So, yeah, new thread. Sorry (I'm not sorry).

tl;dr = it's not perfect but you'll never get access 
control/revision/fulltext searching functionality even if you  spend  
~1000x more.

About using Google Drive....yeah, we're very small ( 115 students!), so 
we're very interested in keeping our over-heads nice and low..
I'm guess I'm old enough to think that 100 GB for $5 a month is a pretty 
good deal, so we started saying "Google Drive is our IR" as a joke, but 
like it's actually turned into a really nice "IR" type thingy.

  We just added a generic "library" user in our domain and bought extra 
drive space it. We try to organize things orderly by keeping things in 
various folders ( "Dissertations", "Articles", "UN Documents"), since it 
makes it easier to recursively apply ACLs. GDrive is like AWS in that 
the folders are not really "folders" like we're used to on a file 
system, but more like if you move a file around, it keeps it 
UUID (and therefore URL), which is pretty nice.

The best part is that since Google Apps uses OAuth, access control is 
really simple both in Google Apps and with external web apps.  We can 
make a document open to the world, grant access to groups/individuals, 
only allow access if they have the URL, etc. This works if they search 
in Google Drive or if they're tying to access a document embedded on 
another site.

The bad news is that there's not much (i.e. none) support in the way of 
descriptive metadata, which is kind of huge. To work around this, we 
currently either have descriptive metadata records kept in our ILS 
(Koha) or in our group Mendeley account. This adds a bit of complexity 
to managing the metadata and also means there's not a discovery 
interface that allows for both full-text searching (which google 
provides) and metadata searching (which Koha mostly provides).
I wrote an app last summer that indexes some of this content into a 
discovery interface, which I'm actually in the process of merging back 
into our Blacklight OPAC so we'll have a unified DS9DE ("Deep Space 9 
Discovery Environment")  . So, there's that...

And slightly less bad news, OCR and the document viewer only supports 
files < 20MB. We're have a lot of very large PDFs, so it's a bit of a 
drag, but the students just have to download the PDF, so it's not so bad.

And the Google Drive desktop client can be buggy and crashes if you try 
and sync large collection. And you still have to figure out preservation 
(or not).

But yeah, despite all that BS it's been pretty great. And since Google 
gives the CIA unlimited warrantless access, I assume that someone out 
there (i.e. DC metro area)  is reading our content.

Any questions, please feel free to ask me...