How to Move Security Up the DevOps Priority List
If you are in the information security business like me, you have probably improved your frequent flyer status recently. Indeed, May-June are when most industry events occur. Like birds, we fly when spring arrives.
In this blog, I’ll share some thoughts based on conversations I had during my own journeys, including those at the global OWASP conference in Tel Aviv, Israel.
The audience was mostly split between developers and researchers, and then me, supposedly the only marketing guy within a mile radius. Since the event was held in Tel Aviv–an information security innovation hub–the vendor/customer ratio was higher than usual.
DevOps Least Favorite Word is “Security”
According to Radware’s C-Suite survey, 75% of organizations have turned information security into a marketing message. Meaning, executives understand that consumers are looking for secure products and services, and actively sell to that notion.
But do developers share the same insight, or accountability?
By nature, information security is the enemy of the agile world. In an age where software development has shifted from 80% code writing and 20% integration to 20% code writing and 80% integration, all DevOps have to do is assemble the right puzzle of scalable infrastructure, available open source modules and their end-to-end automation and orchestration tools for provisioning, run-time management and even security testing.
[You may also like: Are Your DevOps Your Biggest Security Risks?]
In other words, there’s no need to start from scratch today. Being familiar with more tools and how to efficiently navigate in Github (and other open-source communities) can yield more success than coding skills. Moreover, it yields faster time-to-market, which seems to be everybody’s interest.
Agility is the Name of the Game
As I mentioned, the global OWASP event attracted many vendors. However, will pitching ‘best of breed security’ do the trick? If you are the only one that can block rare attacks that only sophisticated hackers can carry out, is there a real business opportunity for your start-up to grow?
Well, DevOps says no!
And they are right. Running applications in the public cloud is all about efficiency and scale. Serverless and micro-services architecture fragment monolithic applications to components that are created, run and vanish without any supervision or visibility of the developer. It is done via end-to-end automation where the main orchestration tool is Kubernetes.
[You may also like: DevOps: Application Automation? The Inescapable Path]
This is agility.
Building Secure Products and Services
Both efficiency and agility are legitimate business objectives. Why would security interfere with their list of ‘what if’s?
Ironically, success doesn’t depend on how well an application security solution detects and mitigates attacks. It correlates better with how well the solution integrates into the SDLC (software development lifecycle), which essentially means it can interoperate with these orchestration and automation tools.
Before building security features, vendors should think of hands-off implementation, auto-scale, zero to minimal day-to-day management and APIs to exchange data with other tools in the customer environment.
[You may also like: How to Prevent Real-Time API Abuse]
Once all that is in place, it’s time to proceed to
security and start building the algorithmics of the detection engines and
Keep in mind security can’t be static anymore, but rather dynamic and evolving. Solutions must be able to learn and profile the behavior of traffic to the application and create policies automatically, adjusting the rules overtime when changes are introduced by the dev side. This is key for CI/CD because the last thing they want to hear about is going back to the code to reassess and test its logic, because every wrong decision translate to either a customer left out (false positives), or an attacker allowed in (false negatives).
Self-sufficient algorithmics reduces TCO significantly by
reducing the required management labor – a plague in old application security