Security Posture is a Thing
With Chris Romeo, Izar Tarandach, and Matt Coles
What is security posture? Izar was at a conference in Amsterdam, where he was asked to define security posture and how to measure it.
About this episode
What is security posture? Izar was at a conference in Amsterdam, where he was asked to define security posture and how to measure it. Is security posture qualitative or quantitative, and can it be compared across teams, organizations, and departments? This led us down this rabbit hole; what is security posture, and is it even possible to measure?
Security posture is multi-dimensional, differentiating between organizational and system security postures. Security activities that are reasonable to a company's level of risk acceptance are essential. Leadership changes could impact security posture; the departure of a CISO, for example, doesn't immediately affect the security posture as the policies and experiences built up over time remain.
Tools and processes assess security posture. An organization's security posture doesn't necessarily reflect the system's security posture. You must understand where a design is starting regarding security and where it is now.
The episode concludes with a call to listeners to share their thoughts on security posture and contribute to the ongoing discussion. The hosts express their interest in learning from different perspectives and experiences in security.
Follow AI Security Table:
➜ Home: https://securitytable.ai/
➜ X: https://x.com/AISecurityTable
➜ LinkedIn: https://www.linkedin.com/company/ai-security-table/
➜ YouTube: https://www.youtube.com/@AISecurityTable
Transcript · 45 min conversation
0:09Chris Romeo Hey folks, welcome to another episode of the Security Table. This is Chris Romeo, joined by my friends Matt Coles and Izar Terandash. And the topic for this week is security posture. And so I'm going to turn it over to Izar and let him explain what he means when he says security posture, kind of set the stage for this conversation so that we can get rolling. So, Izar.
0:36Izar Tarandach Okay, so let me wash philosophical here for a second. A couple of episodes ago, we spoke about how we don't know everything. And I actually just met the other side of that coin. And the fact that there are some things that we, me at least, speaking about myself, take almost as axiomatic and think that everybody knows what I'm talking about and that everybody agrees that that's the definition and that it's such a basic thing that nobody should actually raise a discussion about it. But a couple of weeks ago, I was in Amsterdam, had a lot of fun in the GOTO Amsterdam 2023 conference. And I had the opportunity of meeting a PhD student of one of the Netherlands universities and a very, very smart person. Apparently they only have that kind of flavor over there, very smart people doing PhDs and teaching. And that actually posed a challenge to me after my presentation. It was something like, okay, so you're talking about this security posture. Thing over and over and over again. What, what do you mean by it? And how do you measure it? Is it qualitative or quantitative? If it's quantitative, it's some— is it something that you can compare between teams, organizations, departments, whatevers? And that got me thinking, you know what, I don't know. And Usually it's like security posture. Sometimes you talk about reactive, about proactive, about detective. I've always thought about aggressive security posture or defensive security posture, and I really wanted to hear where you guys are at with that and, uh, if that makes sense.
2:39Chris Romeo Yeah, I think it makes sense. I mean, let's think about what is security posture first, right? Before we even think about how do you measure it, let's see if we're all on the same page about what we even think of as security posture. And so, from my perspective, when I think about posture, it's my stance. It's how am I approaching a given— how I'm securing something. So, my security posture is, How good of a job am I doing? It's my approach. It's how I'm doing it. Now, Matt, for you, does security posture mean the same thing or does it mean something different?
3:21Matt Coles I think very close. I think I'm— well, I'm going to take the compliance road a bit here. I don't mean to, but I think about it in terms of residual risk, right, and defensive capability. I look at it in those terms, I guess, right? So, security posture, I think you said it right, is what's your stance? What's your— not necessarily something that you're going to— you as a— we have to put ourselves in, I guess, in our system's shoes here for a moment or for organization's shoes for a moment, right? It's not something that our organization is going to go out and do, but the ability to resist intrusion. And so, what are my— what risks do I have? And what is my defensive ability to resist those intrusions? That's what I think of when I think of posture.
4:14Chris Romeo And we can put a qualifier on it too, right? So, I can say that institution has a weak security posture or that institution has a strong security posture, but it's not really like a qualitative— I'm not I'm not— it's, it's more like a gut feeling when I say that. I don't, I don't have, I don't have metrics to say that it's, well, they're an 8 out of 10, which means they're strong.
4:43Matt Coles It's, it's actually interesting. You could have a measure, you could measure, you can measure things that contribute to posture, but there is nothing different. There's no current yardstick that I'm, at least that I'm aware of, that says you're strong or you're weak, right? Because there's no, there's no universal standard for that today, right? And that's why, that's why I look towards risk. I can, you can, I can tell you how much risk I have. I can tell you what, what risks, risks I'm worried about, what threats I'm worried about, what threats I know I can't defend against. And that would contribute to saying I'm strong or weak. Now that may be more of a subjective feeling because I don't know if I'm strong or weak compared to you or Isak.
5:28Izar Tarandach Wait, before we go comparing different organizations, the way that you put it, so we have a collection of metrics, right, that express a lot of stuff. By the nature of math, I, math, not Matt, I want to believe, I want to believe that there is a way to bring all those numbers together into one big number. Right? So that one big number is the quantitative, let's call it score for the lack of a better word, for that specific thing, place, organization, whatever that we're measuring. My gut feeling says that if I put that score measured over some period of time and I start looking at the slope of the graph that connects all those data points. If it's going up, then I say, hey, this place has a positive security posture. They're getting better over time by these metrics. And if it's going down, then they have less than positive. I won't call it negative, but less than positive.
6:35Matt Coles But that's great for over time. But how do you know where you're at in any given time?
6:40Izar Tarandach So that's a great question. Does it matter where I am at any given time or does it only matter over time? Is this something that makes sense to measure right now or something that to be understood and to actually have some value as a measurement, it has to happen over time?
6:55Chris Romeo It matters if you don't want to get compromised.
6:57Matt Coles Well, if you don't want to get somebody who's going to—
6:59Izar Tarandach is my posture a measurement of not being compromised?
7:02Chris Romeo You could be improving. If you have a weak security posture, then you are susceptible to some types of attacks. And we know that From experience, if you put a Windows box unprotected on the internet and unpatched, it'll be compromised in less than 60 seconds.
7:21Izar Tarandach But wait, wait, wait, wait, wait. Is that something that factors into the security posture calculus? Because what happens if I am in an organization and they have an amazing security posture? The CISO just left to go be amazing somewhere else. And the next guy in line is a bit of an idiot. What's my security posture right now?
7:44Chris Romeo I was good yesterday. It is a measurement. It's a measurement of where you are at that moment. Yes. But you're not going to— just because the CISO left doesn't mean that they erased all the policies and all the experience that have been built up over a period of time. Yes, it could go on a downward trajectory, but the CISO walks out the door. The security team doesn't just fall apart and die. They continue operating, the SOC keeps looking at things, cloud security people keep enforcing the policies of IAM and access control, and you know, you don't break in one second.
8:20Matt Coles And this applies whether you're talking about a system, a product, or a component, an organization, a nation, right? Presidents change all the time. The national security infrastructure doesn't stop working just because you get a new president. Same thing with a CISO, same thing with a new kernel. You update your kernel, it doesn't necessarily change your security posture immediately. Well, maybe it's a bad example, but, you know, an OS, you may do an OS update, it doesn't necessarily break everything all at once. So your posture changing, there may be a measurable effect. You could potentially measure a change in your posture, but it's not like you're redoing everything just because a person or a component change if there's a good underlying capability. Now, it's interesting, you said, you said improving or, or improve— improving or decreasing, or, you know, increasing or decreasing posture. If you start from zero, you can go up 100% from zero and still be Swiss cheese, right? So there is a notion that you—
9:23Izar Tarandach in order—
9:23Matt Coles and in fact, in order to know that you're improving, you have to be able to say, where was I and where am I? Right, so you have to be able to measure what your posture is at any given time in order to know if you're—
9:33Izar Tarandach so it is a time function—
9:35Matt Coles unimproving. I think there is a way to put it into a time function, but I think you, you have to be able to measure at any given time. You have to be able to say— you have to, you have to be able to say, I have certain characteristics that I'm looking at that tell me I have a posture And that posture then can be measured and then measured over time.
9:57Chris Romeo And until you reach a base level, you don't really care about the measurement. If you're in the basement of your security and privacy program deployments, meaning you don't have any protections, it doesn't matter if you're a 0 or a 2, you're still in big trouble. You still have to put together teams that are going to go fix things and and build new processes and new improvements, add new technology to get yourself to that base level. So I would say measurement is important, but going from, you know, 1 to 2 on a scale of 100 doesn't matter. If 35 is our base level, meaning if you're below 35, you're in trouble, measurement doesn't matter. But now I do want to measure how long it took me to get from 0 to 35. How long does it take me to get from 35 to 100 over a period of time? But I'm just saying there's a minimum bar that gets drawn across somewhere that says, we don't care until we get to a minimum level.
10:59Matt Coles Now, so what's really interesting about this, just thinking about this logically, right? There's this famous, you know, quote, and I'm going to misquote it here, but, you know, a former executive leader in the United States said, there's a set of known knowns and known unknowns and unknown unknowns and whatnot. Well, At some point, you may have a posture that you don't or cannot measure that maybe actually—
11:27Izar Tarandach okay, okay, no, no, wait, wait, wait, now we're right.
11:34Matt Coles Well, so, okay, so if you— I guess if you deploy a— if you, if you have a system and you don't have it connected to the internet, your posture is pretty good. But maybe you can't measure it because you can't actually touch the system, right? So you don't know what its actual capabilities are, what its actual defensive options are.
12:00Izar Tarandach But it's not going to be—
12:01Chris Romeo Wait, wait, wait, wait.
12:02Izar Tarandach I have to clarify this for me and for the listeners, if we have any. So now the security posture is the characteristics of a single system, or are you using that system as a—
12:14Matt Coles I'm just using, I guess, I'm using it as an example, but also you could measure a posture of a thing, a collection of things, an organization, a person, a collection of people, an organization, a system, a collection of systems, the collection of all of those things, all the way. I mean, it's, it's, uh, a posture does not exist for just a single entity. Or the collection of things in that entity. I think you can measure across all those.
12:47Izar Tarandach Is it a measurement of that security posture, the place where we want to go? Or is it a stepping stone to get there? Meaning, once I can measure posture, can I say I have an understanding of my security at this point in time?
13:08Matt Coles I think if you can measure it, so if you have things to measure and you measure them, you can make certain claims or have certain understanding about your posture. You could potentially guess or predict those as well. So, there's direct measurement, like, what—
13:27Izar Tarandach and if it's—
13:27Matt Coles this is where— this is why I think, still think, uh, risk is a good way of describing this, right? Or threat, threats. Let's use threats because we all know threat modeling, right? If I'm— if I can I measure what threats I have, that I think I'm going to have, and I measure, do I actually have defenses against them, I can make certain— have certain understanding or make certain claims about my posture. I'm resilient to attack or resilient to these threats.
13:53Izar Tarandach But you just measured residual risk.
13:57Matt Coles I did, yes. But I could also predict that, right? Oh, we're going to implement new features. I can predict where I'm going to be without measuring it. But I only have confidence in things I measure. I don't have confidence in the things I predict. So I think you, when you measure something, when you, when you measure in this context, when you measure it, you can know what your posture is, you can understand what your residual risk is, but you can also look ahead and predict what it is going to be as a means of making change. Maybe that's a future concept for us to talk about. If we're still stuck on the concept of what is a security posture, uh, yeah, I'm pretty stuck there because we're talking about measuring things.
14:49Izar Tarandach But let's go back and say, okay, what's the equation to get to my security posture right now?
14:58Matt Coles I, I don't think that that's a universal— there's no universal equation for that. So I don't think. I think it's—
15:04Izar Tarandach so, so let's, let me, let me throw an option here.
15:08Matt Coles So if we use risk as an example, you may use, uh, some arbitrary scale that you're like, well, you may use the risk, risk management framework from NIST as your means of assessing or analyzing or understanding risk and, and recording risk. You know, it's a 0 to 5, there's 0 to 5 measurements on various things and you have harms against the organization. Whatever. Or you may use the FAIR model, right, to calculate a dollar amount of risk. Both are valid. Both inform, or rather, are informed by your security posture, or may actually express your security posture.
15:48Izar Tarandach Will those measurements always agree with each other?
15:54Chris Romeo I think we're crossing— I don't think we're mixing multiple things up. I'm not convinced on this line of argument that says risk is the way to get to posture. I think risk is above posture.
16:12Izar Tarandach The way you address risk to me is an indicator of what your posture is. And, that's why I used aggressive, defensive, reactive, detective. Yeah.
16:25Matt Coles The way you address risk or the reason you have for everyone to have risk?
16:28Chris Romeo No, the way you address risk is to have risk. Yeah, I mean, there's risk everywhere. Maybe we need an example. We need to work through something. We need to put a scenario out here on the ether and into the ether and try to describe what's the security posture.
16:49Izar Tarandach It's starting to feel to me like One of those cases of, I don't know what it is, but when I see it, I recognize it.
16:55Chris Romeo That's the expert angle of us being security professionals with decades of experience though, right? Like you show me a network architecture, you show any of us a network architecture, within 5 minutes we'll get a feeling for the security posture of that environment because we're looking for things.
17:15Izar Tarandach But that's where the Right.
17:17Matt Coles And what are we looking for?
17:18Chris Romeo So, let's use— let's pretend— let me define the scenario. Okay. Let's talk about we all have been in and we've been around startups and small, you know, new companies that are building something new and they're deploying it to the world. So, let's just imagine we've got a 1-year or some number of years early-stage startup. They're deploying their application in the cloud. It's a standard backend API written in TypeScript and a JavaScript frontend for whatever their SaaS-based solution is that they're providing. So, they come to us and they say, "Izar and Matt, as wise security professionals, do an assessment of our security posture here." So, what's the first thing that we do? Like, let's figure out if somebody said that to us, Hey, look, can you assess our security posture? What would you ask them for first? What would be the thing you would say, show me something? What would it be first?
18:26Matt Coles Your architecture. We're going to want to see your architecture.
18:27Izar Tarandach I want to see it.
18:28Matt Coles That's number one. We want to see what the architecture is.
18:31Izar Tarandach I want to see their AppSec. Well, no, no, no. I'm assuming that there's a need for an AppSec function. And we discussed last week that there isn't always.
18:39Chris Romeo Yeah. And I would say in that early stage startup, there's, they're not going to be an AppSec person. They're not going to have an AppSec team. Yeah. Right. It's, let's say they're, they're a a 10-person or 20-person startup who's mostly engineering. So, they've got high-end senior developers that are cranking out and building a vision for a new product.
18:56Izar Tarandach Oh my goodness.
18:58Matt Coles The first thing I'm going to want to see is what are their inputs and outputs.
19:03Chris Romeo On the network architecture level though.
19:05Matt Coles No, no, no, no, no. Of the app or the— build the app and the network.
19:10Chris Romeo Is there a distributed—
19:11Matt Coles How the components stitch together and where the data's going.
19:14Izar Tarandach I would just see the list. Of security activities that they are putting in place that is reasonable to their level of risk acceptance. Yeah, there's that word again.
19:29Matt Coles They're going to laugh you out of the room when you ask that question.
19:32Izar Tarandach No, they won't. Because, well, if they do, it's not going to be the first time, but they won't. Because again, we're coming out of the reason, reasonability thing that that startup can only do so many things. And that to me as a security, in this example case, consultant, it's going to indicate that their readiness, their eagerness, eager, how eager they are to do security. And that's going to reflect to me their security posture.
20:10Chris Romeo Let me answer your question. I'll pretend I'm on the other side of the table. I'm representing the startup being assessed here. Okay, so, you know, our VC people told us that security was important and we needed to do some things for it. So, we wrote an incident response plan based on one that we found on the internet. We added a software composition analysis open-source tool. We found an open-source tool that works for JavaScript, so we added that to our build pipelines. We're running npm audit in the build pipelines to at least alert us and break the build if there's any known security vulnerabilities. And we're looking at this open-source static analysis tool that we think we can put into the pipeline with minimal effort. So that's how we approach security so far.
21:01Izar Tarandach So that to me tells me that they took all, most many reasonable activities at the level that they are right now to get them to a better place in a security journey than they were yesterday. So to me, that indicates a positive and growing, and I would even call aggressive security posture because they are actually doing things.
21:30Matt Coles Can I ask a—
21:32Chris Romeo Definitely.
21:32Matt Coles Can I ask a follow-up question? So, to you, as the leaders of this company, so you implement all these tools, what are you doing with those tools?
21:41Izar Tarandach Oh, yeah, I assumed that. That's a very good question. Yeah.
21:44Chris Romeo So, we break the build with NPM. So, if NPM audit runs and throws some type of a dependency vulnerability, that does break the build. The individual engineer will then chase that down, whoever committed the PR. Trying to commit the next PR, we'll just grab the issue and fix it because their PR is not going to go through the pipeline. We do have a filtering capability in case there's some that aren't working. There's some that they don't have— we don't have fixes for yet. That's in the JavaScript infrastructure. Same thing with that we're looking at with the SaaS tool is we, you know, we'll, we'll, we'll break the build at the individual PR level for any new code that's been added. So that engineer would then have to go back and fix it before the PR will be accepted.
22:32Izar Tarandach So Chris, you described the what to do with the results of all the tools and processes you put in place, as Matt pointed out. And then my next question is, under what circumstances are you willing to let go of doing those things? Meaning, when is something else more important than those security measures that you are taking?
22:56Chris Romeo Yeah, I would say, since I'm not a security person, I would say that, you know, we do, we do, we are considering both the business risk to the market, in the marketplace, and the security risk involved. And so we have it, we're trying to do the best thing that we can for security. But we're also a small startup who, if we lose our innovative edge, will go out of business and be gone in 3 months or 6 months from now.
23:28Matt Coles Do you know what you're trying to defend against?
23:33Chris Romeo Bad attackers on the internet is our primary focus of what we're trying. We're trying to prevent— we're trying to protect our customers is our primary focus because we know that we have sensitive data from our customers on the financial side. And so we, we, we don't, we don't really have a good view of our adversaries other than everybody on the internet who wants to steal our customer data.
24:00Izar Tarandach So you said to me, it keeps mounting and mounting to something that feels like a positive security posture. I have no idea how you're making that judgment call because the tools are going into place. The results of the tools are being used. They are— they have a limit, an appetite that says, I'm willing to wear risk because of this is my business interest. And they are doing this proactively by actually discussing and keeping things in mind, like there are bad people out there who want what we have and we have to defend against that. So to me, that sounds like people who are posturing themselves in a positive manner. So let me ask you— As you pointed out correctly, if they just had the tools but they weren't doing anything with the result, then that would not give me the same gut feeling.
25:01Matt Coles So let me ask you a clarifying— I guess let me throw out a point here. Are we assessing the risk, the security posture of the people in the organization and the business, or the system that they're building? Because what you— what was described about how they run tools and they break builds and etc., um, to me doesn't give me an understanding of what their system's security posture is. It tells me what their organization posture is, but not what their system security posture is.
25:39Chris Romeo So it sounds like there's multiple pieces then, is what you're—
25:42Izar Tarandach I'm going to sort of disagree there because they are showing me processes and tools that they are putting into place. Whatever the security status of the system is at any given point, those tools and processes are going to give some result that they already shown that they are willing to take into consideration and act upon. So, in my head, that plays as the quality of the security of the system is going to grow over time.
26:15Matt Coles I would agree with you, but we don't know where they're starting from and where they're at now. So, they could be—
26:21Izar Tarandach I assume that they started from zero.
26:23Matt Coles Well, but they could have— Yes, they certainly did. But so, they're going to improve over time from zero, but we don't know where they are because we don't know code coverage. We don't know what rules they're running. We don't know what their ability of the developers are to fix the issues. They can make the tool shut up, but that doesn't mean that they can be secure.
26:40Izar Tarandach True. But see, that's the thing. They said that they are taking care of the results. That lets me assume that they are not shutting the tool up. But you are right. You're completely right. And here you are putting together the quantitative meters that I was asking, can we sum them up and get to a security posture quantitative measurement?
27:05Matt Coles But I think it's really important that we have dimensions of posture. Organization versus system are just two that we're exploring here.
27:14Chris Romeo Well, they're two inputs to the same thing though, right? Like, I wouldn't describe an organization's system security posture different than I would what their— how their people are approaching it that are building that system. It's all one answer.
27:29Izar Tarandach I think it's both. Yeah.
27:31Matt Coles I don't think it can be one answer. Well, it depends. It depends. I think it needs to depend on— there's other information you need there, right? If I'm a product company, I don't deploy my product at my company. I deploy my product at somebody else's company. So, so I may measure two different postures.
27:50Chris Romeo But posture then is also your attitude, I think, is where Izar was leading us here. Like, he was measuring my responses based on my attitude towards security. And a strong, you could say, for a startup that has a limited amount of money, they have a limited amount of time, my attitude, my positive attitude towards security was a demonstrator that I have a greater than, You know, I'm somewhere on the positive side of my security posture.
28:23Matt Coles I don't disagree with you. And what I will throw out there is now I think we're reinforcing that there's subjective and non-subjective measurements.
28:31Izar Tarandach Yes.
28:32Matt Coles For posture.
28:34Izar Tarandach And that takes me to the next step. Let's assume for a second that we can put a number to security posture measurement. Because of the nature of all the different variables that we have been talking about for the last couple of minutes, would, would there be any sense then to compare security postures between different organizations?
29:00Chris Romeo Yeah.
29:01Izar Tarandach And that's where I'm going to go and say, personally, I don't see it.
29:07Matt Coles I think I can see it in one area, and I'm going to throw out another term I've talked about in the past. Assurance or confidence. And actually, confidence is the word I want to use, not assurance.
29:17Izar Tarandach Yeah, yeah.
29:18Matt Coles The subjective measurement that you asked about, like, what's your thought process, what processes you have in place, how do you handle results, that gives confidence in the response. Meaning whatever posture we might measure, we have confidence that that is a good measurement and you're on an upward trend. And with that, I can, I can compare Company A and Company B, or, you know, Org A, Org B, of is one giving more confidence than the other. That doesn't tell me what their posture explicitly is. That doesn't tell me whether they're defend— they can defend against the latest malware ransomware attacks. But we can say one is improving and the other one is maybe not improving.
30:02Izar Tarandach And how is that different from having a number of consultants show up in a number of different companies, query them about all kinds of practices that they have along all kinds of measurement axes, and draw a radar graph and say, this company is more security-minded, security whatever, than this company in this vertical, and end up with an, with an ordered list of—
30:33Matt Coles yeah, the spider chart is a, is a wonderful thing. In this context.
30:40Izar Tarandach So how, how is that different? Do, do all of a sudden—
30:45Matt Coles I don't know, but there's a scale that they end up— in order to get that, in order to get that post, in order to get that assessment, they have a consistent scale. They ask the same questions of different companies and they get results and they measure those results on a similar scale.
31:00Izar Tarandach But do you agree with me that doing the same things in a 5-people startup and in a 5,000 global organization. The, the, the, the sheer idea of comparing those things in the same scale is— makes no sense to me.
31:24Matt Coles So organizationally, maybe I would agree with you. System-wise, I think it wouldn't matter. Right? The technology, for any, for like technology, they're going to have similar problems and similar, similar ways of addressing it. The number of people in that, in my, in my mind, the number of people involved, if, if you're deploying a Kubernetes system, you know, Kubernetes to the cloud and, and you're using it, you know, on, on AWS or Azure or whatever, it doesn't matter if it's a 5-person startup or a 5,000-person multinational. Your abilities to manage it change as an organization, and the processes and all the places where you have to look may change, but the technology doesn't change. The capabilities of the technology don't change.
32:13Izar Tarandach Static code analysis, you mentioned coverage, how much coverage do they have in those things that they put in place? The 5-person startup, by nature, is going to have to possibly be easier to get a higher coverage than the 5,000-person organization.
32:29Matt Coles How do you—
32:30Izar Tarandach how—
32:31Matt Coles we don't know what their lines of code are. We don't know what their code complexity is. We don't know how they manage their source code.
32:37Izar Tarandach They are 5 people trying to coordinate one single activity against 5,000, right?
32:44Chris Romeo Nobody goes into a startup and says, we're going to have 15 products across 5 product lines and we're going to get it all up by Christmas. Yeah, they go out of— I mean, you know, so it's, it's very much a focus of the 5-person startup is going to have a more simple world.
33:06Matt Coles So again, let's be careful here about making— I mean, what you're saying is because they were successful, they got 5,000 people, and that means they most likely branched out from the thing that they were building in the first place.
33:19Izar Tarandach No, that's not what I'm saying. What I'm saying is somehow they survived.
33:24Chris Romeo I mean, that's basically the— I mean, tell me a company that didn't go through that life cycle.
33:30Matt Coles I mean, SpaceX built one thing.
33:35Izar Tarandach Yeah, but they don't care if things explode.
33:37Matt Coles I mean, they went from 5 people, but now they build—
33:41Chris Romeo but now they've taken that one first rocket that they built and turned it into 5 different product lines of different rockets and payload trailers and all kinds of other things. So it doesn't— it's not the fact that they look at the camera and smile, so it's not like they're worried about it. I mean, it's, it's, it— they, they didn't start by building all of those lines of, of rockets that they do today. They started with building an engine, I think, if I remember correctly. I think that was the first thing they built was an engine that, um, somebody else could, could use. That was—
34:12Matt Coles they built a flamethrower.
34:14Chris Romeo That's the Boring Company, isn't it? Or the flame—
34:15Matt Coles I don't remember.
34:17Izar Tarandach But anyway, that's the Boring Company.
34:18Matt Coles So, so what you're suggesting is a company grows because they diversify, not because because they just need more people for building what they build.
34:27Chris Romeo I mean, either that single product grows in complexity, you know, let's use Salesforce as an example, right? Where are we going with this? Salesforce, I have no idea where I'm going with this. I'm just talking.
34:39Izar Tarandach What about the security posture? We haven't answered that question yet.
34:42Chris Romeo Yeah, that's true. I mean, it's—
34:44Matt Coles No, no, but it was important because you called out, well, a smaller company will have an easier time managing their codebase. And what I'm suggesting is, Whether you have 5 people or 5,000 people, if for like codebases, it won't matter.
34:59Chris Romeo Yeah, but there's no way it's gonna be a like codebase. You can't have— it's not gonna be a like codebase. If you have the same codebase in a 5,000-person company that I do in a 5-person company—
35:08Matt Coles Maybe you're dedicating everything to sales. Maybe those 5,000 people are selling.
35:12Izar Tarandach No, no, no, no, Matt, Matt, Matt, Matt, Matt. Let's assume that we have a startup that's just got like 30 million lines of code, codebase from somewhere else, and they are building a product out of that. Okay, but they have 5 people. Now let's say that we have a 5,000-people organization with 100,000 lines of code.
35:31Matt Coles They're not my codebases.
35:33Izar Tarandach Wait, wait, wait, wait, wait, wait, and 3,000 developers. The fact is that coordinating security between 5 people is always going to be easier than coordinating security with 3,000 people.
35:46Chris Romeo I agree with you. Because you can say, hey, everybody, come here for a second. Come on.
35:49Izar Tarandach Exactly.
35:50Chris Romeo Listen, listen real quick. I need everybody to hear this. With a 3,000-person company, you need 10 people who just communicate. Like, that's their job is to ensure everybody gets the word.
36:01Matt Coles Yeah, I agree with you. But now you're talking different systems, different codebases, different postures.
36:06Izar Tarandach No, different sizes only. Like, I just don't want to get the assumption that the startup, for being a startup, has like this, this pizzi one. One repo, 10 lines of Python thing. I'm saying it's brilliant, but it's a people person now.
36:24Matt Coles But I mean, posture. So, posture, again, posture of the organization versus posture of the thing that's being built, right? If they're both building databases, the posture of the database won't matter how many people they have.
36:39Izar Tarandach Okay, so let me ask this differently. If you're taking this down to the level of products, In the one company that has 250 products, they end up with 250 postures. Do they have one posture? Do they have one posture that reflects the 250?
36:59Chris Romeo I think so. That's— I mean, ultimately, okay, case in point, I'll use the historical case because why not go into the past? 2004, William Gates sends the Trustworthy Computing memo. Yeah, right. As a, as a result of a lack of security across everything Microsoft was doing at that point, that's the one flashpoint that, that generates a positive step forward for them as a— and so you would, in those days, you might— I mean, we're all old enough to remember, like, Microsoft security was a joke. Like if someone said Microsoft and security in the same, but Microsoft was probably making something at that time that was perfectly adequate. There was some product somewhere that its security was actually okay. But I would say, you know, earlier 2004 and earlier, Microsoft security posture was junk.
38:00Matt Coles But I think, but that was an overall consideration. Think about it though. Windows 3.1 versus Access versus Excel, each of those had their own security posture, that the collection of those together—
38:13Chris Romeo sure, was— but I'm saying the collection is what people knew. Nobody knew about Access. I didn't even think about Access in those days as far as its security posture. I just knew, and we as an industry said up to that point, Microsoft security has been bad. And Bill Gates said it because he wrote the Trustworthy Computing memo. We got SDL, we got threat modeling, we got everything that came from that.
38:36Matt Coles But that was the perception. But the perception was driven by real-world information. That real-world information drew a perception both of like from their customers as well as internal, and they needed to address that, right? Because their perception was driving their bet, was driving poor sales.
38:53Chris Romeo I mean, maybe that's— maybe there's a lesson there that security posture is really about perception and attention.
39:00Izar Tarandach And then, then it's a subjective thing.
39:02Matt Coles All the way.
39:05Izar Tarandach And if it's perception, then it definitely doesn't make sense comparing among different things that you are perceiving.
39:14Matt Coles I don't know if I could go out there. That's a bit of a stretch, I think. I think there are portions that matter. Perception drives risk in a different way, but to the organization. But the posture drives behavior in another way with respect to attacker behavior, right? If I have a product that has the Swiss cheese on the network, I'm going to get attacked more. I may drive a perception of bad security, but I can actually measure I have bad security because I have wide open APIs and I have, you know, plaintext encryption keys, whatnot, right? So my posture can be measured, and then the effect of a bad posture can be predicted or understood. Symptoms were—
40:01Izar Tarandach and corrected.
40:02Matt Coles And corrected, yeah. Perception is a lot harder to correct. People, people get that view of all Microsoft security was bad. Well, it took them 20 years for people to go, okay, I trust it now, right?
40:17Chris Romeo Yeah. And they've— yeah, I mean, think about where they are now. Like, it's— nobody questions Microsoft security now. It's, it's a gold standard in the industry.
40:25Matt Coles Except for the latest release of, uh, how China backdoored everything, and you know, it's just all, all bad, right?
40:33Izar Tarandach So who could possibly do that?
40:38Chris Romeo So I don't know if we answered your question after all.
40:41Izar Tarandach I, I don't think so. I don't think so. And the more I think of it, it— I, I go back to the fact that it feels like we, we can't really, uh, define it. We can say I know what it is when I, when I see it, but we, we are not currently able to define it. And we keep using that, we keep using the term, and I don't think that it means what we think it means.
41:04Matt Coles Well, out of the three of us, we have— out of the three of us, we have three different, three different opinions of what a posture— what posture is and whether you can measure it and compare.
41:17Chris Romeo So I think if we added more people, we would just add more opinions and more disagreement as far as what it is.
41:24Matt Coles But you want to take this to a LinkedIn poll?
41:26Chris Romeo I mean, let's, uh, I think let's leave it, let's leave it where we are right now. I think we, I think we, I think we, the exercise was good though. I think, I think we didn't have the same conclusion. I wouldn't have had the same conclusion when we started as where we've landed here as far as, yeah, it's actually pretty difficult. To define what it is, to compare these things amongst different entities and different size companies and whatnot. So, I think we learned something on the pathway of just going down and exploring the example and everything. So, I think we made something good for the world.
42:02Izar Tarandach Oh my God. We didn't see it, but NIST has a definition of security posture.
42:10Matt Coles You know, you could have started with that, right?
42:12Chris Romeo You could have done your own research here. Yeah.
42:14Matt Coles Come on.
42:16Izar Tarandach We didn't Google things. Damn, we didn't Google, we didn't ask ChatGPT.
42:21Chris Romeo Read what it is if you've got it in front of you. We'll end with that.
42:25Izar Tarandach The security status of an enterprise network information and systems based on information security resources, people, hardware, software policies, and capabilities in place to manage the defense of the enterprise and to react as the situation changes. People, I don't think that we We're very, very far away from that.
42:43Chris Romeo Yeah, but that's a very infosec-flavored example.
42:47Izar Tarandach It's the NIST.
42:49Chris Romeo And that's, yeah. So, I mean, we could tear that apart too and say what's— it's not perfect. It's not a perfect definition based on— but the funny thing about this type of an exercise is the three of us are reflecting a definition off the decades of experience we have. We're not just reciting a textbook. We are remembering, our brains are processing all of the things we've seen and all the times we've come in and looked at something and gone, "Oh, the security in this sucks." How do you know? "I can just tell by looking at it because this is here." Well, no, because you don't have this, you don't have this, you don't have that, you don't have that. All of these things on your architecture are blank. You forgot all the things that provide security and build confidence in the architecture. And so, I think that's where we can leave it.
43:41Izar Tarandach Yeah. And the final most important question here is, after you gave your example and Matt and I assessed it, where do we send the invoice?
43:51Chris Romeo That would be on your Linux box. There is a special file called /dev/null, a file descriptor. That you can send any invoice you ever have for me directly to that. And I promise we will answer within 1 billion days. That is our guarantee to you. 1 billion days, we'll get back to you. So folks, thanks for listening to Security Table. We hope you— we hope you got something out of this journey that we just went on here. And tell you what, if you have other thoughts on security posture, put your comments on the LinkedIn post. That you use to find this episode. And let's, let's keep— let's get your side of it too. Feel free to send that to us anytime you want. We'd love to get other perspectives. Once again, other people have had different experiences than we had and come from different areas of security. And it's good. We can all learn from each other along the way. So thanks for listening to the Security Table.
Transcript supplied by AssemblyAI. It may contain transcription errors.
View the original transcript on BuzzsproutOriginally published as The Security Table. Part of the AI Security Table archive.
