DevOps Security and Privacy—FP Complete’s comprehensive, easy to understand guide designed to help you understand why they’re so critical to the safety of your DevOps strategy.
I’m the Founder and Chairman of FP Complete, where we help companies use state-of-the-art tools and techniques to produce secure, lightning-fast, feature-rich software, faster and more often.
Before founding FP Complete, I was an executive at Microsoft, where I served as program manager for distributed systems, and general manager of Visual C++, the leading software development tool at that time. Also, I architected MSN’s move to Internet-based server software, served as the full-time technology adviser to Bill Gates, and I founded and ran the company’s Productivity Tools Team for complex software engineering projects.
Okay, so enough about me. Let’s begin this discussion by recognizing our industry’s unfortunate—but preventable—reality:
We all know how bad the state of the world is within security and privacy right now. Projects are getting very complicated. And I—just as a sample—want to point out that this is a very typical breach. Monzo said that for six months unauthorized people had access to people’s secret code numbers, their pin numbers. I’m not singling them out at all, but rather saying… “This is very typical.” They’re a bank, and they compromised this type of data for months and months.
How does it happen? It’s not only because of logging and monitoring not being in place, although that can be a big factor. It’s because of complexity. Honestly, we’re all trying very hard to do our jobs, but users keep asking and executives keep asking for new features. And that integration just creates point after point where problems can happen, and things get overlooked.
I would argue that today’s applications are more about assembling building blocks than they are about just writing new code. But every time you increase that complexity by adding more building blocks, you increase the number of interface points between components—the number of places where somebody might have done something wrong. And so we’re really creating a system of entry points between component A and component B. But entry points—that sounds like something I would compromise if I were a security violator, right? Furthermore, we’re manually configuring our systems. People aren’t using continuous deployment. And so there is some wizard who’s supposed to go set up the latest server or integrate it with the database or integrate it with the web with a firewall or whatever they’re supposed to do. Every manual step creates further opportunities for penetration, for defects, because people are imperfect. Even the best person in your team doing a process a hundred times might do it wrong, one or two times. An automated scanner is going to find that time, and it’s going to break into your system before you know it.
DevSecOps—DevOps with security stuck right in the middle. And I think that’s a good way of looking at this problem. We want to integrate all the different parts of our engineering into one pool of automation, and include security and quality assurance as part of that automated process. We talked earlier about automated testing being part of our builds. But we want to go much farther than that, as technical teams. We want to start from the beginning of our projects, talking about how secure they need to be. What are the risks that they’re supposed to defend against or or not create? We want every member of the team to understand that system downtime—because somebody broke in and trashed it, or even worse privacy violations which you can never undo, because when people’s personal information has been published, you can’t unpublish it—we need to let our team members know that these are priorities and put them on the to-do list for the project. And we can’t call it something done if the security part isn’t done. It’s not something we tack on at the end. We don’t build in unsecured, crazy, poorly architected apps, and then at the end, ask someone to build a brick wall around them. Because as soon as one little person gets through the brick wall, it’s open season. So, we want the engineers to know everything they do should be checked for security. That’s a culture change to say that it’s everyone’s job.
We need to integrate quality assurance with security, which means somebody is checking the software we wrote for weaknesses; somebody is trying to break in or, at least, trying to run tools that will show us common ways to break in and weather their presence.
And we need to inspect our cloud systems that are running to make sure that our deployment, and our system operations and administration, is as secure as we meant it to be. Did somebody omit a step? We want to discover that right away and fix it. Or, ideally, automate the way we set up all of our systems using, for example, an orchestration software package to automatically configure our servers, so it isn’t the case that late in the day, people are more likely to make a mistake. Because, well written scripts do just as good a job even when they’re tired.
And we want to make sure that all of our systems are updated and patched and not tell people that security is a waste of time and they should get back to work on features.
To do all this, we need to have a simple design. And I would encourage people to focus on the idea that simplicity and modular design are great ways to make a system easier to check for security holes.
We want to make sure that credentials that are used in our modular systems—where one piece of software is logging into another service or another piece of software database—are kept in properly secured credential storage. A common form of security violations is you look at somebody’s source code and… Oh look! There’s the password for the database server right there …because the app had to connect to the server. That’s inappropriate design. There are special credential storage services—your team should use them.
And we want to make sure that quality control remains central to our culture, as developers of software, and that includes DevOps, that includes system administration. Too often, we have a good piece of software, and then it’s deployed incorrectly. And that’s where the problem occurs. So if you’re going to test whether your code is written properly, maybe also test whether the servers configured properly, from time to time. It’s time well spent.
So how can you move forward on security? The good news is, while it may sound like a scary and intimidating area, there are lots of practical steps you can take right now, and you don’t even have to take them all at the same time, you can take them incrementally. Here are some great steps though that I highly recommend.
One is that—in your engineering team, and if you have multiple teams—in each engineering team somebody is explicitly the security person. Somebody knows that it’s their job to keep an eye out for security issues and prevention and that if there’s a problem they’re the person who’s going to hear about it. They should have the power to look into anything they need to make sure there isn’t a security hole in the system.
Use best practices from other companies. This is a great idea throughout all of DevOps, including DevSecOps. You don’t have to reinvent anything. You can learn best practices and get a checklist together of what other companies have found helpful to look for to find opportunities to secure your system incrementally. We just piece by piece chip away at the risks that are present in our systems. We don’t have to wait until some magic day when all of security happens at once.
Teach your people about security. A lot of security problems happen because one person didn’t realize… Who didn’t know that you’re supposed to not put passwords in the source code where everyone can see them? Well, one person typed a password into the source code, but now it’s there for everyone. So be sure that training and security, and how important it is, and how to do it is available to everyone in your team. And make sure that there’s a checklist. Who took the security training? Who’s not been to security training yet?
Scary but true fact: You should, according to Price Waterhouse Coopers, if you want to be a normal IT operation, be spending 11 to 15% of your IT budget on security overall. That’s a significant number. And I think we can all agree that with more internet work and more importing of modules and stuff, we, if anything, could be worried that that number is going to go up. So automation through DevOps is really a way to keep a lid on that number. But I wouldn’t think of it as a way to make that number drive down towards zero. Security is everyone’s job, and it’s going to remain that way.
Beyond that, I’d say use it use the other techniques we talked about earlier in this presentation. You don’t have to be the next Equifax, of having no monitoring. You don’t have to allow silly mistakes by having no automation. And you don’t have to create more security holes by reinventing your own tools and processes using components. Reuse is your friend.
I won’t spend too long on this, but I wanted this for people who are more hands-on or the people who are supervising hands-on engineers. These are some practical steps that you can take to start turning on pieces of security, right now. Every one of these—except perhaps service-oriented architecture—is something that literally you could task somebody to do this week or next week.
These are straightforward tasks.
It’s a winning formula!
So, in short, you have a choice to turn on DevOps to use a lot of technology that’s been solved, a lot of best practices and engineering techniques that have already been solved and tested at numerous other companies—clients of ours, famous internet companies, everyone. When I say “everyone”, the truth is the minority of companies are already using proper DevOps. But enough companies that you don’t have to be the first, you don’t have to be the Pioneer. DevOps is a winning formula that will get you to market faster, and more reliable, and with better security. Or you could be the next Equifax and the next Capital One, which is the default situation.
FP Complete offers corporations its DevOps Success Program which offers advanced Privacy and Security software engineering mentoring among many other moving parts in the DevOps world.
Do you like this blog post and need help with DevOps, Rust or functional programming? Contact us.