Public concerned about security flaws in government open source code

Earlier this month, the White House released a draft of an open source code policy for public review which would require agencies to share code with each other and with the public, but some experts are concerned about possible security implications.

Under the proposed new policy, new software developed specifically for or by the federal government must be made available to other federal agencies. In addition, under a pilot program, a portion of the federally-funded code must be released to the public as open source.

“We can save taxpayer dollars by avoiding duplicative custom software purchases and promote innovation and collaboration across federal agencies,” said Tony Scott, the U.S. chief information officer, in the announcement. “We will also enable the brightest minds inside and outside of government to review and improve our code, and work together to ensure that the code is secure, reliable, and effective.”

[ ALSO ON CSO: The state of open source security ]

Some members of the public are already expressing concern about the security implications, however.

“The world does not need more insecure code,” said John Pescatore, director of emerging trends at SANS Institute, in a comment submitted via the project’s GitHub page. “This process needs to as a minimum require common process for code to be tested for basic app development hygiene before being submitted or at least before being accepted.”

Code should undergo quality review before it is used, agreed John Scott, founder and co-chair at Washington, D.C.-based Open Source for America, an effort to encourage more government use of open source software.

But agencies shouldn’t be required to use code review tools before releasing it to the public, he added.

“Testing of the code before release will just create a bottleneck and could stop progress,” he said in his comment.

Many commenters pointed out that public scrutiny by itself will help catch security problems.

But vulnerabilities are regularly found in open source projects, even ones that have been around for a while.

In addition, open source code poses two additional security problems, said Mike Pittenger, vice president of security strategy at Black Duck Software. “Open source projects are often ubiquitous, so if there’s a vulnerability it creates a target-rich environment for attackers,” he said.

Attackers who spot a vulnerability in a popular tool can simply keep trying it against different organizations until they find one who uses it.

The other problem is that open source code often doesn’t have an organization behind it to push out updates, Pittenger added.

“With open source, there’s a lower up front cost, and you don’t have to pay for support,” he said. “But it’s up to you to monitor the open source community and see if there’s a vulnerability and pull the patch. You multiply that by the number of open source projects and components used in an environment, and you quickly have a problem.”

Black Duck is currently tracking more than 1.5 million different open source projects, he added.

“And we’re seeing the pace of development increasing because it’s more and more accepted and no longer viewed as a scary, bad thing,” he added.

The public comment period ends on April 11.