CloudForest is an on-demand computing platform for GSB faculty and their collaborating researchers.
Using your personalized Dashboard (Stanford login required) you can request AWS EC2 instances for your computing and analysis needs. Instances are more or less akin to the
yen compute servers at GSB: Ubuntu (16) machines that have current versions of Python, R, Stata SE/MP, and MATLAB. DARC currently pays the costs for these instances, but we transparently show you cost rates and running totals right in your Dashboard.
Instances with up to 64 (logical) CPUs and 976 GB memory are available, as are instances with hardware acceleration (i.e., GPUs). You can access these instances with standard
SSH tools using standard Stanford two-factor authentication. Faculty control who has access and what level of access they have through simple, web-based group management tools. You can also connect to notebooks directly through the Dashboard, and thus launch and compute interactively in the cloud without ever leaving your browser.
|If you are new to
Need additional, temporary computing power for your work?
Need access to a lot of memory, or an access guarantee?
Need your data or machines to be isolated from other users?
Need to run with special software we can't install on-premise?
Need to run on specialized accelerator hardware, like GPUs?
Need to run the same code over (possibly many) different inputs?
CloudForest is built entirely on top of popular web technologies:
Amazon Web Services (AWS),
jekyll. Most of the stack is free: the only thing that costs money are the cloud services you use, either in AWS or for mongoDB (or both). At Stanford GSB, the DARC team pays for approved use.
CloudForest is building up to be an Open Source project. You can view and contribute to any part of the code, including this site itself and the project documentation in it. We're aiming to fully open-source the project in the early Fall of 2019.
Obviously the actual platform supported by DARC is reserved for GSB faculty and their collaborators.