University researchers  have found a way to bypass cloud security, something that will greatly concern the likes of Amazon and Microsoft who have been pushing cloud-computing services as a low-cost way to outsource computing power.

In experiments with Amazon's EC2, researchers at the University of California, San Diego, and the Massachusetts Institute of Technology, have shown that they could pull off some very basic versions of what are known as side-channel attacks. A side-channel attacker looks at indirect information related to the computer - the electromagnetic emanations from screens or keyboards, for example - to determine what is going on in the machine

The researchers were able to pinpoint the physical server used by programs running on the EC2 cloud and then extract small amounts of data from these programs, by placing their own software there and launching a side-channel attack. Security experts say the attacks developed by the researchers are minor, but they believe side-channel techniques could lead to more serious problems for cloud computing.

Many users are already reluctant to use cloud services because of regulatory concerns - they need to have a better handle on the physical location of their data - but the side-channel research brings a whole new set of problems, according to Tadayoshi Kohno, an assistant professor with the University of Washington's computer science department. "It's exactly these types of concerns - the threat of the unknown - that is going to make a lot of people hesitant to use cloud services such as EC2."

In the past, some side-channel attacks have been very successful. In 2001, researchers at the University of California, Berkeley, showed how they were able to extract password information from an encrypted SSH (Secure Shell) data stream by performing a statistical analysis of the way keyboard strokes generated traffic on the network.

The UC and MIT researchers weren't able to achieve anything that sophisticated, but they think their work may open the door to future research in this area. "A virtual machine is not proof against all of the kinds of side-channel attacks that we've been hearing about for years," said Stefan Savage, associate professor with UC San Diego, and one of the authors of the paper.

By looking at the computer's memory cache, the researchers were able to glean some basic information about when other users on the same machine were using a keyboard, for example to access the computer using an SSH terminal. They believe that by measuring the time between keystrokes they could eventually figure out what is being typed on the machine using the same techniques as the Berkeley researchers.

Savage and his co-authors Thomas Ristenpart, Eran Tromer and Hovav Shacham were also able to measure the cache activity when the computer was performing simple tasks such as loading a particular web page. They believe that this method could be used to do things such as see how many Internet users were visiting a server or even which pages they were viewing.