There are in general two possibilities of handling stuff on Apache:
make configurations for folders one by one using htaccess file
to fully resign htaccess and put all the rules into the httpd.conf
Usage of htaccess is a load time issue. I want to know, how expensive in terms of load time is htaccess usage? Are there any tests?
The recommendation from the Apache project is:
In general, you should only use .htaccess files when you don't have access to the main server configuration file. ... a common misconception is that user authentication and mod_rewrite directives must go in
.htaccessfiles.
So please both set AllowOverride None and all your other directives in the main httpd.conf (and/or the subsections you Include)
When Apache is not configured with AllowOverride None you already occur a (minor) performance penalty, regardless of wether or not any .htaccess files are used.
This because for each and every request apache will need to check for the presence of a potential .htaccess file in every (sub-) directory leading down to the requested resource. For example when a file is requested out of a directory /www/htdocs/example, apache must look for the following files:
/.htaccess
/www/.htaccess
/www/htdocs/.htaccess
/www/htdocs/example/.htaccess
And so, for each file access out of that directory, there are 4 additional file-system accesses, even if none of those files are present. (If AllowOveride is set for /)
The system call used for that (man 2 stat) by itself isn't that expensive and typically the file-system cache is used rather than polling the actual disk, limiting the actual IO requirements, but still it can add up, as this article argues.
When one or more actual .htaccess files are actually present, apache still needs to open and read it (triggering another IO read operation and usually also an IO write operation to update the filesystem atime attribute) and parse it before the logic therein can be applied.
Unlike when your directives are in the main httpd.conf, which only needs apache to parse them once, at startup, each .htaccess file needs to be interpreted again for each and every request.
How expensive, in addition to the IO operations, parsing the .htaccess files is depends on their complexity.
That will take careful benchmarking to determine.
The expense depends upon how powerful the server is in particular it's storage I/O subsystem and whether there is any caching involved.
The way to test it is to configure your system with .htacess and load test it in the same way as you would any web server. Then configure the system without .htaccess and run the same tests.
Compare and contrast the results for your configured systems.
I don't believe there is a single right answer for your question. Are many variables which can affect htaccess load time:
more info:
I have tested with and without .htaccess and could really find no difference in speed or overhead whatsoever. Granted I only run 16 sites and it's a super fast server so in my case tracking milliseconds is almost impossible. On a server running hundreds of sites the overhead may be noticeable so as @Iain said test and see.
I personally do everything using the .conf file for a virtual host and not the .htaccess it gives me absolute control over everything, it's cleaner and it's loaded once into apache's memory. Even my wordpress rewrite rules are in my .conf files so it can never break.
