Fixing Vulnerabilities On A Shoestring

RSA CONFERENCE 2012 -- San Francisco, Calif. -- It takes less than 10 minutes at the keyboard to fix a cross-site scripting vulnerability, but any company that expects a remediation project to follow a simple calculus is in for a surprise.

On average, the actual programming only accounts for one-third of the total time in any effort to fix vulnerabilities, application development expert Daniel Cornell told attendees here today. A principal at the Denim Group, a software security consultancy, Cornell tracked 15 vulnerability remediation projects and discovered that fixing the vulnerability accounted for anywhere from 15 percent to nearly 60 percent of the total development time, with an average of 29 percent.

"It's not just the time spent on the keyboard that you have to watch out for -- there is a lot outside of that as well," he said. "The [vulnerability] may only take five minutes to fix, but then there is a lot of other hoops to jump through for, in some cases, the other 85 percent of the time."

Setting up the development environment took up to a third of some remediation projects, and confirming the fixes took up to 44 percent of project time. While the data makes up a small sample size, it can help companies take a better approach to fixing vulnerabilities in their code and reduce the time -- and thus, the cost -- spent on projects.

With more than two-thirds of a remediation project spent off the keyboard, a good way to reduce costs is to make nonprogramming activities more efficient, Cornell said. Having an automated way to stand up a development environment can cut, on average, one-sixth off the project time, and efficiently confirming fixes can eliminate another quarter of the overhead.

"In some cases, confirming the fixes and doing quality assurance took longer than fixing the actual vulnerability," Cornell said.

While actual development time accounted for the minority of most projects, be ready for outliers, Cornell warned. In its study, the Denim Group ran into an authorization flaw that took longer to fix than all the other flaws combined, he says.

To speed up development times, the security group should work with developers to plan out the remediation process. Working together can eliminate misunderstandings and combine expertise to more quickly fix flaws, he said.

The worst thing that security teams can do is bring in a voluminous PDF report from an automated static code-analysis tool, slap it on a developer's desk, and tell them to fix the issues, Cornell said. You are telling them to take time from developing features to work on fixes, a task for which they might not have the background.

"There is a lot of friction in organizations when security jumps in and tries to dictate what the development team does," he says.

In the end, taking a team-based approach helps both the developers and the security teams. A badly managed project helps no one, Cornell says.

"The result is that the business guys have no new features, which is really what they wanted, and they have half-fixed or non-fixed vulnerabilities," he said. "If you were the person that spearheaded that first remediation effort, good luck getting your next remediation effort funded."

Have a comment on this story? Please click "Add Your Comment" below. If you'd like to contact Dark Reading's editors directly, send us a message.