Yesterday I travelled (via a slightly circuitous route) to London and the Magic Circle Headquarters to attend the JISC/Eduserve event Repositories and the Cloud. Billed as a technical and policy event I was a little concerned that it would be too high level, but as things turned out it seems my understanding of such things is better than I thought. I can only capture my main thoughts on the day, as 10 minutes in my netbook power died and I had to drop back onto my HTC Magic to continue my Web 2 participation.
The day was split into three main chunks.
In the first session Michele Kimpton (DuraSpace Foundation), Alex Wade (Microsoft Research) and Les Carr (EPrints) talked about their repository platforms, and how they are looking to engage with the cloud. I was especially interested in what Michele had to say; since DSPace and Fedora are now under one umbrella corporation I anticipate big changes in what we use currently as our repository platform. Talking to other delegates at the event it seems the smart money is backing Fedora to emerge as the sole winner, but I guess we’ll have to wait and see. I have to confess I’d never even heard of Windows Azure before today, and as it turns out I wasn’t the only one. That Microsoft are making a play into the repository market is fairly significant, as increasingly it seems that proprietory software platforms may be in all repository manager’s futures.
A lot of what they focussed on (from my P.O.V.) was the using of off-site cloud storage for large chunks of repository contents; for preservation, for reasons of perceived economic savings and maintenance of access. I was quite surprised as well to hear just what a big player Amazon are in this market; it seems their ambitions in the library and information sphere goes far beyond the supply of cheap books.
After a lot of lunch time discussions we moved onto the second session with Terry Harmer (Belfast eScience Centre). Terry’s presentation seemed more technical, and so I must confess some of the finer points were probably lost on me. However, he did raise a point about trans-national data storage and security that interested me. The fact that under British law it’s not permissable to just host any data about people off-shore is a definite challenge to moving into the cloud. Then again, how many of us are already doing such a thing if only in a small way. He touched on the need to use either EU based servers, or ensure that (for example) US hosts had signed up to safe harbour (or harbor I imagine) agreements, whereby they would agree to respect key elements of UK and EU data protection laws, where they would otherwise not apply.
He also touched on the risks involved with using a major host (for example once again) Amazon – by being a bigger target as such there was a greater risk from griefers or hackers in general for DDoS attacks (taking out access to your data) or worse. As he pointed out, most IRs are far too small to bother with; although I know that this doesn’t make them invisible or inviolate from assault.
The third part of the day broke out into two discussion sessions – one technical and one policy. Paul Miller facilitated the policy one I attended, and there followed two hours of wide ranging discussions about using cloud resources for repositories. A straw poll in the room revealed that most participants were using cloud resources already on a daily or regular basis, and a sizable number were doing this for work based activities and even with the blessing of their management. However, a lot of the group noted a certain corporate or personal reservation about trusting to the cloud to the Nth degree. One point that particularly resonated for me was the comment that “We’ve been spending a lot of time and energy advocating local repositories as something controllable, accessible locally mounted to placate certain academic worries – to suddenly shift to trans-national locations might well start ringing alarm bells for many of them.” A feeling that was expressed by some in the meeting, and in the discussions that followed, that while cloud computing options were certainly exciting the repository field itself is still insufficiently mature to go for them wholesale just yet.
The event concluded with a reception where the discussion continued into the evening, or until the Magic Circle threw us out. I spent some of the time planning some possible collaborative work with the RSP team in among sharing experiences and feedback on the day.
Overall this was a challenging day of thought and discussion. Am I convinced that cloud computing is the saviour of repositories? No. Do I think real economic savings can be made with them? Not yet. Do I think they’re relevant to the future of the repository community? Almost certainly. It will be a field well worth the watching, and doubtless there will be more about it in the months or years to come.