[00:00:00] James Christiansen: I mean, I think you have to be an innovator to be a leader. You have to keep challenging the status quo. You have to keep challenging yesterday's thoughts. That's what we really did when we sat down as a team started listening to our colleagues and taking that input along with their own cost to really develop out these principles, challenging the way we've been doing things and really thinking about, how does this digitalization is changing us and our organization.
[00:00:34] Producer:Hello, and welcome to security visionaries hosted by Jason Clark, chief security officer and chief strategy officer at Netskope. You just heard from James Christianson, the vice president, chief information security officer at Netskope on this show. You'll hear from world-class practitioners and thought leaders like James on how they stay on top of the game in networking and cloud security. You're about to listen to the second half of a two-part discussion on the principles of security transformation. In this half, Jason and Erick are joined by colleagues, James Christianson, the Vice President, Chief Information Security Officer at Netskope. James Robinson, deputy chief information security officer at Netskope and Lamont Orange, Chief Information Security Officer at Netskope. The following discussion and the security visionaries podcast are part of the security transformation playbook, a set of new resources from Netskope and some of the industry's most forward thinking leaders examining the most important issues in security today, before we dive in, here's a brief word from our sponsors.
[00:01:38] Sponsor: The Security Visionaries podcast is powered by the team at Netskope. Netskope is the SASE leader. Offering everything you need to provide a fast data centric and cloud smart user experience at the speed of business today. Learn more at netskope.com
[00:01:57] Producer: without further ado, please enjoy episode two of security visionaries with your host, Jason Clark.
[00:02:05] Jason Clark: In the last episode, Erick and I talked about the genesis for the security transformation project and explained that there are several principles for the future that we should work on, right. 10 principles that we're going to do a deep dive on specifically today. And so I'm joined first by Erick, Rudiak, Erick, how are you?
[00:02:24] Erickk Rudiak: Hi Jason. Glad to be back. Thank you so much for having me on.
[00:02:29] Jason Clark: Happy to have you. And Lamont Orange.
[00:02:31] Lamont Orange: Hey Jason, thanks for having me on the show. I look forward to the conversation.
[00:02:36] Jason Clark: And James Robinson.
James Robinson: Hey, happy to be here. Thanks.
Jason Clark: And James Christiansen.
James Christiansen: Let’s rock and roll!
Jason Clark: So guys, welcome. Welcome to the conversation. How you guys doing? Are you ready?
James Robinson: I'm ready. Let's do it.
Jason Clark: Awesome. Lamont, I know that you, uh, this morning had to, uh, you already were on stage on a panel conversation. You had to race to this conversation. So thank you for that, but it's gonna, it's probably easy and nice to do back to back.
[00:03:06] Lamont Orange: Definitely. So my pleasure.
[00:03:09] Jason Clark: So we altogether, you know, over the last couple of years had worked on. Really with the industry, right. Spending time on, you know, hundreds of, of round table dinners and, and workshops and, you know, surveys and one-on-one conversations, right. Trying to collect. What is the future of security look like? And in this new world, Kind of digital transformation is just happening period. Right. For every organization. And security is kind of in this upside down world where we're trying to, you know, security teams are being stretched beyond belief. Right. And, and trying to keep up. So how are they going to be able to perform and gain leverage. In this new model, right? Cause they're obviously stressed and, um, they're looking at the legacy technology architectures and, and in the end, kind of these, these past ways that we've done stuff for the last 25 years, that we've all been working in this industry or more, we spent time together on 10 principles for the future. Right. And then obviously the rest of the security transformation playbook as part of, what does, if I need to get to by 20, 25 and beyond. So, you know, all of you have helped tremendous amount in this, right? And so just look for each of you to, to give, you know, your context to each principle, as we talk to them and experiences that you've had in these conversations, when these, you know, I'd say over, gosh, probably well over a thousand CISOs and CIO conversations that we. On this. Right. But also in your experiences as CISOs in any, you know, past lives as, as operating other organizations, as CISOs write about the past, moving to the future. So with that, you know, I'm going to start off with the principle one, right? Principle one is challenge all your existing principals, right? So what, what principles do you believe needs to be challenged from the past?
[00:05:04] James Christiansen: Yeah, you know, I really am. I always challenge everything. I think every day. I mean, I think you have to be an innovator to be a leader. You have to keep challenging status quo. You have to keep challenging yesterday's thoughts. And I think that's what we really did when we sat down. And as a team started listening to our colleagues and taking that input along with our own thoughts to really develop out these principles, it was really about challenging status quo. It's about challenging the way we've been doing things and really thinking about how business digitalization is changing us and our organizations, and certainly the, the quick movement to work from anywhere and what that's doing to the things we have to do as we look for.
[00:05:51] Jason Clark: What about the principle of, if it's not broke, don't fix it.
[00:05:54] James Christiansen: Yeah. I don't know. I've always been really good at breaking things, so I'm not the right guy to ask.
[00:06:00] Lamont Orange: And when you look at digital transformation, honestly, you're probably sanded. It's already broken. From how you want to move the organization forward so that, that you have to say that it is broken already, and it does need fixing because we're still having an escalated amount of attacks from attackers. They’re still being successful and are being successful at a high velocity rate. So we must come back to, it's already broken. Now it's how do we plant a seed and move forward?
[00:06:32] Jason Clark: So James, you’re a product security and application, you know, expert. And in my view, right, you've been doing it for a very long time for very large organizations. Well, how are the principals changed? From an app sec or product security or software pipeline standpoint.
[00:06:50] James Robinson: Yeah, one, the one that came to mind that, uh, I think it also got looped into a different principle was that trust, but verify that was one that for a long, long, long time we relied on the all through out the product security AppSec domain. And I think that now that really gets challenged a lot. That's one that was actually a very comfortable principle. Um, and one that, that I've relied on for many, many years that now it just totally gets broken, totally gets challenged and know that there's a lot of conversations about zero trust principles and it having its own. But it's really the zero trust architectures
[00:07:24] Jason Clark: That’s a good point. Trust, but verify has changed into zero trust. Right? That's that's a very good point. So principle two. Stop buying black box solutions and buy open and integrated. So I'd like to kind of say that in general vendors have bought a lot or technology companies, right, they’ve bought a lot of companies and integrated them, or they claim to integrate them. But generally the integration is a price list and the sales person selling to you. So overall, what's your guys' view on how the, how the industry needs to change in the way that we procure technologies?
[00:07:59] James Robinson: I'm jumping on this one first, because this, this was actually one that I saw that I've been talking internally a lot with Lamont about. Um, and it's that idea. I love the idea of open, you know, open NDR, open XDR, open cloud. Yeah, those, those types of things, we have to be able to, you know, make that almost requirement number one. In many ways, you know, we know the sum of many things is better than the sum of One, if you buy the black box, that's what you're getting is that sum of one, or maybe a sum of a few, um, and you have to, you know, build that intelligence by being open. That's really where it comes into. I'm a huge believer you'll out of, uh, out of some of them, you know, I know we've only talked about two, but right now this one is a probably ranked higher for me then principle one that we, that we talked about with challenge, everything, right? This one is, you know, this one is, is core. I think for us to be able to succeed with the future,
[00:08:54] James Christiansen: You know, I'm Jason I've bought best of breed products a lot through my career. You know, we, we very much went after with, uh, very aggressive companies like these, uh, that I was working for, you know, but today's world, you know, I have to look at best of breed platforms. I just can't afford the manpower it takes to manage all these different solutions. And the complexity it's brought to the organization, you know, just leads to human errors, leads to patches, not getting applied versions, not getting updated. So I've really had to move away and my thought process away from best of breed and started looking at best of breed platforms. Now, what can give me the best tightest integration, like you said, in your opening conversation, it can't be somebody with a lot of skew numbers, it has to be a truly integrated platform to solve the real problems.
[00:09:50] Jason Clark: So Erick, any thoughts from you on open and integrated? Why every solution we buy from this point forward, it should be more part of the ecosystem versus being the black boxes of kind of that we procured in the past.
[00:10:03] Erick Rudiak: Yeah. Great question, Jason. So like when I think about open and integrated and why it's so important, our systems are so interconnected. If there's no API to create visibility, like the complexity and interconnectedness of our systems kind of demands that signal from one defensive system, uh, be available to others so that they can orchestrate a response nearest to where the attacker is, and also so that a coherent user experience results. Um, and it becomes very, you know, both, uh, difficult for defenders to manage that. And candidly creates a drain and demands a level of complexity to weave those systems together that open an integrated, uh, is just a superior pattern for them.
[00:10:51] Jason Clark: Brilliant answer. Love it. So principle three is focused on foundational technologies that integrate with your entire security ecosystem. So I'll start with Lamont, you know, you've had the opportunity to build a Greenfield security program. What does that, you know, when, when you look at that right. What was the first stack that you built? What were the five core kind of foundational technologies that were part of this Greenfield infrastructure?
[00:11:16] Lamont Orange: So I think that's a very important question because when you talk about the transformation that security must go through you, you have to look at it. Not only from what tools are in my stack, but it's what capabilities we want and make that more aligned to the outcome. So I'd say the first capability That I wanted. It is around visibility. I had to see what was happening and order to affect the risk level of the organization and be able to put controls around that and tools that help you to understand what is happening would, would be a tool that looks at your usage of even legacy applications, as well as SaaS, IaaS, and PaaS technologies, you also have to take into consideration your identities. Many companies are struggling with identity as they have several IDPs. They have several managers of those identities, whether they are production systems or, uh, corporate systems. And what you want to do is have some sort of governance around it. So identity was one of the other areas that I focused on. And then you look at data protection. Well, we're all in it to protect our data. That is the crown jewel for the bad actor. We have to understand what valuable data we have and what data we like to protect. And then you look at where is that going? So you want to understand the data protection from the user to app to end point. So you have to have back to a comment that James Robinson made about being open with XDR NDR and whatever we put in front of. That DR capability, you have to have that understanding around the configuration of that device and even the organizations that may be using that data. And then there's one other capability that I think is very important to help organizations scale. Uh, when they're looking at a Greenfield, you need something that's going to manage your configuration automation and orchestration. And I think those are solutions that can be kept in one, one area, but they need to have the following those three capabilities in order to be effective.
[00:13:29] Jason Clark: Erick, on, on this, uh, principle around kind of foundational technologies. You know, landed in the other organizations. And now as a CTO, what would any thoughts on this one from an ecosystem standpoint around security,
[00:13:43] Erick Rudiak: There's a couple of things. So identity is one. Um, and you know, I think about that in terms of kind of the various levels of assertions that people in systems can make about who the human is at the other end of the line or who the system is that another system is working with. So that it kind of encompasses everything from multi-factor to kind of directory services, like that's absolutely vital to get right. I agree with Lamont, uh, that having data protection, having visibility into kind of data at rest and data in motion. Is another, and then, uh, it nowadays it's not particularly exciting, but the, kind of the very basics of encryption and configuration management and, you know, in thinking about configuration management, kind of incorporating both, uh, kind of config drift golden builds, you know, system hardening as well as, uh, vulnerability management, which I consider an instance of that class. Kind of pulling that all together. Those are among the first places that I have looked myself for, the kind of assurance that the basic blocking and tackling that the outrage factor of, uh, getting those wrong is managed and minimized for any organization that I’m part of.
[00:15:04] Jason Clark: Okay. So principle four only buy cloud powered new technologies. Right? So essentially everything you buy should be cloud power to cloud enabled or cloud born. You know, there's a Gartner paper. That's. The future of security is in the cloud, as we were doing this tour, there was a, a sentence said that, you know, cloud is the perfect reset for security programs, right? Because you get to kind of start fresh and do things, right. What do you, when you, so overall for all of you, what curious what you all think about this, this principle and, and why we wrote it.
[00:15:34] James Robinson: I think the principle is good. I think for us to take advantage of and, and to meet, you know, one of the things that we say internally is meet your customer, where they're at, you know, if the customer is, is in the cloud or they're moving to the cloud, right, which is even better to be where they're going to be, you have to adopt this principle. It has to be a foundational component for anything that you're looking at. ou know? And, and if it's, you know, cloud assisted, you may look at it. For instance, cloud assisted EDR. Okay, get it right. It's got to have something on the end point. It can't just all be cloud, but you know, for, for anything that's on the network stack for anything that's in the application stack server stack, um, you know, anywhere in between, you know, it definitely has to, has to carry with this principle.
[00:16:17] James Christiansen: You know, uh, Jason, this is the third major transformation I've been through it in my career and, and, um, technology and security versus movement from mainframes to client server. Then from client server to cloud enabled system. And now with business Digitas, digitalization, you know, we see these transformations and I think, you know, the further you resist them and don't recognize them the further you drop behind. So as you start thinking about cloud and cloud enablement, I talked to many CISOs every year, I mean, three or four hundred. And when we talk about what their plans are, where they see as SASE came out or are now secure service edge, And you start looking at definitions. When I start working with an organization and start looking where they're at, they're already somewhere down the path. They may be further in the maybe far, uh, ladder down, but they're, they're already on that path, which just, it just says we're the most common security folks recognize this is the pattern to go to. This is the direction. The only new investments I see are really just renewables. Cause they, they can't get moved off quick enough to the new cloud enabled, uh, technologies. But I think, you know, from a people process and technology perspective, all three it's about training our people on how to work in these cloud technologies. It's building out the processes that support those and the things, and finally implementing the technology, to enable those people and enable those process to provide the level of controls we need nowadays. But certainly the business is going there. We've seen the acceleration of that movement in the business and us as security professionals.
[00:17:59] Jason Clark: I mean in the end, isn't it, every bit as businesses going there and therefore your security needs to be where the data is and where the businesses. And also whenever you have a mobile workforce, you need to leverage the cloud to be able to secure that workforce because you can't just do it from your data center, right? Like in its simplest form, you just changing a leverage point in your scale.
[00:18:18] James Robinson: I love what James said, because when you talked about, you know, moving from, you know, from mainframe to client server, you know, we saw that major shift. If anyone's part of it, I definitely saw it and was part of it as well. And watching that happen, I could not imagine if you didn't make that shift, or if you did make that shift and you still tried to maintain, remember how hard it was to maintain in client server, those controls that you had and, and deliver those via the mainframe. It was almost impossible to do. In fact, it was impossible, which is why everyone's shifted and you saw the market change.
[00:18:49] Jason Clark: So the next one is principle five and. It potentially has more weight than many of the others. It is protect business data with security controls that follow the data everywhere. Right. So, which essentially to me is, you know, this is, this is the grand strategy of security it's for many companies or organizations why security exists. Right. It's just protecting the information. So, you know, maybe let's talk a little bit about, we named this a principle for a reason, right. Because I think we felt that people were historically not mature enough in data protection. So, Erick, what's your thoughts around kind of why we established this as a principle and in how people need to transform from a, how they look at data security or data protection?
[00:19:34] Erick Rudiak: Yeah, I think there's two elements to this one, Jason, like one way to look at it is the security industry has had a history of negotiating against itself. On controls like this of saying, oh, well, this data used to be in our, you know, in our data center, um, you know, protected by perimeter controls, but we want to move it into the cloud and, you know, the cloud just doesn't have quite the same thing, but boy, the cloud is so compelling, so we accept less. Um, and so, uh, one way to look at this principle and they're not mutually exclude. Is never accept less like don't compromise on, you know, uh, on this kind of very basic idea that no matter where your data goes, Your controls matter, um, because they certainly, uh, the, you know, the attacker doesn't look at the world that way. And so that is part of the cost of doing business. That's part of the economics of, uh, you know, of handling sensitive data. And so the obligation to the person whose data you're, uh, you're trafficking in doesn't change. If you've allowed the data to move from one environment to another. I think another way to look at this one is that it's immensely aspirational. So I think any of us who have done threat modeling, we'll look at that guiding principle and say, oh, Jason, uh, Surely, uh, surely you can't protect the data when I do this to it when I use steganography or when I use covert channels or when I, you know, when I take advantage of, you know, pick your favorite row and column out of the MITRE attack framework. Um, and I think aspirationally, it's a challenge to us to say no to say. Uh, we are going to put controls in place that kind of reaffirm our obligations as custodians, that wherever the data goes, we are going to know about it. We're going to assert our controls, that you know, that the attacker doesn't get to win by using these common patterns.
[00:21:42] Jason Clark: Yeah. It's um, it's, uh, you know, fair. If we can get the controls to follow the user and the data that is Nirvana, especially whenever we don't have to invest heavily to do it, but love. Principle six, prioritize your business and risk reward when making security strategy right. Or decisions. Right. And building your tactics, et cetera. Right. So essentially as the business constantly changed, right. So it's always a constant trade off and we need to be in tuned to the business to be able to make that decision. Right. But enable them. And I love, you know, the saying. This isn't about eliminating friction, right? It's the right amount of friction to the business. So curious, what, what, uh, each of you, James, maybe we'll start with you, uh, Christiansen on, you know, your thoughts on, on that principle.
[00:22:29] James Christiansen: Yeah. This, this one for me is an important one because more successful CISOs. They're not focused on stopping data protected. They're focused on enabling the business and they understand the role of enabling the business to find that right balance between risk and reward. Right. And can't get too far skewed on either side or you're not, you're not doing the company and the stockholder value. You've got to live up to your role, but it's about enabling the business. And when I think about this in, in terms of change and how fast things change. I really think about, you know, the seven forces that are on a security strategy and these forces are really something you should be looking at because they're constantly changing. And as they change, you should be in tune with that change and think about how has that affect my stress. Has that really make a difference in what I was planning to do versus what I should do now. And the seven forces, we put out a white paper on it recently that I would encourage you to go read, but it seems like the culture of the company, a merger and acquisition might change. The culture might change in many ways. So how does that change the way you're looking at the economy and economic forces, government forces, conflicts, things that aren't directly related, but have these impacts on your overall security strategy. And as they change, you need to be agile as an organization. You need to be able to shift quickly and shift effectively to stay up with the business, because remember where we started this it's about enabling. The business and that's what you should always keep in mind.
[00:24:16] Jason Clark: So anybody else, you know James, Erick, and Lamont to add to that the prioritization of the business risk and reward.
[00:24:23] Erick Rudiak: So when I think about what's happened in the last year and a half, you know, we all woke up on March the 13th of last year and all of our annualized loss expectancy curves were garbage. Like they, uh, they were fundamentally flawed because there had been this injection of a completely new risk and totally new risk reward needs in terms of ALE. And so when I think about where our industry was 10 years ago, um, you know, a lot of us practitioners would say, you know, don't spend a thousand bucks, you know, to protect 10. And I think what has happened since then is our counterparts in the C-suite, the boards that govern us have become much more sophisticated. And so they expect us, as security practitioners, to understand business’ risk tolerance. And to come in with a narrative that says, look as your CISO, here's how closely our loss expectancy curves match our risk tolerance. And here's what needs to, you know, what we need to do in some places we may need to invest, in others we may actually have so much control that we can afford to take a little bit more risk, um, and you know, in adjust and be business positive and support additional capabilities to improve productivity. And so we've all, kind of through a forced Sentinel evet, had to learn that skill very quickly over the last year and a half, you know, the silver lining for us as security practitioners is, um, that skill is going to really come in handy for us. As we get back in front of our boards, as we get back into, uh, into the C-suite and discuss with our peers at the table. Not why we are saying yes or no to particular controls, but how our entire control environment matches our risk tolerance. And that's, you know, uh, just a maturation of the dialogue that we engage in and it's, you know, it's a really exciting one. It gives us a seat at the table.
[00:26:29] Jason Clark: Yeah, I think that it is, we look at it. We have to think about it again. It's just, we always say this, but as business leaders and, and how is this, what this activity, the business is trying to do, going to, for companies focused on, you know, acquiring, retaining profitable customers, right? Which is any way that's in the business of making money. And how are activities, the business doing, going to accelerate that right. And to what degree, what, how much do I doing need to protect them? Or can I let them go and then apply protections and, and just, just truly understanding the things that can move the needle for the business. And I mean, I, I personally deal with it every single day. And looking at marketing stuff and looking at security stuff. Lamont and I are in the conversations daily, right? Or about, about what's, uh, wanting to open up something that we want to do versus the balance of security of it. Right. And, and the marketing team wants to, you know, always wants to, wants to push forward, but we want to make sure that partner we're signing up to access our Salesforce information it's secure. Right. And it's all about the outcomes. And then how long do we have to. To apply the security controls. So I think it's, uh, it's not something that every security team's built with today, they think of things as well. There needs to be an exception. Who's going to sign off on the risk versus rationalizing it. Like you just said, Erick. So principle seven is build threat models and use them in every architecture decision you make. So Erick, you know, when you first heard James Robinson talk about threat models. At one point, I think you kind of came to me and said, uh, Jason Knight, I want to hire James Robinson away from you.
[00:28:08] Erick Rudiak: That sounds like me. Like I would have brazenly just said none for you all for me. That's totally on brand,
[00:28:16] Jason Clark: But we'd love your thoughts on principle seven here.
[00:28:21] Erick Rudiak:No, I mean, I think, uh, boy, it sure is hard to defend against something that you've never seen or thought about. And so like, everybody likes to use military analogies for this, but like I think about, so I've got a kid that loves to go to the neighborhood tennis courts with me and, you know, hit the ball around, um, in boy, uh, the first time that I threw a slice backhand at that kid, it was a predictably hilarious. Like they'd only ever seen the top spin shot and they just kind of swung and missed as soon as the ball landed and skidded away. And so I think that's, you know, that, that practice for our defenders and our defenders are everyone. Right? It's our infrastructure folks. It's our cloud, uh, engineering. It's our, uh, software engineers and everybody in between it. So user experience folks, it's our, you know, it's our database and middleware folks. They have to have seen both the top spin and the back spin in order to be able to effectively do that. And so that's, you know, that's why red teaming and attack simulation are such a vital part of any world-class, uh, information protection organization is because that practice is just so important.
[00:29:34] Jason Clark: Principal eight expand security operations automation. Right? So ultimately just a hyper-focus on automating your operations. So Lamont, can you maybe talk a little bit to why this principle song.
[00:29:46] Lamont Orange: I think it's a very important principle given that I think I spoke earlier about the increased velocity of attacks that we see. And I think it was also mentioned about some of the complexities that we have in our organization. So when you think of all of that, and you think of our, our teams, our skill, how skilled our teams are. And just the tooling resources that we have, we have to keep up at the speed of our attacker. And in some ways. We have to go and propel ourselves further and faster than the speed of our attacker is a better way to say it. And I think the way to do that is through automation. We have spent lots of time, I think, in the security industry, working with the security automation, tools and resources, formerly known as SOAR. We've spent some time with, with, uh, orchestration tools and confusing them as SOAR tools. And I think we, we really just need to take a step back and say automation is something that we should focus on when we're looking for scale and flexibility. And that will allow the velocity to be reduced to signal, to noise ratio, to be reduced. And then we further can react to these alerts that come through. Some of these things come in the form of having some of the foundational components built such as runbooks. And understanding how you are to execute a standard process throughout your SOC. That's something that you can automate from a maturity standpoint, but you can orchestrate a response in the very beginning, understanding how you triage. Triage is something that can be orchestrated, and then you can automate the response and control. When you talk about also automating and using orchestration within your SOC, you're also updating your incident response and management procedures. The incident management practices that we've used in the past don't necessarily scale very well to the attack, the type of threats that we have, the type of interaction that we have, have to have what I workforce, our suppliers and vendors. We've extended our attack surface, which I know we'll talk about in a different principle, but the extension of that attack surface is really something that we have to look at when we're thinking about how automation helps us again, achieve that scale and flexibility. I think there's another piece of this that we can't always run through our teams and say, well, it's contained into security. As long as security says, it's good, then we're fine. We also have to communicate what we're doing to our team members so that once they understand why we do some of the things we do, they will also be a willing person. And that's part of that transparency that I, that I believe that's part of automating and building out that, that SOC. And that's why it becomes the heartbeat of your security and response programs.
[00:32:54] Jason Clark : Oui. C'est drôle quand on regarde ces principes, n'est-ce pas ? Ils sont tous tellement liés les uns aux autres, n'est-ce pas ? Et dépendent les uns des autres pour être vraiment, vraiment bons. Avant de passer au neuvième principe, quelqu'un souhaite-t-il ajouter quelque chose sur les opérations de sécurité que Lamont.
[00:33:11] James Robinson : J'en ai une qui, euh, est plus une question pour elle ou pour tout le monde. L'une des choses auxquelles je pense de plus en plus, et je regarde certains de mes, euh, constructeurs, développeurs, ingénieurs, ingénieurs logiciels, frères, et j'ai travaillé avec tous les jeux et l'idée de runbooks et de playbooks et, et ainsi de suite, vous savez, en dehors de l'architecture et de la conception, vous savez, ou, ou l'une de ces choses, Je me demande si, où en êtes-vous en ce qui concerne l'automatisation, est-ce que vous exigez toujours que votre équipe construise les runbooks, les playbooks et ces différentes choses, et c'est ce qui régit et met en place l'automatisation pour la mise à l'échelle, ou est-ce que vous adoptez une approche différente ?
[00:33:52] Lamont Orange : C'est un débat intéressant. Honnêtement, je pense que très, très peu de programmes sont passés à l'automatisation et à l'automatisation fidèle. Pour revenir à ce que j'ai dit à propos de l'orchestration, je pense que l'orchestration sera l'exécution des tâches qui doivent se produire. Je pense que les runbooks, et j'utilise ce terme de manière très large, sont des outils de travail. Il peut s'agir d'un processus standard que vous utilisez pour répondre à un certain type d'incident. Ou un certain type de question posée à l'équipe de sécurité, une certaine exigence posée dans le cadre du développement ou avec n'importe laquelle de vos équipes auxquelles vous imposeriez des exigences si elles traitent certains types de données ou si elles ont certains types d'interactions avec le système, je pense qu'il est possible d'orchestrer ces réponses. Nous devons encore mûrir sur le plan de l'automatisation, car je ne suis pas sûr que tous les programmes fassent encore confiance à l'action automatisée. C'est là que réside le défi.
[00:34:53] Erick Rudiak : Je vais vous donner un point de vue d'ingénieur logiciel sur cette question. Lorsque je pense aux caractéristiques de performance d'un logiciel, c'est vrai. Il y a une sorte de, vous savez, en comp, comp sci, nous pouvons tous avoir appris sur, O , la notation, d'accord. Ou notation O ou N. J'ai donc eu récemment une excellente conversation avec notre RSSI qui m'a posé la question. Pour passer en revue certaines de nos vitesses et de nos flux, par exemple combien de dépôts GitHub avons-nous ? Combien d'opérations de validation sont effectuées par jour ? Vous savez, combien de fois, vous savez, combien de fois lançons-nous nos pipelines par jour ? Nous avons eu une grande conversation sur la question de savoir si nous étions capables d'adapter notre sécurité à O de N ou à O de log N. Et c'est là que l'automatisation est absolument vitale, vous savez, elle répond au défi du talent et à certains défis de la diversité dans la sécurité de l'information et l'automatisation est le meilleur moyen pour nous d'avancer. Pour faire passer nos contrôles d'un O de N à un O de log N ou même quelque chose de plus petit. La capacité à supprimer l'élément humain ou à le canaliser vers la résolution de problèmes plus complexes et plus intéressants est donc vitale. Je ne sais pas comment une entreprise peut faire face à l'ampleur de la demande. Tant du côté des attaquants, avec de nouvelles techniques et l'évolution des logiciels malveillants, des ransomwares, euh, des techniques d'attaque, etc., que de la capacité d'une organisation à étendre sa surface d'attaque grâce à des services en nuage faciles à fournir, etc. Dans le livre, nous parlons de l'idée qu'à l'époque, pour certains d'entre nous, nous connaissions nos serveurs par leur nom. C'est vrai. Ils étaient traités comme des animaux de compagnie. Ils étaient aimés. On s'attendait à ce qu'ils aient une très longue durée de vie, contrairement à l'idée que les serveurs d'aujourd'hui sont comme des briques de Lego, produites en masse et dont la durée de vie est, volontairement ou non, extrêmement courte. Et il n'y a tout simplement pas de moyen vraiment efficace de mettre en place une organisation de la sécurité qui réponde aux besoins actuels. qui n'ont pas d'automatismes, un principe de base
[00:37:10] Jason Clark : J'adore l'analogie, Erick. Et, euh, c'est une grande section du livre que tout le monde devrait, euh, devrait absolument lire. Je commencerai par James Christiania, puis je m'adresserai à Erick, car j'ai eu les meilleures conversations sur les risques avec vous deux. Mais le neuvième principe est d'exiger une visibilité et une évaluation des risques permanentes et chaque contrôle de sécurité, essentiellement en temps réel. L'évaluation des risques et les contrôles s'opposent à ces évaluations ponctuelles, une fois par an, des risques. James commence par vous.
[00:37:43] James Christiansen : Oui. Merci. Vous savez, c'est toujours un sujet brûlant pour moi et vous savez, je vais prendre quelques directions, mais je serai bref. Vous savez, l'idée d'une visibilité continue. Oui, c'est vrai. Le premier principe en matière de risque est que vous ne pouvez pas gérer les risques que vous ne connaissez pas, vous devez donc absolument obtenir ce risque ou cette visibilité. Et ce que nous avons vu, c'est un changement : nous sommes passés du client-serveur au nuage, puis à ce nouveau monde, et nous avons perdu cette visibilité. Nous avons perdu les capacités dont nous disposions auparavant. Et nous devons les récupérer. La visibilité est absolument essentielle, vous savez, et l'autre aspect de cette visibilité est la continuité. Nous entendons souvent parler de confiance zéro, mais c'est un terme dont je ne suis pas vraiment fan, principalement parce que, vous savez, avec les équipes de direction, il évoque la mauvaise connotation de Oh, vous êtes le responsable de la sécurité. Vous ne faites confiance à personne. Et il y avait un, vous savez, les gens de la sécurité, ils pensent à, vous savez, des jetons. Je m'en suis occupé. Eh bien, je préfère penser qu'il s'agit d'une confiance adaptative continue. Ou un programme de sécurité en constante adaptation qui examine automatiquement les risques et qui est capable de mesurer les différentes télémétries de risque et d'ajuster la réaction à ces risques. Vous savez, pourquoi avons-nous besoin du même niveau de contrôle ? Si, si quelqu'un va sur ESPN et, et vérifie leur, hum, sécurité ou leurs scores de football, comme ce serait le cas, s'ils vont regarder. Vous savez, le matériel pour le conseil d'administration avant la publication, vous savez, évidemment des niveaux de risque très, très différents. De nombreuses données télémétriques sont transmises par l'appareil, l'application, la localisation de l'individu, le trafic, et tout cela peut désormais être mesuré. Et c'est là l'essentiel pour les auditeurs. Il y a la capacité. Pouvoir le faire maintenant et prendre ces décisions adaptatives en temps réel sur le niveau de risque, cette voie de contrôle et être en mesure d'adapter ces contrôles. Je pense donc qu'il s'agit d'une période très stimulante pour pouvoir examiner ces risques et s'adapter au fur et à mesure que nous avançons. Oui, c'est vrai. Il y a toute une autre conversation sur les risques de tierce partie et les risques de quatrième partie que nous devrons garder pour un autre jour, mais c'est un autre grand sujet. Je vous laisse la parole, Jason,
[00:39:59] Erick Rudiak : J'essaie de trouver un moyen d'argumenter contre la visibilité constante. Et je n'arrive pas à en trouver un, euh, vous savez, euh, comme si je préférais avoir plus d'informations ou moins, si je préférais attraper les attaques et progresser plus tôt ou plus lentement ? Euh, est-ce que je préférerais un temps d'attente long ou un temps d'attente court ? Oui, c'est vrai. Je préférerais gagner. Et donc, vous savez, je pense à cette idée de visibilité continue, euh, peut, cela me ramène à Peter Sandman et aux communications de crise. Hum, et Peter, le marchand de sable donne une définition non orthodoxe du risque, qui est la somme du danger et de l'outrage. Et donc le facteur d'indignation de ne pas savoir. À l'ère moderne, nous avons la capacité de surveiller quelque chose en permanence et de ne pas choisir, de ne pas savoir qu'un contrôle est défaillant, de ne pas savoir qu'une configuration a changé, de ne pas savoir qu'un mauvais acteur s'est introduit dans nos systèmes. Le facteur d'indignation est tout simplement, hum, insoutenable
[00:41:09] Jason Clark : Donc, dernier, euh, dernier principe, nous savons ce que c'est. Il s'agit, encore une fois, des 10 qui nous ont semblé les plus importants pour préparer les gens à l'avenir. Nous pourrions probablement, vous savez, nous avons commencé par un certain nombre d'entre eux, nous en avions 20 au départ et nous les avons ramenés à 10, mais le 10ème est de réduire la surface d'attaque en utilisant les principes de la confiance zéro.
[00:41:31] Erick Rudiak : ] Je viens d'entendre que nous n'aimons pas la confiance zéro. Nous allons devoir le modifier.
[00:41:36] Jason Clark : Je pense que le secteur est confus en ce qui concerne la confiance zéro. N'est-ce pas ? Qu'est-ce que la confiance zéro ?
[00:41:43] Erick Rudiak : Pour moi, le truc, vous savez, le truc avec la confiance zéro, c'est que la confiance se dégrade avec le temps. Et donc, quand je pense à, euh, l'exemple le plus classique de cela pour moi est le délai d'inactivité de l'écran, comme le contrôle classique, il fait partie de PCI. Cela faisait partie du NIST et de pratiquement tous les cadres de sécurité depuis des décennies, où l'on faisait une approximation qui disait, oh, eh bien, si quelqu'un connaissait le mot de passe de Jason. Il l'a entré dans le terminal de Jason, puis dans le PC ou l'ordinateur portable de Jason et enfin dans le téléphone de Jason. Ensuite, on suppose que pendant un certain temps, il s'agit probablement encore de Jasonn. Euh, et, euh, cela défie la logique, euh, surtout quand on pense, vous savez, aux facteurs humains et à la conception et à la façon dont les gens travaillent réellement que, vous savez, 14,9 minutes après que ce mot de passe a été saisi, nous devrions avoir le même degré de confiance dans les données provenant de ce terminal, de ce PC, de ce téléphone, qu'au moment où Jason a saisi son mot de passe. Et donc, quand je pense à la confiance zéro et pourquoi. Pour moi, c'est parce qu'il reconnaît ce mensonge que l'industrie de la sécurité a toléré pendant des années, parce que la technologie était trop difficile, parce qu'elle était trop chère et que nous avons fait ce compromis. Nous avons répondu par l'affirmative. Nous faisons confiance à ce terme. Pendant 15 minutes au maximum, nous attribuons le même niveau de confiance aux données, aux paquets que l'appareil génère à la neuvième minute qu'à la première. Il est donc temps de passer à autre chose.
[00:43:23] Jason Clark : Vous savez, le débat sur la confiance zéro, c'est que, vous savez, il y a évidemment des vendeurs qui diront, oh, c'est tout ce qu'il y a à dire. C'est vrai, ou c'est une question de point final, ou je pense que ce n'est pas une chose binaire, ce n'est pas "on" ou "off". C'est vrai. Il y a des gens qui prétendent que la confiance est activée ou désactivée. Pour moi, c'est comme si vous pouviez entrer dans ma maison par la porte d'entrée et que je vous laissais entrer, mais cela ne veut pas dire que vous pouvez aller dans toutes les pièces. Cela ne veut pas dire que vous pouvez entrer dans mon coffre-fort.
[00:43:44] Erick Rudiak : Ooh, il va falloir qu'on parle de la dernière fois que je suis venu chez vous, Jason.
[00:43:49] Jason Clark : Et, vous savez, juste en fonction de votre comportement, n'est-ce pas ? Comme vous l'avez dit, depuis combien de temps vous êtes là, etc. Plus vous êtes ici depuis longtemps, plus je vous fais confiance parce que j'ai observé votre comportement. C'est vrai. Je pourrais donc dire le contraire, peut-être du point de vue de l'authentification, mais plus vous êtes ici, plus je vois votre comportement, plus je vous fréquente. Plus je vous laisse aller dans d'autres pièces. Je pense que ce que j'essaie de dire, c'est que pour moi, il y a une confiance dans un dispositif. Dois-je faire confiance à votre appareil et dans quelle mesure ? Et à ce moment précis, il s'agit en quelque sorte d'une évaluation des risques en temps réel. Est-ce que j'ai confiance dans le réseau que vous utilisez ? Le réseau Iran. Est-ce que je vous fais confiance en tant que personne à ce moment précis, sur la base de votre comportement, est-ce que je fais confiance à l'application que vous allez faire ? Parce que ce n'est pas seulement vous, c'est l'application, votre acte avec lequel vous interagissez et qui est potentiellement compromis ? Les données que vous essayez de télécharger, quel est leur impact ? C'est vrai. Pour moi, il s'agit d'une sorte de décision de contrôle conditionnel en temps réel, basée sur le risque, qui nous convient. Comme le Nirvana de la confiance zéro dans mon esprit, je vais soit vous donner zéro. C'est vrai. Je ne vous donne aucun accès ou je vous fais entièrement confiance. Appelons cela le niveau zéro et le niveau cinq, et ensuite il y a 1, 2, 3, 4, 5, d'accord. Pour moi, trois au milieu signifie que je vous donne accès aux données, mais qu'il s'agit d'une édition en ligne uniquement, et non d'une édition en ligne. C'est vrai. Il peut s'agir de plusieurs sauts supplémentaires au cours desquels vous rédigez un paragraphe expliquant pourquoi vous avez besoin de ces données et qu'il s'agit d'une urgence ou autre. Je considère donc cela comme une confiance zéro, comme l'endroit où nous devons nous rendre pour réduire la surface d'attaque, tout en permettant aux gens d'obtenir plus de confiance par leur comportement, si nécessaire. Ainsi, nous n'exerçons pas de pression sur les personnes occupées. Mais Lamont. Je sais que je suis curieux, car nous avons beaucoup parlé de la confiance zéro, mais la première partie de ce principe consistait à réduire la surface d'attaque en utilisant la confiance zéro. C'est vrai. Donc, vous savez, vous défendez votre organisation, n'est-ce pas ? Chaque jour, comment appliquez-vous ce principe ? Réduisons-nous les attaques ?
[00:45:46] Lamont Orange : Je pense que le premier point sur lequel nous devrions nous pencher, c'est la visibilité et l'analyse, la visibilité. Je suis désolé, l'analyse, parce que si vous ne pouvez pas la voir, il est difficile de l'appliquer et de la contrôler ou de la protéger, et ce sera la visibilité pour votre utilisateur, la visibilité pour votre infrastructure, pour vos données. Ensuite, vous pouvez commencer à faire le lien entre ces cas d'utilisation et ce qui est le plus important pour l'organisation à ce moment-là. J'aime les niveaux et la façon dont vous les avez décrits. Mais nous devons aussi comprendre quel est l'actif le plus important et nous, nous pouvons nous mettre d'accord et dire que ce sont les données, mais quelles données, quel ensemble de données et vous regardez ces contrôles à travers cette visibilité et cette capacité d'analyse pour appliquer ces contrôles à ce jour.
[00:46:32] Jason Clark : Je pense que nous avons également couvert les 10 principes. Merci d'avoir écouté les Visionnaires de la sécurité et soyez à l'affût des nouveaux épisodes qui paraîtront chaque semaine avec des professionnels de la sécurité de l'ensemble du secteur. Merci donc à Erick Lamont et à James pour le temps qu'ils nous ont accordé. Et, euh, c'était amusant
[00:46:50] Sponsor : Le podcast des visionnaires de la sécurité est alimenté par l'équipe de Netskope à la recherche de la bonne plateforme de sécurité dans le nuage pour faciliter votre parcours de transformation numérique. Le nuage de sécurité Netskope vous aide à connecter rapidement et en toute sécurité les utilisateurs directement à l'internet, à partir de n'importe quel appareil et de n'importe quelle application. Pour en savoir plus, consultez le site Netskope.com.
[00:47:11] Producteur : Merci d'avoir écouté Visionnaires de la sécurité. Prenez le temps d'évaluer l'émission et de la partager avec quelqu'un. Restez à l'écoute pour découvrir les nouveaux épisodes qui paraîtront toutes les deux semaines. Nous vous donnons rendez-vous au prochain épisode.
Pourquoi Netskope 
){kind=icon}
