36:03This recording is not available yet.
Ensuring Developer Intent Matches Reality
Download resourcesCategories
About this session
Mike Milner, VP of cloud technology at Trend Micro and a twenty-year security veteran, asks how a security team can tell whether what is running in the cloud is what developers meant to run. He contrasts clear-cut code bugs like SQL injection with context-dependent findings, such as an unconstrained query or a public S3 bucket, that cloud security posture management tools flag without knowing intent. Because developers vastly outnumber security staff and the industry is converging on cloud-native application protection platforms, he proposes practices that restore that context: define all infrastructure declaratively as code and deny hands-on changes in production; version-control it so code reviews and an audit trail exist, including an automatic second reviewer for infrastructure changes as Trend Micro does; and tag every deployed resource with team, repository, branch, immutable commit hash and issue queue so a runtime finding traces back to source. Fixes then land in code and redeploy rather than being tweaked in the console. Two scenarios show the payoff: locating vulnerable Log4j deployments across images, VMs and functions, and shrinking the investigation after a developer workstation is compromised to one person's changes in one time window.
Developers have the control. They write the code, they write the infrastructure-as-code, they deploy it all continuously. Ensuring what’s running is supposed to be running can be difficult, but the cloud provides extreme transparency through APIs to see exactly what exists in your environment. So when you find a problem, how do you tell if something was changed maliciously, or if the developer did it intentionally? Mike will share practical advice on how to find problems, how to compare with developer intent, and how to fix it properly in source code. Key Takeaways: - Version control everything to ensure changes are reviewed and audited - Ensure running resources match what is in version control to prevent drift - Look for issues everywhere, but fix them in source code.
Key takeaways
- Define every cloud resource declaratively as code and remove or restrict write access to production consoles so drift cannot creep in by hand.
- Version-control infrastructure alongside application code to get code review, separation of duties and an audit trail; automate a second reviewer for infrastructure changes.
- Tag deployed resources with owning team, repository, branch, immutable commit hash and the ticket queue so any runtime finding can be traced to its source.
- Fix problems in source code and redeploy rather than patching the running environment, otherwise the next deployment reintroduces the issue.
- Run posture checks on infrastructure-as-code templates in the pipeline so misconfigurations are caught before they reach production.
Speakers
Mike Milner
Mike Milner is the Head of Product Management at Trend Micro. He’s experienced the breadth of opportunities technology and data intelligence have created for businesses and governments. Between fighting cybercrime for the Canadian government and… Read moreRead less
Mike Milner is the Head of Product Management at Trend Micro. He’s experienced the breadth of opportunities technology and data intelligence have created for businesses and governments. Between fighting cybercrime for the Canadian government and working for security agencies overseas, Mike has developed a deep understanding of the global security landscape and how the underground economy dictates hacks and ultimately drives breaches. Prior his current role, Mike was CTO and Co-Founder of IMMUNIO, which was acquired by Trend Micro in 2017. He has also held technical roles with Salesforce, Canonical and the UK and Canadian governments.
