Last year I was asked to join the organising committee for the Eduserv JISC ‘Repositories in the Cloud‘ event that was held at the fantastic Magic Circle venue near Euston Station. The sell out day was a great success, the speakers giving an excellent overview of the current state of the art for cloud computing applications in the area of repository storage in particular. ‘Compute’ in the cloud was also discussed as one of main benefits of cloud technology in helping to reduce bandwidth by placing the compute next to the storage.

It was clear from the event that it’s still quite early days in the use of cloud technologies for repositories, and many have the usual concerns based around the security, control and licensing of the data that you often hear for cloud storage in general. The idea of going for a hybrid approach sounded like a sensible option, where you may keep critical and important data on your own servers, and use the cloud for less critical data, or perhaps use it more as a backup service.

Video recordings of the presentations by Michele Kimpton, CBO DuraSpace, ‘DuraCloud – Open technologies and services for managing durable data in the cloud’, Alex Wade, Director Scholarly Communication, Microsoft Research, ‘Cloud Services for Repositories’, and Les Carr, EPrints, University of Southampton, ‘EPrints Cloud Visions’ are now available on the Eduserv event page.

I caught up with most of the speakers, and a number of the attendees for some quick video reactions, thoughts and commentary. These are available on the event page and I’ve included them here:

Michele Kimpton	Alex Wade	Les Carr	Brad McLean
Ross MacIntyre	Michael Guthrie	John Salter	Rob Sanderson
Simeon Warner	Kevin Ashley	Jane Stevenson	Dave Tarrant

• Go with the Web
  Mainstream is good. Even if it doesn’t seem to fit, it’s probably a good idea. SWORD went with the Atom Publishing Protocol. It was developed for blogs.
• Do you really need to standardise?
  Maybe de-facto is good enough? Why waste time and effort? The DSpace, Fedora, EPrints, Microsoft Zentity and Intralibrary repositories all ship with SWORD in their current releases. Microsoft have adopted SWORD as their de-facto standard for deposit and have implemented it in their ‘Article Authoring Add-in for Word‘. ‘Nuff said.
• Prove a point
  Develop some test implementations and demo clients. Show the thing works.
• Be agile
  … and don’t be too prescriptive.
• You can never be too simple
  Do one simple thing well. Don’t try to do everything. It’s got to be clear.
• Get the message out
  Some say that the marketing is the most important thing.
• Don’t just say it, do it
  Practice what you preach. You know the quote, there’s too many standards and specifications. Re-cycle. Be strong, don’t re-invent.
• Allow for serendipity
  … and embrace it when it happens.
• Don’t go it alone
  Get people on board. Bring people with you.
• Having great developers makes life easier
  It means you can get things done.

I mentioned a new UKOLN project called ‘RepUK’ that will be providing a Linked Data interface to an aggregation of scholarly materials from a number of UK repositories using the Talis platform. I then outlined a few issues around Linked Data, as well as mentioning the new Manchester OpenData project. Following on from discussions at the CETIS conference, I highlighted the difficulty of convincing IT managers and VCs that providing Linked Data interfaces to institutional systems is a worthwhile venture. The slides below provide a pointer to the full range of issues I raised.

After lunch Thanassis Tiropanis gave us an overview of the SemTech Project roadmap and recommendations (pdf).  Following this there was a general discussion about the way ahead for the project, but I’m not sure there were any clear decisions from the day. Nevertheless, it was a useful day for myself and hopefully a productive one for CETIS in determining where to go next.