A Scientist's Guide to Cloud Computing

New tools (some commercial and even public), have made it so that dealing with the cloud and running large-scale processing can be rather easy and efficient. Cloud's appeal for science is clear: simplicity, elasticity (that is, the availability of large resources on the spot by launching as many instances as needed), true reproducibility (the virtual machine and the code running on that machine can be made public together with the data, when necessary), the ability to cover a large span of open questions previously unattainable due to a possible lack of computing power, and most importantly, democratization of science (since anyone has access to large computing power). PiCloud (www.multyvac.com) is one of the first commercial entities with a special focus on making scientific computing in the cloud simple for the users. PiCloud provisions AWS instances transparently to the user, acting as a middleware between AWS and the user. Their provisioning technology allows PiCloud to compete for the lower cost spot instances on Amazon.