Steven Abrams discussed the work the California Digital Library is doing to figure out the price they need to charge their customers from the University of California system and elsewhere for preservation services. They use a very detailed breakdown of cost elements for the whole preservation process, together with some relatively simple assumptions about their economics. The details are here.
Matthew Addis described work under the auspices of PrestoCentre that is intermediate between the work at CDL and my work. He uses a somewhat less detailed breakdown of the preservation process than CDL, but like I do uses stochastic models. He models the cost vs. reliability tradeoff using specific events with associated costs such as integrity checks and media migrations. The details are here.
I gave a talk about our work modeling the economics of storage, based on my talk at PDA2012 with updates from my recent blog posts on storing everything in the cloud and the discussions at the Storage Valley Supper Club. You can find the resources I used by following those links.
There was a surprising level of agreement between the three of us on the major points, including:
- The very great uncertainties attached to any economic models of preservation.
- The fact that the bulk of the cost of preservation is the one-off, up-front cost of ingest and the related metadata generation.
- The difficulties posed by different organization's differing approaches to accounting (for example, the way University budgeting's focus on current cash flow makes long-term investment hard).
- That the economic situation and the probable flattening of the Kryder's Law curve would make it increasing difficult to justify preserving content in the future.