Robin Hood Gone Evil: Loophole Leading To Cloud Pickpocketing Identified
Researchers at North Carolina State university and University of Oregon have proposed a jaw-dropping price tag for performing heavy duty cloud computing task – as low as zero dollars. Experiments reveal that cloud-based web browsers can be exploited to hijack the underlying computational power, and that as well, in total anonymity.
The result could be as unforgivable as cloud computing time theft of mammoth proportions. The pickpocketed resources, once fallen into the wrong hands, can be used for just about anything, including brute force password crack attempts, denial of service attacks and other genres of cycle-hungry attacks.
Contrary to relying upon the end-user’s device to perform the number crunching, cloud-based browsers make the most out of cloud resources to process and deliver web pages. This functionality of cloud-based browsers (likes of Opera Mini, Amazon Silk and Puffin) can be imitated by creating customized variants that have the potential to trick servers into performing word counts, string parsing, text search and other tasks for free. The above is accomplished by a neat hack termed as the browser MapReduce, BMR.
BMR spawns from Google’s MapReduce, an alternative mechanism to manage parallel processing of utterly large datasets. In simple words, Browser MapReduce operates by amassing free JavaScript processing cycles, in unison with a punctilious scheduling plan to effectively work around the processing bounds enforced originally by the cloud-browser providers.
The team has proved their point by saving chunks of data on URL-shortening sites, effectually deceiving them and the cloud browser providers into processing about 100MB of data for free. “What we were able to do was chain together a bunch of requests to make a larger computation“, Enck, the primary research investigator, explained.
Things are not all gloomy though. The team also presented ways to fix the cloud exploitation problem, the most effective requiring a check on the number of requests that can be directed towards the core server cluster originating from a single user. A user-authentication mechanism built into the browser should do the trick pretty well. Enck pointed out that “Instead of allowing anyone on the Internet to make requests of their servers, end users should have accounts.”
Such a methodology would allow for the service providers to notice whenever one account is generating requests that are enormously volumetric for a genuine human user. The team is all set to present their research findings at the Annual Computer Security Alliance summit to be held in the first week of December 2012.
The title of the research work, “Abusing Cloud-Based Browsers for Fun and Profit” almost says it all – cloud security measures associated with mobile devices require further fortification.
Loopholes of such sort continue to assist the bad guys in using cloud computing horsepower for not-so-noble purposes. Its about time that cloud-browser service providers take note of such weak links in the mobile cloud computing chain before the tables are turned on them.
By Humayun Shahid