Host own data

Hi,
I've searched the documentation and the forum without finding an answer but is it possible to host the data you collect using Zotera on your own site/server in some way?
  • Could you clarify a little what you're trying to do and why? In general, the answer is yes, but the level of technical difficulty ranges from very easy to very involved depending on what exactly you want to do.
  • edited June 12, 2014
    I collect references along with my colleagues but I want the collections of references I collect to be stored on a server within my network, not in the cloud/server at Zotero. Setting up access rights on my data at Zotero is not what I mean.

    There could be security reasons for it or even technical issues when you work at a company that restrict you from using cloud services so that's the background for it.

    Does that make it somewhat clearer?
  • Unfortunately that's the "very involved" case. I think the most complete set of instructions is here: https://github.com/sualk/dataserver/wiki/Installation-Instructions-for-Debian-Wheezy You may find the following discussion helpful: https://groups.google.com/d/msg/zotero-dev/yEuHw4o4VWs/caiA1R3hnpcJ There is not much support offered for installing/maintaining/troubleshooting local Zotero servers so you're mostly on you own. You're welcome to post on zotero-dev mailing list and some community members or maybe even main developers may respond.

    Good luck!
  • Alright, thanks for the quick response.
  • edited June 12, 2014
    Somewhat. The details really do matter here, so I wish you'd be more specific about what exactly _your_ concerns are, not what hypothetical concerns in general are. I'm aware of people in all types of different situations—ranging from economic to privacy to network issues to DoD contracts—that may have very different solutions.

    But if you do really want to host a Zotero server locally that is possible—the server code is open and available:
    https://github.com/zotero/dataserver
    but it's a significant undertaking, requires patched versions of the Zotero software as well and will receive no support from Zotero's core developers.
    If you look on zotero-dev https://groups.google.com/forum/?fromgroups=#!forum/zotero-dev you'll find recent discussion among people who have successfully done that. There are a couple of gits with their set-up, but I don't currently know which the most up-to-date one is, you'd have to ask. But have a look at the discussion first to see if this is at a technical level you feel comfortable.

    (edit: crossposted with aurimas above)
  • I collect references along with my colleagues but I want the collections of references I collect to be stored on a server within my network, not in the cloud/server at Zotero.
    I've asked this of others, so I'll ask you too, mostly out of curiosity: would the option of end-to-end encryption in Zotero make a difference here, or would you still need/want to use a local server?
  • I collect references along with my colleagues but I want the collections of references I collect to be stored on a server within my network
    Important note that I forgot... you will not be able to use both zotero.org and your own server using the same client (you could have two different Zotero versions with separate profiles installed though). Your collaborators would have to use the patched Zotero client as well.
  • I've asked this of others, so I'll ask you too, mostly out of curiosity: would the option of end-to-end encryption in Zotero make a difference here, or would you still need/want to use a local server?
    You didn't ask it of me, but I'll be happy to answer for my institution....a majority of my references are allowed on the public web. For those references that are not, only a tiny portion of Official Use Only references would allow encryption (and it would need to be FIPS certified & the easiest way to allow that would likely to be to allow us to make use of encryption packages we already use.

    For other data (probably smaller than the public data, but larger than the OUO data), we have requirements to have an airgap between networks & so encryption features are insufficient.
  • OK, right. So it seems like we're just fundamentally not going to be able to address situations where institutional policy comes into play (at least through encryption — an easier-to-deploy data server is a separate matter, and that may become more realistic when we drop classic sync support).

    If we did offer end-to-end encryption (which, I'll clarify, isn't currently on the table), it would likely simply be to address general resistance to cloud-based services, in the same vein as, say, Firefox Sync.
Sign In or Register to comment.