So you want to run R in the cloud so you can set your Gibbs sampling off, forget about it, and not be paranoid about power cuts and reboots. Andrew Gelman hosted a good debate on the pros and cons of R in the cloud on his blog.
The consensus seems to be RStudio and EC2. P.S. If you haven’t got Andrew’s book Data Analysis Using Regression and Multilevel/Hierarchical Models (2007) go and steal one. Seriously. Anyway, here is a lovely intro on how how to go down the R EC2 route
There you have it. If you want to run R on a 64-bit server with googles of power, that is your answer. So long as you know your cat from your whoami and you don’t want to use JAGS or BUGS. Really? ok, so it wouldn’t be too hard to get JAGS talking to RStudio, but is there more to web services R than this?