The number of coding mistakes on websites continues to fall but companies are slow to fix issues that could be exploited by hackers working with improved attack tools, a security expert said.
The average number of serious vulnerabilities introduced to websites by developers in 2011 was 148, down from 230 in 2010 and 480 in 2009, said Jeremiah Grossman, chief technology officer for WhiteHat Security, which specialises in testing websites for security issues. Grossman spoke on the sidelines of the Open Web Application Security Project conference in Sydney on Monday.
The vulnerabilities are contained within custom website code and are not issues that can be fixed by applying patches from, for example, Microsoft or Oracle, Grossman said. According to WhiteHat Security statistics, it takes organisations an average of 100 days to fix about half of their vulnerabilities.
The risk is that vulnerabilities which haven't been speedily remedied could be found by a hacker, resulting in a high-profile data breach such as those that affected Sony, the analyst firm Stratfor Global Intelligence, and AT&T.
Hackers are honing their skills and are becoming better focused. They are using a wider array of improved tools in order to find coding problems in websites. "Offense gets better every year," Grossman said.
Security analysts in Grossman's company constantly try to hack websites belonging to major financial institutions and other companies -- with permission. Developers in those companies don't tell WhiteHat when they roll out new features or make changes. WhiteHat's hackers go to work, trying to find cross-site scripting flaws, SQL injection or information leakage vulnerabilities.
"We are constantly smashing [websites]," Grossman said. "We're LulzSec or Anonymous 24/7. We don't stop."
Companies decide whether they want to fix the problems, which often involves reassigning a developer working on a new feature that the business needs to roll out, Grossman said. It's a gamble whether or not to fix, since the vulnerability may never be found by a hacker but could cost the company dearly if it is.
"Do you take the developer off that [project] and put them on correcting a vulnerability that they know they have but may or may not get exploited and may or may not cost them any money whatsoever?" Grossman said.
The best scenario is to write good software from the start, with a keen eye on security. "We're not going to get perfect at software, but we can get economically good enough software," Grossman said.