Cloud Adversarial Models

Last April I attended a workshop organized by the NSF on cloud security (see here for an overview from the Computing Community Consortium blog). The goal was to get a few people to think about the most interesting and important future research directions in cloud security. The attendees came from a wide range of backgrounds: operating systems, networking, computer architecture, security and cryptography. It was really interesting to see how people from different communities think about and approach the same issues.

A Key to the Cloud

At this point most people would agree that cloud computing represents a major shift in computing. Consider, for example, that many tech companies are entirely cloud-based. This includes the more established like Netflix and startups like Instagram and Pinterest. Instead of building and maintaining their own computing infrastructure, they can scale on-demand, save money and focus on improving their services. Scientists and engineers whose research was limited by the computing resources they had at their disposal (usually whatever their departments could afford) can now use the cloud to process huge amounts of data and run large-scale simulations cheaply 1.