Number of files in root - performance

how is performance affected by the number of files

[beuten]

We are in the process of recreating out website. As part of the SEO side of things, we want to keep all the pages in the root directory, which means some 8000-9000+ static html pages. We are concerned that this will have a major affect on our performance.
IIS 6.0 running Coldfusion MX and MSSQL
Hardware to be determined (purchasing new)

Any comments on the performance with so many pages in a single directory?

Is there a way to have the URLs show up as root, but all the pages in subdirectories to divide it up?

Happy New Year to all!
Beuten

[BwanaZulia]

I know nothing about IIS... but....

I would think from the SEO and usability side of things it would be better to put things in directories.

This:
domain.com/keywordpage1.html
domain.com/keywordpage2.html
domain.com/keywordpage3.html
...
domain.com/keywordpage8000.html

or this:
domain.com/keyworddiretory/keywordpage1.html
domain.com/keyworddiretory2/keywordpage1.html
domain.com/keyworddiretory3/keywordpage1.html

Seems like the directory model would help with keywords.

BZ

[MichaelBluejay]

I don't know anything about performance, but I believe you're mistaken if you think you'll get some kind of SEO advantage by having all your files at the root level. I can't see why that would make any SEO difference at all.

[beuten]

everything that I have read until now has shown that root files have a higher ranking than subdirectories.

When looking at our keywords, the top 5 also have ALL their files in the root. this just backed up what I have read.

Am I understanding you correctly that you see absolutely no advantage to having everything in the root (for SEO purposes?)?

[dmorison]

Any comments on the performance with so many pages in a single directory?

There would be a hit, of course, because whatever indexing method NTFS uses to access files it WILL take longer the more files there are in a directory.

At 10,000 ish, my gut feel is that it would not be a problem as far as serving files is concernced, but your problems will come from other angles. The moment somebody uses Windows Explorer to navigate into your HTTP root directory they can go away and make a cup of tea whilst your server sits on 100% processor utilisation; likewise anybody trying to FTP into that directory remotely will have the same opportunity for a coffee break.

I think you will find that number of files in one directory difficult to manage anyway; however if you are worried about the SEO aspect you could consider other ways of making all your files "appear" to reside in the root directory.

Using the Apache web server, you could use MOD_REWRITE to convert, for example, /1_2_3_456.html into /1/2/3/456.html. I don't know of the IIS equivalent of MOD_REWRITE, but i'm sure it has been discussed all over the web.

[jdMorgan]

There is common confusion between "directory levels" and "click-path depth." As far as SEO goes, the major concern is "how many clicks are needed to navigate to a given page from the "home page" of the site -- in other words, how many links must be followed to get to that page.

You can have pages at any number of subdirectory levels, but if all those pages are linked directly from your home page, then the subdirectory level is not going to make any difference because the click-path length for any of those pages will be equal to one.

Directory depth is more important when you consider site usability and direct URL type-ins. As long as you keep your URL-paths short, then neither usability nor indexability will suffer by using a few subdirectory levels to help you keep all those pages organized.

Again, it is critical to understand that your URL system need not bear any direct relationship to your filesystem, and that subdirectory levels are usually quite different from click-path depth. dmorison touches on this by suggesting mod_rewrite on Apache or ISAPI Rewrite on IIS -- both of these modules clearly demonstrate that URLs and filenames need have no fixed relationship.

Jim

[MichaelBluejay]

everything that I have read until now has shown that root files have a higher ranking than subdirectories.

Well, that's one on me since in all the years I've been doing this I've never read that before (except from people who confuse directory structure with click-depth, as jdMorgan mentioned). In any event, I wouldn't believe everything you read.

When looking at our keywords, the top 5 also have ALL their files in the root. this just backed up what I have read.

I'm sorry, this is just really amusing that you draw that kind of conclusion. If all five had blue backgrounds would you think the blue backgrounds caused the top 5 rankings? Never draw conclusions from such a tiny sample.

Am I understanding you correctly that you see absolutely no advantage to having everything in the root (for SEO purposes?)?

Yes, that's exactly what I'm saying. I could be wrong, but that's my position until I see good evidence that suggests otherwise.

[beuten]

I'm sorry, this is just really amusing that you draw that kind of conclusion. If all five had blue backgrounds would you think the blue backgrounds caused the top 5 rankings? Never draw conclusions from such a tiny sample.

How did you know they also had blue backgrounds!? :-) Point understood.

It looks like we need to go back and look at the site structure some more before we move forward.
Thanks for all the helpful insight!