Online attackers have started to experiment with embedding malicious code or links to such code in different video formats.
On Tuesday, anti-virus firm McAfee warned Windows users that the company had discovered a worm, dubbed W32/Realor, actively infecting Real Media files. The infected video files do not contain an exploit for the RealOne or Real players, but a hyperlink that points to a malicious website. When infected files are opened, the victim is referred to the web ite, which attempts to compromise their computer using a previously patched flaw in Internet Explorer.
There are numerous disadvantages to using video files to carry malicious code, but using the technique may allow attackers to take advantage of users' expectations, said Craig Schmugar, senior threat researcher with McAfee's anti-virus emergency response team.
"A chunk of people generally regard video files as safe, where they might treat screensavers and Office documents with some caution," Schmugar said.
While W32/Realor had not spread far, the incident came the same day that Microsoft distributed a patch for five security vulnerabilities in Adobe's Flash Player - software that is frequently used to play video streamed from popular internet sites. A week earlier, users of the social networking site MySpace attempted to use links in video files to surreptitiously install adware on visitors' computers.
The attention is unsurprising. Vulnerability researchers, for one, have increasingly focused on media players. In 2006, 19 medium and high-severity flaws were found in Apple's QuickTime Player, two in RealOne and Real Player, another two in Microsoft's Windows Media Player, and three in Adobe's Flash Player, according to the National Vulnerability Database. SecurityFocus sought comment from all four companies. Apple and Microsoft did not respond to the request, while RealNetworks could not provide a spokesperson in time for this article.
To date, actual video files have rarely been used as a vector of attack - typically, video plays only an incidental role. Many mass-mailing email viruses, such as the Kama Sutra or Blackmal worm, attempt to lure victims by offering an attachment that masquerades as a video. In other incidents, a Windows virus shipped on Apple video iPods and the virus - again, Blackmal - sent out to subscribers of Google's Video mailing list.
Yet, the increasing popularity of video downloads and streaming internet video - as demonstrated by the $1.6bn valuation that Google placed on internet video startup YouTube - will likely mean that online attackers will increasingly find ways to utilise the digital media as a method of compromising PCs, security experts said.
"It is my belief that most malware targets the 'large audience'," said Val Smith, co-founder of OffensiveComputing.net. "So following that, I do think that YouTube is, and will be, a target...As soon as someone comes up with a good and simple video malware kit - if they haven't already - then I think we start to see this become a problem."
Video and other media files to which people frequently link could use unique methods to boost infection rates. Malicious code could use true viral marketing, for example, using the reputation systems of community-oriented video sites such as YouTube to attempt to make infected videos more popular. The MySpace worm Samy used such techniques to build a massive friends list for the MySpace user, "Samy."
Google has processes in place to make such attacks difficult, the company said in a statement emailed to SecurityFocus.
"We work constantly to prevent people from misusing our services to distribute malicious software," Google said in the statement. "When we become aware of an instance where this happens, we take immediate action to limit user exposure.
While video files, as entertainment, evoke less suspicion on the part of users, they also have a number of disadvantages.
In the world of data, video files are the lumbering cruise liners, the Titanics. For standard TV quality video, a 30-second clip may only require downloading 2Mb to 3Mb, but a 23 minute TV episode weighs in at a third of a gigabyte, and a full-length movie can easily top a gigabyte. A video of any appreciable length contains enough bits to slow down a home internet connection to a crawl, though more highly compressed video takes up less bandwidth.
The sheer size of such files make them a less appealing target, said OffensiveComputing's Smith.
"I think users are a bit more likely to open a video file but they are also a little harder to transport around because of their size," he said. "Often malware authors want something small."
The cross-platform nature of video files also makes them less appealing, Smith said. An attacker would either have to target an exploit for a single platform or attempt to encode the file in such a way as to work on different systems. In the 40,000 samples of malware that OffensiveComputing has in its database, none use video as an infection vector, he said.
Not only is video harder for an attacker to use, the files are easier for a defender to secure, said Adrian Ludwig, senior manager for secure software engineering at Adobe.
"The threat models for video are far more well understood," Ludwig said. "It is either a good piece of video or it's not."
Run-of-the-mill Flash content, which contains scripting and is event driven, can be far more complex, Ludwig said. Flash-creator Macromedia - and then Adobe, which bought the company in April 2005 - has increasingly focused on the security of the multimedia technology as its popularity grew.
"We have been responding to the issues for many years now, and I think we put out releases as fast as anyone else," Ludwig said.
In the end, while virus writers are currently experimenting with video files, other media files will likely become more popular, said McAfee's Schmugar.
"As more users adopt any technology, that's where there attacker will go," McAfee's Schmugar said. "It is a little early to say it is the beginning of a trend, but it's safe to say that we expect to see continued attention on media files as a vector."
This article originally appeared in Security Focus.
Copyright © 2006, SecurityFocus