As DevOps Accelerates, Security's Role Changes
There remains a disconnect between developers and security teams, with uncertainty around who should handle software security.
DevOps adoption rates have increased, with 25% of companies reporting three to five years of practice, and another 37% reporting one to three years. The jump has accelerated development but driven what researchers call "a clear disconnect" between developers and security teams.
As part of its 2020 Global DevSecOps Survey, GitLab researchers polled more than 3,650 people on their DevOps successes and challenges. They learned the accelerating adoption of DevOps in general and implementation of new tools has led to changes in job functions and tool choices, as well as organizational charts within the developer, security, and operations teams.
"One of the biggest changes is a majority of respondents indicated that even today their roles are changing dramatically," says Jonathan Hunt, vice president of security with GitLab. "Over 60% of developers indicated they feel their role is changing and about 80% of security teams feel their roles and responsibilities are changing as well, with respect to DevSecOps strategy."
No longer is DevSecOps a futuristic concept or cutting-edge strategy people don't know much about, Hunt adds. Businesses are subscribed to the idea that DevOps and DevSecOps provide a significant advantage into developing code faster and identifying vulnerabilities sooner. These thoughts are echoed in a Dark Reading study focused on secure application development: 75% of organizations surveyed credit their development team with being knowledgeable about application security, and 70% say security is involved in their software development efforts.
Many organizations continue to experience a disconnect between developer and security teams. Dark Reading data shows 30% of developers are promoting code without security's involvement, and 30% of respondents consider the relationship between the teams as "neutral," "poor," or "nonexistent." More than 25% of developers in GitLab's study feel solely responsible for security, compared to 23% of testers and 21% of operations pros. Roughly one-third of security pros say they own security; 29% think everyone should be responsible for it.
Shifting Left
Nearly 75% of GitLab respondents say they have shifted testing left, meaning it's closer to the development process. Security is a tricky subject "no matter where you sit in a development organization," researchers note in their report. Nearly 40% of respondents rate their security efforts as good, while nearly 30% describe theirs as fair, and 20% as strong.
Security pros are experiencing changes in their day-to-day responsibilities as security becomes a higher priority. Nearly 28% say they're part of a cross-functional team focused on security, and 27% say they were more hands-on and involved in daily development activities. Most (65%) of security pros say their organization is bringing security into the development process earlier.
"I think the organizations are ready, and developers are ready, to take on more responsibility and participate in shared responsibility of security, but they need the tooling and guidance to do so," says Hunt. Data indicates both guidance and tooling is lacking, he continues, and right now "it's very indicative they may not know what tooling is needed," he adds.
Why Security Testing Is Tricky
Security testing remains a significant source of frustration for infosec teams. More than 42% say it happens too late in the development lifecycle; 36% say it's difficult to understand, process, and fix any vulnerabilities discovered. More than 30% describe prioritizing vulnerability remediation as an "uphill battle," and nearly 30% say it's hard to find someone to fix the bugs.
Even though most organizations perform code reviews, only 25% of vulnerabilities are caught prior to production, Hunt points out. Part of the reason, he says, is a lack of accountability.
"On average, developers are not evaluated or rated against the quality of code they produce but for the amount of code they produce," he says. "This leads to shortcuts and inattentiveness to code quality which leads to the growth of bugs and vulnerabilities being produced."
Another factor is lack of training: there are more than 600 categories of flaws, Hunt notes, and those doing code reviews are unlikely to have experience in detecting most vulnerabilities. There is also lack of automation with the right tooling. Most code reviews are manual or check-the-box requirements because PCI or another framework mandates it. "There are a number of SAST/DAST tools that can be integrated into CI/CD pipelines that will do a relatively good job at identifying vulnerabilities and your level of risk prior to production push," Hunt explains.
Related Content:
Learn from industry experts in a setting that is conducive to interaction and conversation about how to prepare for that "really bad day" in cybersecurity. Click for more information and to register.
About the Author
You May Also Like
The State of Attack Surface Management (ASM), Featuring Forrester
Nov 15, 2024Applying the Principle of Least Privilege to the Cloud
Nov 18, 2024The Right Way to Use Artificial Intelligence and Machine Learning in Incident Response
Nov 20, 2024Safeguarding GitHub Data to Fuel Web Innovation
Nov 21, 2024