The improved security in Microsoft's newest software products may convince some security researchers to look at other leading products such as iTunes and Acrobat Reader for work.
That was the message that some security professionals took away from BlueHat, an event last week on Microsoft's campus that allows security researchers to mingle with Microsoft developers.
"One of the messages we got was to look in the future for [our products] to not be so successful," said Pedram Amini, manager of security research at 3Com's Tipping Point division. That's because Microsoft is applying a lot of the technologies used by security researchers in house, making the third-party techniques not as effective, he said.
For example, he said that Microsoft Office has been susceptible to fault by fuzzing, an automated technique for finding software faults when access to the code isn't available. But Microsoft has recently put more effort into using fuzzing itself, so now third-party fuzzing technologies are unlikely to be as necessary for Office 2007.
One well-known researcher who goes by the name Halvar Flake called Windows Vista "arguably the most secure closed-source OS available on the market", in a blog post about BlueHat. "As a result I think that most of the security researchers will move on to greener pastures for a while. Why try to chase a difficult overflow out of Vista when you have Acrobat Reader installed, some antivirus software with shoddy file parsing, and the latest iTunes?"
But the security researchers don't expect to have time on their hands just because Vista and Office 2007 are more secure than their predecessors. "It's not like our industry is done now," said Dan Kaminksy, director of penetration testing services for IOActive. He pointed to weaknesses in web-based services and technologies such as virtualisation.
Others agreed. "There's always something that can be improved on," Amini said. Some researcher will come up with a new approach to bug hunting or they'll focus on different technologies, he said.
While the advent of the first BlueHat event in 2005 marked a shift at Microsoft to become more open to the security research community, this BlueHat, only the second since the release of Vista, reflected another shift, Kaminsky said.
He has seen a change in Microsoft toward considering security as an engineering problem. "If you look at security as an engineering problem, then the message from the security researchers stops being 'you bad horrible people, you write bad code,' and starts becoming 'here are changes in the engineering landscape that you need to be aware of'," he said.
That attitude change was apparent at the conference last week, he said. At some earlier BlueHat conferences, there was some antagonism among the researchers and Microsoft employees. Kaminsky remembers a presentation at the first event that took Microsoft to task for learning about certain bugs in one piece of software and then failing to prevent the same bugs in different applications. He didn't see those types of presentations this time.
Neither did Amini. "Everyone appreciated what everyone else is doing," he said.
BlueHat typically happens twice a year and Microsoft does not allow members of the press to attend.