0

When you host a PDF on your own website and server, and then link to that PDF in an article, does the size of the PDF file matter? Or is the size only a factor when clicked, because of the time it takes a client to download it?

Secondly, what is an appropriate file size to request from a client for a business/technical PDF document, with images, from 10 - 50 pages, hosted in this context?

The background

Note: in the future, we'll have guidelines for clients, but this is the first time it's been an issue.

I've been asked to upload a 7MB PDF file (44 pages, lots of images & graphs) for a client (a very big multinational company) to go underneath a few paragraphs saying 'read the full whitepaper' or similar.

I've been told by our designer that they have created it on Photoshop and the text is not 'text' it's actually more like 44 photographs.

We can't ask the client to change it for various reasons, one of which is that we've already asked them to bend once on this and since we're not a big multinational company, I wouldn't want to be presumptuous.

We run Drupal and the default is 2MB uploads. We'd be hosting the files ourselves (though in the future I'm sure we'll improve our file structures, possibly over different servers). I'm hesitant to 'just increase' the file size upload limits, though it can be done. And I'll probably circumvent this and upload it on this occasion.

Are there pros/cons for large PDFs in this context? My assumption is that because it is just a link to a file there is not much HTTP request overhead unless the user clinks on the download link. So I assume that here, having a large file is not a problem.

I've done some reading and this is quite a good post on maximum PDF files for public websites, which advises that the more specialist the subject matter (and ours is specialist), the bigger the file size users will accept. It was written in 2011 though, and based on the graph on this answer the size of what we ask users to download could be twice as much in 2015.

  • Are you asking in context to user experience (download time) or SEO best practices (ranking implications from page asset size)? In regards to max file size, i don't think there is a mod for this, and this is prob not realistic, but perhaps if the user is admin-ish the controller for the node edit area could temporarily set max file size to a larger number. As long as there was a master cap in your PDF serv as well as in Drupal, and lesser users (or pub front end) didn't have file upload facets, it may work to briefly set cap higher. Or, it may not since that sounds slightly insane haha. – dhaupin Aug 26 '15 at 14:50
  • Thanks @dhaupin - interesting thinking but out of our means atm, to put it lightly :) I'm thinking with regards to both your points and more: UX of loading the page that contains the link to the PDF, and the downloading of the PDF itself (which ultimately I assume is just time); and SEO best practices with what you described as page asset size (will be googling that now...) Also, since our files and Drupal all run from the same apache instance I'm also thinking about the server's load and how adverserly affected it might be from this sort of thing (particularly if it was more commonplace) – ja_him Aug 26 '15 at 15:27
0

If its possible, try redoing your PDF files so they are smaller in size.

Smaller files can be downloaded faster than larger files and the server will then be ready sooner for more users.

In a server setup, there is a limit on the number of processes that can be handled by the server at any one time and in apache, this is defined by MaxClients value.

Each request to download a PDF file counts as a process and the process is in use until the PDF download is complete plus at least a few seconds to clear the connection entirely.

So if your server is set up to handle up to 50 processes, and 100 people want to download the 7 MB PDF file and it takes 1 minute to download 1 MB then the first 50 people can download the file right away but the remaining 50 people have to wait at least 7 minutes if a queue is setup (apache calls this the listen backlog) or failing that, those 50 people will be denied access temporarily. As processes are freed, more users can then be accepted. That's why your server can be slow.

So if you can't make the download size smaller, then consider looking at your server settings and increasing the number of clients it can handle. Also, consider adding more memory or upgrading the server itself.

  • Thats a great explanation @Mike, thank - really clears a lot up. Can I ask - does the size of the PDF matter at all simply as an asset on the page, when someone visits the page, but hasn't yet clicked the link to download the file? Would that impact user experience? Or would there have to be an element of pre-loading before that becomes a factor for a client? – ja_him Sep 1 '15 at 9:27
  • Since many internet users have high speed connections, advertising a file size for a PDF isn't an issue unless you decide making one that is several dozen gigs in size. No action will be done on the PDF file that is already stored on the server without user initiation which usually is a click to the PDF from the website by the user. – Mike Sep 1 '15 at 23:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.