We're probably the folks who brought up S3 back when. We have a RAID server in-house that will ultimately hold about 6GB (which is enough for current projects once we get them all in manageable, preservable forms) and are backing that up to S3. My experience with removable media and/or optical media is sufficiently negative that our policy is first back up to RAID inhouse, then S3, then LOCKSS with other media as convenient (in our case, probably DVD and minimal tape). We are also using this with Subversion for the accompanying documentation (transcripts, interview logs, metadata descriptions), etc. I'll be presenting on the subject at the upcoming conference in Washington, DC--do come by!
We are currently working on a minimalist Fedora install that is going to be based on Amazon's services so that we never have to purchase (or maintain) yet another server for this purpose. The bandwidth/usage charges are not soooo cheap, but the result, at least in this phase, is a server that is easily accessible, easily secured, easily re-imaged separate from the server content (that archive on S3 and other online services) and for which we don't have to write useful specs at a time when we're not sure what the real requirements will be once things (if things) stabilize to where a commodity physical solution (and its upkeep) is more economical. I am hoping to avoid ever again to be putting in a hard disk that has been sitting around for six months (or six months) hoping to find a way to get it unstuck and working, or hoping that the last copy of the three we had in separate places on separate optical media will work. But we're at least a year off having enough of this implemented in enough places that I'll be able to sleep easily on this account. I may change my tune by then ;-). ari On Thu, Sep 11, 2008 at 3:42 PM, Mary Bloodworth <MBLOODWORTH at folger.edu>wrote: > I am writing to ask if any of you -- like us here at the Folger > Shakespeare Library -- are at small institutions and without major > digital asset management or IT infrastructures but nevertheless are > engaged in active digital imaging? > > > > If so, are you willing to talk (offlist or on) about your backup / > archiving schemes? We are working on establishing scalable systems > architecture and backup strategies for digital images of collection > materials, and would love to compare notes with others who have some > version of a 2-3 tier backup strategy. > > > > Our current situation is this: For each digital image of collection > material, our Photography and Digital Imaging lab produces a minimum of > two images: a ca. 100-120 mb unretouched master, and a ca. 80-100 mb > cropped & color-corrected derivative. We are looking for a solution that > will permit us to archive the masters offsite. We're currently running > tape backups and taking them to a staff member's house. However, tapes > sitting on the bookshelf in a Folger staff member's house isn't good > enough anymore. > > > > What we'll need is at least 1.5 - 2 TB of space. This can be a dark > archive because we won't need frequent access, though infrequent access > would be necessary. I looked at the MCN-L archives and found one thread > from November, in which some spoke of Amazon S3. Any thoughts on this, > or a different service that's cost-effective? > > > > With thanks in advance, > Mary Bloodworth > > Head of Information Services > > Folger Shakespeare Library > > _______________________________________________ > You are currently subscribed to mcn-l, the listserv of the Museum Computer > Network (http://www.mcn.edu) > > To post to this list, send messages to: mcn-l at mcn.edu > > To unsubscribe or change mcn-l delivery options visit: > http://toronto.mediatrope.com/mailman/listinfo/mcn-l >
