It was a good hackfest day! I was involved in the Preservation through Redundant, Distributed and Indifferent Filespace project. We didn’t write any code in the end (it is a huge project), but we did identify what features we would like to see in a software project of this type, along with some of the potential problems that need to be addressed. We found two project that partially met the needs we found. They are DIBS (Distributed Internet Backup System), and OceanStore. It is interesting to note that both of these projects come out of Berkeley.
To be clear, we were looking for software that uses a peer to peer model for distributed backups in an untrusted environment. For example I would install the client software on my computer, and would tell it how much data I want to backup (50 GB for example). The software would then ask me how much redundancy I want (1,2,3,4… copies of the backup). Then it would tell me that for that particular level of backup I would need to make 150MB available on my hard drive for other people to store their backups. The backups would be digitally signed and encrypted for privacy and integrity purposes. This selection of backup computers on the internet would be done automatically by default, or manually if desired.
I’ve installed DIBS on my laptop (on the linux side of my laptop), but have not installed it on another computer. I hope to try it out when I get home. It was also written in Python (which is a language that I’m trying to use more right now), so that is another reason for me to try to get involved on some level in this project.