Hello,
I'm just adding some info about what Justin said concerning the PDF.
So yes, you have the possibility to convert on the fly AFP into PDF with the AFP2WEB component (which contains lots of converter including AFP2PDF).
Then I've some experience with the PDF Indexer, and you need to know that you PDF will be cut in small pieces, and if some pieces are not strict Version PDF 1.3 or 1.4, then it will be discarded.
Meaning that if you have some PDF/A with certificates, some fields where you might have some custom indexes, or anything that makes use of some enhanced capability of PDF, then it will be removed.
The advantage of that, is that the PDF blocks that are stored in a way, that CMOD to display the document will automatically generated a PDF to the actual version. (meaning PDF 1.7 or 1.8, and maybe in 10 years as PDF 4.5)
The disadvantage is that you lose the integrity of your PDF, they won't be 1:1 archived. They will 100% the same displayed and printed, but all internal metadata that a PDF can have will be removed.
So if you want a 100% integrity, you must NOT use the PDF Indexer.
From a disk space perspective, I have done no testing between AFP and PDF, because I had no such material.
What I can say, is with the test files we had we have gain something like 70-80% of disk spaces.
Sincerely yours,
Alessandro