almost non-existent. One shkolota which lizhby hype. Well, all this will be 10 times more expensive than the same WordPress. Ptm up to 1 million hosts per day, it is better to go on a platform that rightfully occupies the 1st place in the world.
Article in the topic https://wpcraft.ru/2017/wordpress-highload/
Looking for a ready-made CMS so that you can then spend as much time on finishing it as on developing your own narrowly focused one ... I doubt that there is at least some profit in this. It makes sense to use CMS on small business card sites, in other cases, you can save a lot of nerves and write everything on frameworks (you can stock up on ready-made developments to do not from scratch).
And their frameworks should be addressed to Symfony2 or Zend2. There are no alternatives in terms of the number of ready-made high-quality solutions. In terms of NoSQL and integration into the clouds, everything is fine with them too.
I'm currently working on a project with a similar theme. There is not even a framework there (historically). Translating all this to a CMS is simply unrealistic, not to mention performance. Just use the framework. By the way, it’s not entirely clear why MongoDB was needed, you can safely do without it. SQL solutions are most often faster.
Write at least on Drupal, at least on Joomla. Any project changes at a frantic pace in the beginning. And I suspect that at this time you will not receive a million visitors. In general, this will be your "brake but easily changeable prototype." Once the requirements and product have stabilized, you will understand the resource and technology requirements. There you can and should switch to self-writing.
I would stop either at NodeJS (but there is still a lot of raw stuff here) or at proven solutions like Zend Framework (there are Zend / Cloud components for clouds), and MongoDB as storage. Doctrine 2 ODM can be used as a model layer. We use it this way.
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question