In this TechStrong TV / DevOps Chats interview, Colby Dyess, director of cloud product management at Tufin, joins Mitch Ashley to discuss security policy management across hybrid cloud, multi-cloud and cloud-native security. Recorded in late 2020, Colby shares some experience from working with customers during the rapid migration to the cloud in 2020.
The video and podcast audio of the conversation is below, followed by the transcript. Enjoy!
TechStrong TV Video
DevOps Chats Podcast Audio
Mitch Ashley: I’m joined today by Colby Dyess who is director of cloud product management at Tufin. Welcome, Colby. Welcome, it’s good to be talking with you. Would you introduce yourself, tell us a little bit about you and of course, tell us about Tufin.
Dyess: Yeah, well, you’ve got my name and my title, so I’m dialing in here from beautiful Southern New Hampshire where it’s fantastic weather. So, I’ve worked for Tufin Technologies, Tufin is really well known in the industry for providing security policy management, mostly in the organizations like the Global 2000 complex environments where you need to make changes to the network environment and you need to be able to do it quickly, safely, ideally you do it automatically and it’s secure. The Tufin product suite helps organizations do that and we’ve done that in the traditional space. Of course, our customers have adopted cloud, cloud native platforms, Kubernetes, and we’ve been taking these same capabilities for securing the environments and bringing them into the cloud, so no matter where you put your workloads, we can help you, and that’s mostly what we’ve been focused on.
Ashley: Good stuff. Boy, the center for the universe for a lot of folks. Kinda speaking of that and just thinking about the timing of where we all are, you know, there’s a lot of statistics. I run an analyst firm as part of MediaOps organization called Accelerated Strategies Group. Anyway, we have a lot of data showing some of the acceleration that’s happened, both in digital transformation and also move to the cloud, just in the last six to nine months. I don’t think that’s news to anybody, right? I don’t know if you need data to say that, because it’s happening probably in most people’s organizations.
Ashley: If I put on my security hat for a moment, we always get plenty of notice when we’re making changes and doing new projects and accelerating stuff—sarcasm—so, I’m sure there’s a lot of security folks who are scrambling still just to not only keep pace, but make sure we’re all doing a good job of securing environments that our developers are taking this into, making sure we’re understanding the services we might be using now that we weren’t before, cloud services that we’ve been in. I’m sure you’ve seen a ton of that happening with your customers.
Dyess: Yeah. It’s no surprise. Organizations have been moving to public cloud environments for a long, long time, at least, technically, for the past 15 years, they’ve been doing it, but it hasn’t always been sanctioned, right? Let’s say over the past five or eight years, organizations have really begun to intentionally adopt the cloud. And now this year, of course, everybody’s moved to the cloud about as quickly as they could.
What we’re seeing, at least what I’ve seen a lot is, security is now scrambling to try and get ahold of some visibility—what’s really going on inside these cloud environments—and trying to put some sort of guardrails around the services that the Dev teams are using. That’s a real challenge because they’ve been focused so much on how to secure most of the on prem environment or use firewalls to protect the cloud space, but now they realize they’ve got to go a bit deeper, because the threats are pretty real out there and they’re really dynamic. And if you’re not sure how to secure the cloud environment, you’re going to leave yourself open to threats.
So, security has really been spending a lot of time trying to learn more about cloud and cloud native security controls. We spend a lot of time helping customers figure out the difference between cloud native security controls and on prem stuff and then how to combine them so they can be assured it’s all secured. We see that a lot, yeah.
Ashley: Well, you know, I mean, we’ve all been doing firewalls and intrusion prevention and access controls and the list goes on.
Ashley: So, the topics aren’t new, but the environments are different and the services. It’s not like you can go into the cloud—no, let’s use all the things I’m normally using in my on prem environment. Maybe you can get some software versions of that. Sometimes it’s just better to use because it’s better integrated, you spend less time making the environment work by using what’s provided by the cloud provider.
Dyess: Yeah. That pattern is the same everywhere. I’ve seen that you’re gonna go to a new environment, you wanna take some of the things that have worked for you in the past and carry them forward. And a lot of times, that makes sense as you try to move quickly into the cloud. If you’re not familiar with cloud native security controls, you probably think, “I’ll just use firewalls and the process for managing firewalls in the cloud.” And there are a lot of options to do that, a lot of the traditional vendors in that space have tried to make a pivot to offer their same services, same capabilities in the cloud environment. So, it’s no surprise that we see a ton of that.
At the same time, those processes around managing all that stuff still takes too long. If you moved to the cloud, you went there for agility, right? You want to rapidly build applications and get them out into the market as quickly as possible. But if you’ve got to fill out a ticket to get a firewall changed and it takes three days for that stuff to be processed, you’re no longer as agile as you used to be when you could twiddle a little configuration in a Terraform template, redeploy, and—poof, you’re up and running all over again. That’s something that security teams—
Ashley: If you take another two days to answer that ticket, you’re probably gonna have a fourth after your third _____ while you’re add it to the mix, right? Some developers are gonna be like, “I’m outta here, let me go set up the next environment.”
Dyess: Yeah, it’s too difficult. You can’t wait. You don’t want to see shadow IT any longer, because IT was taking too long to respond, but you’ve got to respect that IT’s there, the security site of IT is just really trying to help protect the business.
But you know, you also saw Amazon and AWS, they both announced their own firewalls, right? Azure did a little bit back and AWS just did within the past month. And I think that’s really just paying homage to the fact that those are really good technologies. Firewall technologies really are good stuff and it’s more than security groups or NSGs, it goes beyond that, and the businesses really need these. These do address customer problems. So, we’re seeing a lot of folks ask about how do they adopt those new to the cloud provider, new technology, and how do they integrate them into the very agile processes? And I expect we’ll see a lot more of that.
In fact, what we are finding that we do often is try to help organizations see how they can get the most value out of the cloud native security controls, because these cloud environments are really, they’re pretty solid. And like you pointed out, the security controls are integrated into the services. So, it makes a lot of sense to use the services, they’re there to protect your assets and they’re so integrated in the services, you really should be trying to use them, but it’s a learning curve for security, so it means a lot to try and help them get the visibility and then start to affect control over that, so that they don’t have to slow down the rest of the business, swapping out, let’s say, a cloud native control with a legacy control [Cross talk].
Ashley: Mm-hmm. It is that tradeoff, and I think that’s one of those where it’s sort of, I think it puts security teams in a position where, are we gonna stick to our guns of this is the environment we’re gonna have consistent across all of the environments we operate in, or for sake of speed but not sacrificing security, maybe we do look at some other options and can those integrate into the monitoring and management systems as opposed to using the same firewall in every environment that we take with us there, right?
Dyess: Yeah, yeah.
Ashley: Easier said than done. I appreciate it’s not an easy job.
Dyess: Right. You know, whether you decide to use the cloud native security controls or use the traditional kinda controls, the virtualized versions of those or a combination of them, one thing I think we’re sharing is, we see a lot of conversations around how—which technologies to adopt and then how to actually manage those technologies.
And from my point of view, it seems like we’re almost too focused on getting down into the weeds about the specific security control and, as a security person, I think it makes a lot of sense to step back a little bit and try and focus on what’s the overall security policy. Instead of thinking what firewall should I use, should I use one from the cloud or should I use from my traditional vendor, maybe a better question to help organizations remain agile is for security to be able to focus on what the security policies are, what the guardrails should be, and then begin to get the practice of checking cloud environments against those guardrails, against these policies as part of your CI/CD pipeline, the whole buzzword of shift left.
That’s a real thing, right? That’s really important. That security can focus on defining security policies where they have the strength, but the actual implementation of enforcing the security policy is probably handled by the cloud operations or whoever actually twiddles the bits on the security controls. Like, if you can marry those two things together and you can do it in the CI/CD pipeline, then that’s a huge win.
We’ve seen organizations move a lot faster because security focuses on what they do and they get it codified in a way that can play in a pipeline, and now the app teams, they can make changes to an application or a cloud Ops can change the configuration of an environment and get an alert saying it doesn’t comply with the security policy. That’s great. To step away from choosing which sort of security device you want to use and focus on do things comply with the security policy? That’s a major shift, and one that seems to be really hard for organizations to do. I think to kinda get away from security being hands on with the control and now focused on the policy—it’s trust but verify, I think.
Ashley: [Laughter] Back to that old saying, right?
Ashley: You know, in a way, it also comes down—you know, we’re used to, let’s first take an inventory and make sure we know what’s all out there that we’re trying to secure.
Dyess: Sure, of course.
Ashley: But it’s equally just as important to have, whether it’s change management practices or software that can help you with notifying when changes occur or understanding, so when a change occurs, how does that map back to policies that we have in place and are we still in alignment, are there gaps that we have to take care of? Because the environment isn’t—it isn’t an environment we control everything in any more. It changes by itself, because other people are changing it along with us. So, it seems like it’s a more dynamic environment than maybe what we had 10 years ago, if you reflect back a bit.
Dyess: Intentionally more dynamic, right? Businesses have to move faster, we also have to react more quickly to customer demand or competitive threats or whatever. So, the fact that we’ve got these dynamic environments is just a reflection of what we actually needed to have built, anyway. And now, we need to find a way to incorporate that back into our regular practices.
So, you made the comment about being able to track what’s changed and when and all that, but if you adopt more of an infrastructure as code model, you already have that. You don’t need a CMDB—maybe you do, but you don’t have to have a CMDB in place when you could go back to your Git logs and see, oh, here’s the pull request for the change that we made that allowed for this security group to be open or for this Azure firewall to have this particular rule. You can actually see that information, but these are, on the security side, these are new tools. As a security person, when was the last time you looked into a Git repo to see what the configurations were? [Laughter]
Ashley: The repo—right, GitOps and security, right?
Dyess: Right. Right, but we’ve got to get somewhere around there, as a group collectively, whether we’re writing the code or we’re managing the operations or we’re managing the security collectively, we’ve got to be able to collaborate and use these tools that I can’t say are merging, they’ve been around a long time. Terraform and Helm, Ansible, Chef, Puppet, the whole bit—these things have been around for a while, but to incorporate them now into our security practices is a pretty big leap for folks.
Dyess: So, we really try to encourage that the security team gets the visibility that they need, that’s the first thing, as you pointed out, right? You absolutely need to understand what’s deployed, what’s talking to what and whether or not things are properly segmented.
You’ve got to first get that visibility, but I think before you start jumping into let’s throw a bunch of other control points in there, consider what resources are there and how can you, on the security side, participate in automation, participate in these flows that your CloudOps, DevOps, SRE, whomever had set up. Because there’s a really good opportunity now on security to be part of this agile process and to allow the business to be secure, but to do so in a way that won’t break the developer productivity.
Dyess: And that’s, it’s uncomfortable, because it’s a big change, but it’s a change that I know it can be done successfully. We’ve worked with companies that have done that, you know, make the conscious decision.
Ashley: I almost wish we had stepped back and redefined shift left, because if, you know, I were king of the world for a nanosecond, I would think about it as, it’s not just about getting security people to work with the software teams earlier in the life cycle of software—yes, it’s that. But think about it this way of, the old world was, set the policies and then how do we implement the enforcement mechanisms, right, to make sure that those are being followed.
Ashley: I think in the world we’re in now, it’s, security people are, if you were gonna set up a security practice in an organization today, what I would do is, let’s start with how software is created, let’s work with the team about how the entire tool chain is set up, how that flow happens, both for application software, but also for the infrastructure and how that gets created where we’re talking about test environments and production environments and maybe you’re stepping into infrastructure as code if it really, truly is kinda dynamic Terraform or whatever kind of environment.
But start there, because if you understand how that’s done, now you know how to secure it. You know how to work with teams to help them secure it, then the next layer would be getting into the software architecture itself around microservices and things that are more porous than we’re used to, kinda the hard outer shell of monolith applications of the past.
That, to me, is the mental shift, the big mental shift for a security team to not necessarily be an expert at all that stuff, but understanding that’s what you’re really securing now.
Dyess: So, two things that stand out, like responses I can imagine people having right now and that’s, I don’t have time to go and learn all that stuff. [Laughter] I’ve already got 20 other tickets for all high priority things. I’ve gotta get on that right now, so I can’t go and learn all that stuff. Valid, that’s definitely the case in a lot of organizations, yeah.
And another problem related to this is security teams who maybe do want to know, maybe have some time to do that, but struggle that the application teams themselves don’t even know what makes up the application and what connections do they need. So, even the folks who you think would know because they built it, you’d think they would know the answers to some of these questions, but it turns out that they don’t—in a shockingly large number of scenarios, that’s the case.
Dyess: And so, if you’re gonna shift left, you’ve got to address not just the problem of the security team getting visibility in this—or maybe, actually, it’s the same thing. The application teams have to be able to come to the table to say, “This is what the application looks like. Here’s how it’s put together. Here are the resources it needs and here are the services that will rely upon it.”
Dyess: And so, if application groups can come to the table with that, then a security team can look at it usually very quickly, look at a map and say, oh, I can see where it needs to be segmented this way. I can see where he’s a resource that has too close of an access to PII, PCI, right? And so, that conversation could start a lot sooner if it’s part of your normal bill practices, you were generating these views, you had some way to produce an item that a network team or security team could look at, right? They don’t need to look at the source code, they need to look at your network topology, they need to understand what you’re connecting to.
So, if you can bridge that gap, that goes a long way. So, to your point, it’s not just like, “I just need to check if it doesn’t comply with the policy.” Sometimes, it’s even just trying to define the policy at the outset.
Ashley: Yeah, on what you’re trying to secure.
Dyess: Yeah, but also, while security never seems to get the heads up about the things that are coming down, the joke that you made at the beginning, it’s also the case that when the application team is building something that they don’t really have all the rules, they’re not sure what all the restrictions are, and there are certainly plenty of guardrails, guidelines that they could have if you give them access that they could pull from an API to do their own checks.
So, you know, there’s a model. We’ve done this before in the industry, back when I wrote code, we had a separate QA team, so we’d write code, we’d do a little bit of testing, but just enough, frankly, to get us through the build, right? And then we throw it over the wall and it went over to QA. And then thankfully, our development practices matured to a point where we realize we actually need to integrate testing as part of what we do. So, as developers, we had to own some of that testing, some more of the testing, but also importantly, the test team got involved earlier in the application life cycle, the model you were talking about. So, the QA teams could come in and learn a bit more and begin to set up some of the additional tests and participate in the build process, so what they did showed up in NMake.
So, I think security is on a path like that as well. I think security, instead of app teams just throwing it over the wall and asking security to figure out how to lock it all down, I think we’re going to, and in some places, some customers for sure are moving to a place where security is part of the process earlier, and if they can say, “Here’s some of the tests,” so now the guardrails, they can contribute some of that and they can go into this build process, then now you’ve really begun to get that shift left.
But I know this can happen. We’ve seen it happen before. Now QA is just integral to what we do. You can’t even imagine putting a product out without some part of QA being part of the build process and tests happening really early on, test planning happening early on. I imagine security just going in that same sort of loop.
Ashley: Yeah, of course Tufin’s been in this space for a while, has built up, I’m assuming, a body of knowledge about this. If you were gonna advise maybe a new customer or an existing customer on—okay, we’re always playing catch up, we’re really playing catch up now with how much more cloud services that organizations are using in such a short period of time, what are some of the sort of best practices or recommendations of, “Here’s to sort of get the first thing established, here’s the next thing to work on, here’s the next thing to work on.”
Ashley: Any suggestions about that?
Dyess: Sure. We have this whole crawl, walk, run methodology, and it’s designed to help organizations take these small steps towards getting to really secured environments. And the first step right away is just to get visibility, which is kind of a “no, duh” statement, but it’s actually one of the big problems that orgs have. You’ve got to be able to see what’s deployed inside your cloud, what’s deployed inside your cluster, and which resources are communicating with each other, where are the dependencies.
So, understanding that is a really major first step. It’s true for the application teams, but for the security, it’s to know are things properly segmented, where are the biggest risks? So, getting that awareness, that’s really the crawl phase. And in the walk phase, that’s where you take what you’ve learned about what’s deployed, what talks to what, and whether or not things are segmented properly and then you begin to isolate. Here are specific accounts that we need to tackle or here are clusters that we need to tackle—like, hone in on the higher priority items and begin to define a security policy for those cloud native environments.
And as you develop the cloud native policies, which you then share with the application teams or the cloud operators, whoever’s responsible for configuring it. As you develop that sort of, that skill, you’ve gotta get to the run phase, and the run phase is when you begin to automate this. You automate generating security policies, you automate validating security policies, but at the very beginning, where most organizations are at right now is, first get the visibility and then the second step of defining a security policy. Not defining what is my security group definition, not defining what is my AWS firewall definition, but what is my security policy? What are the rules that should govern access to applications, assets, resources of certain categories? Those are two big steps to take right away.
Ashley: I think that’s some great advice. I’m curious, any thoughts as we—I think we all want to turn the page on 2020 and kinda get to 2021, and hopefully, we’re looking at a different kind of a year. But one of the things about 2020 is, we’re not gonna go back, I don’t think. We’re not gonna go back to, let’s rewind to January and start working in that environment again, right? Everything is changed with remote work or how much we’re using the cloud.
What are some things that you’re anticipating for 2021 as you look forward in the next 6 to 12 months. Do you kinda think it will be topics maybe not as talked about yet but will be more topical in 2021? I’m just kinda asking you—
Dyess: Are you asking for predictions of 2021?
Ashley: You know, I don’t see it, but I’m pretty sure, there might be a crystal ball right behind your chair, just look back a little bit there and kinda check back. [Laughter]
Dyess: [Laughter] Well, I think, given the maturity of the market adoption of public cloud environments, I think we’re going to see more security teams getting their arms around how to help in securing these cloud environments without necessarily always relying on the traditional devices. We’re already starting to see that even now. So, I think we’ll see more of the cloud native security control adoption. I know organizations have seen the benefits of doing that to lower cost of operations, just total lower cost of ownership, plus the benefit of being able to say when there’s a problem, I can call my cloud vendor, I don’t have to worry about vendors poking at each other. So, I think we’ll see more of that.
I think it’s the Kubernetes space that’s probably the most interesting, because it’s the least well known for most folks on the security side.
Dyess: So, I’d say maybe over the past year, there’s been a lot of emphasis, just like with cloud, slap a firewall on it and we’ll solve the problems of micro-segmentation later, but later is gonna happen real quick, and I’d expect in 2021, you’ll see securities start to get more involved in getting visibility into the clusters and then looking to define the security policies.
And not—again, not to try and define the policy of an individual cluster, but just the general rules, because there’s no way we can keep up with what will really be thousands of services running inside of a cluster.
Dyess: So, we’ll see that. I think we’ll see more organizations at the security side have to deal with the scale and the rate of change and begin to participate in the automation processes that the Dev teams are using to manage it.
Ashley: Maybe that’s something we’ll hear more frequently. I wouldn’t be surprised if the—let’s think of it as cloud native security, not just as security, sort of think about it as one of you architectural approaches to how you do security.
Dyess: Cloud native. The native controls of the cloud provider—yeah, that’s a big shift for folks, I think.
Ashley: It is. We’ll see if something like that happens. And I think the things you said made a lot of sense. Well, it’s been great talking with you, it’s the first chance you and I have had to talk before, and—
Dyess: Yeah, it was good to talk to you, Mitch.
Ashley: – a lot of fun. Tell us a little bit about where folks can learn more about Tufin.
Dyess: Yeah, of course, you can go to Tufin.com, that’s T-U-F-I-N dot com. On the cloud side, like what’s really relevant to what we’re doing here, we’ve got a lot of resources at Tufin.com/try-securecloud. It’s cloud native security controls. It’s really designed to help you get visibility and control the cloud native security components. Great stuff there, free to use product. There’s a great trial for people to give a go. Tufin.com would be the place.
Ashley: Great. Perfect. Thanks. Well, we wish you the best and a even better 2021, we’ll look forward to that.
Dyess: Yeah. From your lips, man.
Dyess: Good luck in 2021, everybody.
Ashley: There you go. Hang onto that, hang onto that. Alright, well, take care. It was very nice talking with you, Colby—Colby Dyess from Tufin.