Basically we are trying to determine the minimum key size needed to keep secrets encrypted with the RC5 algorithm safe even in the future. Secret messages used came from a challenge set by RSA Labs in 1997. One of the projects that participate in the search is the RC5 project of distributed.net, which is using the brute force method. It's actually the only project left these days, to my knowledge.
Using brute force method is about checking each and every key to find the one that unlocks the secret message. Correct keys were already found for messages encrypted with a 56-bit and 64-bit keys and search took 265 and 1757 days, respectively. Currently we are working on message encrypted with a 72-bit key and current estimation is that it will still take 50000 days until all keys have been tested. Granted, it's possible that we have already found the key or that it will indeed be the last one we test.. :)
Given that estimation, one could argue that we already know that 72-bit is enough. However, I don't think that estimate takes into account Moore's Law (see http://en.wikipedia.org/wiki/Moore%27s_law) about doubling CPU power every few years. And it's also interesting to see how the GPU crunching is changing the picture dramatically. Especially since RC5 algorithm benefits hugely with parallel computing and that's what GPU brings to the table.
For more information and links to additional references, please go to http://www.distributed.net/RC5/.
What comes to more projects in the future, easiest would be to embrace any other distributed.net projects. However, they currently have only one other project, OGR, and that's already part of yoyo@home. And I don't want to duplicate their work and there's also no reason since it's pretty stable project.
Of course, I'm open to any suggestions about new projects that you'd want us to incorporate. :)
I don't have any in mind at the moment.