Video: Finding Threats in Motion: How the Network Is a Missing Piece in Security Strategy | Duration: 2374s | Summary: Finding Threats in Motion: How the Network Is a Missing Piece in Security Strategy | Chapters: Webinar Introduction (0s), Polling the Audience (63.95199999999999s), Network Security Evolution (250.252s), Modern Attack Landscape (538.232s), Network Detection Response (851.8969999999999s), NDR Behavioral Analytics (1189.857s), Network Behavior Analysis (2109.742s), Advanced Alerting Discussion (2204.117s), NDR Solution Packaging (2269.367s), Closing Remarks (2332.302s)
Transcript for "Finding Threats in Motion: How the Network Is a Missing Piece in Security Strategy":
hello, everyone. Thank you for joining us today. I'm Tori, our vice president of marketing here at Thrive. Thank you so much for joining our webinar, finding threats in motion, how the network is a missing piece in security strategy. Our hope is that by the end of this webinar, you'll have a better understanding of the basics of network security, new types of attacks, and a new architecture for infrastructure security. So, let's get into a few things. We do have a q and a panel that is going to be open throughout the duration of the webinar. Please feel free to put those questions in at any time. We will be answering as many as we can at the end of the presentation. Any that we don't get to live today, we will follow-up with you via email on those questions. We will be providing the presentation and the recording after the fact to all registrants, so be on the lookout for that email to come, to you as well. And we'd like to kick off the the webinar today with a few poll questions. This is gonna help aid in the discussion for today. So I'm gonna go ahead and open the first polling question. So what is the average dwell time, time within the infrastructure than it's that an attack spends before being detected or launching an attack? Less than three days, five to twenty five days to twenty five days, 50 to a 100, or over 250? Alright. Give this a second for some votes to come through. 50 to a 100 is pulling away. Alright. Just one more second for this one, and I'll I'll open the next poll. Great. Okay. Our next polling question is, how many different operating environments, SaaS, public cloud data center, do you have across your entire IT infrastructure? Less than three, three to seven, seven to ten, ten plus, or you are not sure? Saw 10 plus hit first right out of the gate. Okay. Three to seven is is coming out strong. Mark, Kevin, is that kind of what you would expect to see, three to seven, or were you expecting something different? I think I think it's changing over time, and and what we've seen is is more and more different environments, more of a hybrid, Mhmm. compared to the past. Absolutely. K. One more second here, and then I'm gonna open the final question. Alright. Final question is, do you know the difference between north south and east west network traffic? Yes, no, or unsure? K. We got a quite a few yeses. Give this one more second here, and then we will get started. Okay. Thank you so much, everybody, for participating in these polls. So that's gonna help aid in the discussion today. And with that, I am going to hand this over to the actual experts on this topic. Kevin Landt, Thrive's vice president of product cybersecurity, and Mark Wojtasiak, who is, Vectra AI's sorry. Get to this slide. Vectra AI's vice president of product research and strategy. So with that, Mark, Kevin, please take it away. Thanks, Tori. So today, as as Tori mentioned, we'll be walking through, network security and talk about how the role of networks has changed, some of the emerging threats and and vulnerabilities that come along with those changes, and then a little bit more about how NDR or network detection and response can enhance things we're doing, today to protect the network. So a little look, at the very recent past here. This may look familiar to a lot of you. It's how networks have have typically been represented. There's a nice, clear, distinct boundary between internal network and external. We can secure that boundary with security tools like firewalls. And then once we've done that, the users and the devices that are inside that perimeter, we implicitly have have trusted those users. They've they've been authenticated. There's there's devices we know about and approve, and so we have less concerns about what's inside the network. Some terminology, a lot of you indicated you you know these terms already, but review for for those that haven't. Communication that's moving inside the network from one device to to another inside the network, that's known as east west network traffic. And then north south traffic is traffic, communication traffic that's moving from the internal network to the outside world or vice versa from the outside to the inside. So from a security perspective, a lot of our focus has been on that North South traffic. That's typically where we expect the threats to come from from from outside the network. So security is all about defense in-depth. At Thrive, we've put together a number of different security and technology layers to help protect the network. We offer managed firewalls and and VPNs so we can have those remote users or outside users authenticate into the network. Endpoint detection and response or EDR, that's something we added a number of years back when we started to see spikes in ransomware and other attacks that were targeting our our workstations and our servers. So EDR was a good replacement for that traditional antivirus that couldn't keep up with it. And then we offer managed detection and response where we've got a 24 by seven security operation center that's monitoring all of these things for us. They're looking at that North South traffic activity through the firewall logs. They're looking for suspicious logins, and we're looking at what's going on on the systems. And that gives us a good sign that that there might be an attack underway, gives us the tools we need to respond to it. Of course, nothing into technology stays the same. It's always evolving, and that's true of networking as well. So been a lot of changes to the network in recent years. The technology has changed. The architecture certainly has changed. As as you indicated in your in your poll responses, a lot of you now have SaaS applications. You have cloud infrastructure. You have remote users. So we don't have that distinct boundary anymore between what's inside and what's outside. It's really a hybrid. And so, of course, how we secure that and the tools we needed and the techniques, that's changing as well on the security side. We've seen attackers shift their focus a bit, and Mark gonna talk a lot more about that in in a minute. But, you know, we've had to respond to that. And and at Thrive, we've we've done that by we've started offering zero trust network access, for example. We can no longer assume that users that are inside the network, are are not risks to us. We can't do that implicit trust anymore. And we have to start looking at that East West traffic because now we've got attackers that might be in the environment, and we need to be able to spot them and and respond. So to do that, we've decided to partner with Vectra AI and and introduce network detection and response as a complimentary layer added to what we were already doing that we that I that I showed on that previous slide. So it's really just giving us better tools, giving our our SOC better visibility so that we can respond to emerging threats. So let me, hand it over to Mark, talk a little bit more about those threats and and how NDR works. Awesome. Thanks, Kevin. Hey, everybody. Great great to be here. I love this topic, in terms of and and kind of a a precursor, you know, we the slide emerging threats from the inside. One thing that, at Vectra AI, we I I love the poll question on hybrid. Right? And three to seven kind of domains that you're protecting. Mark will be more than that. That's that's how we think. Right? We we have a our CEO has a notion that enterprises will be hybrid forever. And one thing that we've realized over the course of the last several years is, like, how what does it take to to protect, an environment that is that is hybrid, that is constantly evolving, changing, etcetera, etcetera. And if you consider, you know, what's happening in the attack landscape, I look at these two. And and these two numbers, the one on the left, two hundred and ninety two days to identify and contain an attack is from IBM. And they track like, cost of a data breach report will highlight, okay, how many days does it does it take to identify and contain attack. And this number has gone up every year, for the past few years. Number on the right is from our friends at CrowdStrike. And this is the time it takes for an attacker to move from once they've infiltrated your environment, to begin moving laterally across that environment. Whether it is north, south, or east, west. Forty eight minutes. So think about the disparity between identifying containing an an attacker of two hundred ninety two days and the fact that attackers are moving even faster and faster and faster. And I'll talk a little bit about how they're doing that and and and and kind of the the the crux behind how we begin to think about approaching that need to move at speed. So the first thing we think about is like, how do these guys think? Right? We have an ocean hybrid enterprise hybrid attack. Right? Modern enterprise modern attack. How do these modern attackers think? Well, I think they think in, you know, much like any criminal. They've got a motive, they've got means, they've got opportunity and they've got a goal. They're motivated to circumvent controls. Right? Whether that's just, you know, disrupting, avoiding, evading, whatever it might be. They're gonna find a way to get into the environment. Their mode or their means is an identity. We have a notion that hackers don't or, attackers don't hack in, they log in. They assume an identity. They figure out and they're doing amazing things in terms of, like, narrowing down, the ability to compromise an account and seize an identity down to like one step. We see this with attackers like Scattered Spider. Right? Calling IT help desk, resetting, account credentials, getting escalated IT admin credentials and then they're on the network. Right? Which is their opportunity. Once they have an identity, they have the ability to move on the network. And this picture, I love the you know, a lot of you answered your most of you I think answered you have hybrid environments. This is what we call the hybrid network. Data center, campuses, OT IOT, remote workers, identity infrastructure in the cloud, multi cloud infrastructure and networks, SaaS applications like m three sixty five, and any number of additional applications that are all cloud connected or delivered in the cloud. This is what, you know, we argue is like the attacker doesn't see individual attack surfaces. They see one giant attack surface that they can move through on a network for the purposes of finding and breaching the crown jewels or causing disruption or interrupting supply chain or whatever they whatever their their objective or goal is. So this is this is how they fundamentally think. How do they operate? And they operate largely in the same way. Is they look at, you know, I wanna get if my objective is to get on the network, and I'm just pulling up my, a different visual over here. My objective is to get on the network and my path is to get an identity. With an identity, I can blend in. I can look like normal. Right? Different isn't bad. Normal isn't good. Right? They can look normal. Right? With an identity, I can move faster. I can escalate my privileges. I can get access to different parts of the environment. The faster I move, the less I'm seen. The less I'm seen, the further I get. The further I get, the more damage I can do. So this is how these modern attackers are operating, in inside the environment and they're doing this extremely fast forty eight minutes if you consider, the CrowdStrike stat. Now we have a belief that if the net if the attacker thinks, get on the network, That's where I can find what I'm looking for. Cause disruption. And my path is to get an identity. I don't need to deploy malware. I don't need to do whatever. I just need an identity. I just need to take over an account. Whether that account is a human or that account is a machine, it doesn't matter. I just need an identity to move on this network. Then why isn't a tax signal that we're delivering to the SOC or our partners like Thrive rooted in network and identity? Why aren't we thinking more like attackers think fundamentally to address the issue? Just keep pace. To understand where they are, what they're doing. How they're progressing. What behaviors are being exhibited by the attacker that are indicative of an attack in progress. Are those behaviors chained together to deliver, an accurate signal? And that's fundamentally what NDR does. NDR fundamentally, at least Vectra AI thinks like the attacker thinks. They're gonna get in. They're gonna circumvent. They're gonna get an identity and they're gonna move on the network for the purposes of getting to the crown jewel. So what does what is NDR's role? What is Vectra AI's role in understanding the attacker's means and their opportunity? What identities or accounts are they taking over, and what are they doing once they've taken over that account? How are they moving laterally? How are they, executing command and control? You call this the cyber kill chain. How are they what are they doing across the cyber kill chain to progress their attack? Fundamentally, Kevin brought up EDR, totally essential. And I think I'm a simple thinker. I'm like, okay. EDR fundamentally, and we know EDR is much more than this, but fundamentally the purpose is to stop the attackers from getting in. Right? See them. They're gonna attack an endpoint. They're gonna try and get into an endpoint. There's a number of different ways they get in, but the endpoint is their primary path to get in. And EDR's fundamental role is like stop them from getting in. But what happens when they're already in? And that's NDR's world. Once they're already in, assume compromise, zero trust, whatever you want to call it. If they're already in, who's watching what they're doing? Who can detect them? Who can string things together and understand what's happening once they're already in their environment? So that's these two things aren't mutually exclusive. In the day, in the world of the modern attacker, you need both. And both these technologies need to work together in order to keep pace with the attack. To see it, to stop it, etcetera. Now, what, you know, if if you look at the, what NDR covers that EDR might miss, there are a number of things that where NDR would fill a gap in the ability to detect an attack in progress. Right? Identity based attacks. Attackers that have seized an active directory account, an Entra ID account, whatever it might be, or they're inside M365. When attackers compromise an account and escalate privileges, NDR would see that. We would see that. Whether it's a human account or a machine account. What happens when we abuse privileges? Right? So privilege abuse. Attackers always wanna escalate their privileges. They wanna get an IT admin account or some account that has privileged access to the environment. EDR doesn't necessarily see that on the network. When attackers escalate abuse privilege, we'll see that even when malware isn't involved. We can see suspicious things happening in Entre ID or whatnot when privileges are escalated. We call it Privileged Access Analytics. It's a it's where we observe privileges and and behaviors of attackers who are privileged doing things that they shouldn't be doing with that privileged access. Lateral movement. East, west, north, south. Kevin alluded to this earlier. When and where Panels. We'll see that as well. When are they moving closer to the crown jewels? When is the right time for, and and and Thrive MDR service would would would understand and see, okay, when do I need to take action? Where has the attacker progressed to? When is it time to do isolation containment? Whatever response, action that that analyst needs to take, when they need to take it. They have enough information and context that yes, I need to act now. It is this is a real attack in progress and I need to stop it. So how does NDR complete your security, kind of strategy and defense against a modern attacker? And and and we think that we call this our our three c's. And I'm gonna break down each one of these and kind of, you know, the role of Vectra AI and NDR in this world. And that first C is fundamentally NDR is about coverage. Observability to indicators of attack across the cyber kill chain, across this hybrid environment. Vectra AI is largely based on behavioral indicators. So AI driven detections that understand attacker behavior and look for that behavior indicative of an attacker across the cyber kill train from infiltration all the way to exfiltration. What is the attacker doing? What behaviors would they exhibit? No matter the domain. The data center, the campus, cyber physical systems, OT, remote locations, identity, multi cloud, AWS, Azure, m three sixty five, SaaS. What behaviors are indicative of attack across those domains at different stages of an attack? Answering questions like who or what's on our network? How are they behaving? Are we exposed or currently under attack? It's continuously analyzing raw network packets to look for behaviors indicative of an attack using metadata within those networks, East, West, North, South, etc. To paint a picture of what's happening on the network right now. Bouncing what's happening off of what we've seen indicative of how an attacker might behave. What do you do with that coverage? So you think about coverage as like I have visibility to what's happening on my network. I've applied that visibility to understand what is indicative of an attacker behavior across the cyber kill train, across these domains. What does that visibility then turn into? We'll begin to like tie it to an identity or an account. This is where, you know, you take an indicator of attack or detection or detected behavior And it's like this behavior is associated with this account, this identity, this host. Oh, this behavior is also attributed to this host or identity or account. And this behavior, those behaviors or detections in correlated and grouped together is indicative of this attack profile. This is an active attack. This looks like Scattered Spider. This looks like Volt Typhoon. This looks like Salt Typhoon. This looks like Black Basta. And paint that picture to say, we are going to prioritize this host or account or this entity that is currently under attack or being used by the attacker to execute their attack. We give context an attack rating for it. Why is this entity prioritized? What context can we give you to understand that this is a real attack and we should take action? Where's the attacker going? How are they moving laterally across the environment? Looking for those abnormal behaviors inside the network that nothing else can see and stitching them together to deliver the clarity the analyst needs to understand and investigate that attack in progress. Coverage gives you the visibility. Clarity gives you the signal to take action. Control is the third and it's about taking that action. We call about we talk about attack resilience both pre and post compromise. Pre compromise would be, you know, exposure management and things like that. But resilience is just as important post compromise. Once the attacker is already in, how resilient are you? How fast can you move to contain and isolate that attack? There's applications and threat hunting. You know, obviously investigation and response. So control is about taking control of that attack. Answering fundamental questions for the analysts. When to mitigate and contain this. Sometimes you'll wanna wait to see if this attack manifests. Sometimes you'll want more information and contextualization to understand is this a real attack and do we need to do something now? That's what NDR is constantly feeding. When do I contain it? When do I isolate and mitigate? Where is there further remediation needed? There's tremendous amount of information and context in what's being seen on the network that can be used to improve security posture. To reduce exposure. The telemetry and the information that NDR delivers helps helps, understand where further remediation is needed. That's where you get into incident response and and other use cases like that. Ultimately, to to, measure and improve posture of the hybrid environment. The cloud posture, there's, you know I think of it as like one giant attack surface. What's our posture exposure to attack, across this entire hybrid environment? So we provide customers the ability to respond quickly to threats as they're unfolding in real time. And with Thrive, NDR, includes all kinds of different response and remediation activities. Host containment, traffic blocking, any number of ways the Thrive team can take action and contain and isolate that that attack based on the behavioral analytics and the signal that NDR is delivering to them. Coverage, clarity, control. You know, this is a fundamental belief that we have in in working with partners like Thrive. Ensuring the safety of an organization, requires coverage for attacks that EDR might miss. Okay. Removing uncertainty with signal clarity rooted in network and identity. Think like an attacker. The only way that we are going to be able to build resilience to modern attacks on a modern hybrid environment or modern hybrid enterprise is with signal clarity. That's rooted in how attackers think and how attackers operate. For the purposes of stopping the breach from happening long before it happens. Giving defenders, the Thrive team, the MDR team, control of that attack. And control not only to stop it, but the control to then further remediate it from happening again in the future. And I think that we're gonna open it up to q and a. Thank you, Mark. That's right. Great presentation, Mark and Kevin. Thank you. Yep. We're gonna start taking a few questions. So I will, reiterate because I did see it come through a couple times in the q and a. The presentation will be shared, with the recording as well. Alright. I'll I'll go ahead and grab one here. So someone asked, does Vectra use agents for remote entities? So, the the way Thrive is gonna be deploying the the the NER services is through a few different ways. So, for the on prem piece of the network, we will be providing a hardware appliance that that Thrive will manage, that can connect to, the SPAN port on a switch or to a tap if you've got a network tap. If you're not sure what these these things mean or stand stand for, our solutions architects will walk through that with you, get a get a good sense for for what you've got available in your network, how we should configure that. We can also do virtual sensors in cloud environments. So AWS, GCP, Azure, where we can collect their their virtual traffic mirroring, and feed that in as well. And then that identity piece that that Mark was emphasizing, we would collect that through API. So we can collect, API direct directly with with enter ID, Azure, AD, and and do that integration. So no need to go install individual agents. We would we would handle it through through that type of architecture. Here's another good one. How does NDR handle encrypted traffic? Mark, you wanna take that one? Because I know that's a the question that comes up, and it's sort of a difference between NDR and the more old fashioned IDS world. Yeah. It it's it there's different there's different, approaches, NDR approaches when it comes to encrypted traffic. Some NDRs need to decrypt in order to detect. Vectra AI doesn't. So because of our AI and our the behavioral models, we don't need to detect traffic in order or decrypt traffic in order to detect an attacker behavior. So, you know, encrypted, nonencrypted traffic, it doesn't matter to Vectra AI, but we don't need it to decrypt it to detect. Alright. What else we got here? Feel feel free to grab any. You you see there, Mark? Yeah. What kind of analytics or patterns can show that an attack is in progress? We have, we have a 170, more than a 170 now, behavioral based detections. I talked about across the cyber kill chain. So from, like, initial access all the way to exfiltration, you know, some there's a there's a MITRE attack framework for this. There's, like, cyber kill chain standard like, industry standard stuff. Like, an access and we and we I talked about how, you know, you can chain we chain things together across the cyber kill chain or behaviors together. So we see things like, suspect compromise access, TOR activity, unusual scripting engines, suspicious OAuth. So we'll see like access. Behavioral indicators of access. What the attacker might do to persist. Right? To kinda move. Account manipulation, admin account creation. There's a trusted IP change. There's another thing. Then they assume command and control, the third step. We'll see different things like hidden HTP tunnels, novel external ports. We might match to a threat intel feed and then escalate and evade. Like those certain behaviors an attacker will do to do that and then move into recon and discovering assets. So like we look at behaviors. We have 170 of these detections, behavioral detections across this kill chain. And then we begin to map those behaviors to again to an identity or an entity, an account. And then bounce that off of like 14 different attack profiles that we've seen to say, Oh, this set of behaviors that we're analyzing in real time, near real time, this set of behaviors is indicative of this type of attack. Right? So. you can almost know where they're going next, based on what they've exhibited thus far. And sometimes this can take, you know, minutes. Sometimes an attacker will move slow. Low and slow. But we're constantly kind of piecing those pieces together, to send a signal to you guys so that your analysts can take action on it. Right. And and a big difference from from, intrusion detection systems or IDS where those were really those were really rule based. You had to write a you had to write a pattern to detect. each different type of attack. And so you're always playing keep up with attackers because, obviously, they're gonna change their their patterns. Yes. Yeah. Whereas the the the NDR approach, because it's behavioral, we don't necessarily need to know the exact attack pattern. We can start to piece together these different behaviors and recognize they're using techniques that an attacker would use, not that a normal user would would would use. And so, you know, if it's a zero day or some new technique, we've pretty good chance we're gonna pick up on it because we're gonna identify those behaviors that we know their motive, we know their objectives, so we can see. that they're working towards it in the environment even if they're doing it at a slow drip Yeah. over time. Yeah. And then and it's it's like when and when we see a new attack technique, right, or MITRE releases one or whatever it might be, we see them real time in terms of, like, based on customer feedback, based on that sort of thing. We'll roll out a detection in a matter of a couple weeks. The ability to detect that behavior and it's constant. Right? So it's like we're trying to keep pace with new techniques we're seeing on the network, in the cloud, in m three sixty five, in identity or with an identity, and keeping those behavioral based detections real time. So there isn't a need to write a new rule. Right? And. and, and which, you know, that's a huge challenge for detection engineers. It's like, oh, I got all kinds of threat intel coming in. I got to write rules. I'm seeing this behavior. Here's a new technique I need to write a rule for. I got to maintain that rule. Oh, this rule is creating too much alert noise. I got to tune the rule. That's all gone with behavioral based AI detection. Right? And triage. And and the stitching and correlating of all those different behaviors. It's all about how accurate and fast is the signal that we're delivering to the SOC. The faster and more accurate it is, the more confident the action the Thrive team Thrive team can take. Exactly. Alright. We've got one asking about how NDR compares to other network security tools. They mentioned firewalls, VPNs. And and we just talked about how it's a a bit different from from IDS. Firewalls, really meant to to enforce that that boundary, whether it's between two segments in the in the network or with that external network. It's really gonna be looking at the source, the destination, the port, checking my rules. Is that allowed? It's not really, let's say, smart when it when it comes to those. It it's really just just checking its access control lists and either allowing or denying. NDR is a little bit more what what we're looking for is those clues to the behavior. So that's how it would be different than than a firewall. You know, really really not, equivalent to a VPN VPN, different purpose, allowing users to to access network resources securely from from outside the network. But we would wanna look at the behavior of those users after they're they authenticate because we have seen at Thrive, one of the vectors for for attacks lately has been compromising somebody's VPN credentials. I get their VPN password. I can log in, and now I've got access to to move around the network wherever I want. And I look just like them. gonna detect, Right. but. something that. that in. the air. Yeah. Yeah. Think about a think about compromised IT admin account, which ScatterSpider is very good at and known for, is the behaviors they begin to exhibit, to other network technologies may look completely normal. Right? And NDR's role is, like, string things together. It's like we've seen this behavior, this behavior, this behavior. This is indicative of scattered spiders, you know, operandi. This is how they operate. It's like, this is real. Now send a signal to the Thrive team. This is real. This is this we think this with this degree of confidence, a score of 80 or 90, we have prioritized this entity. You guys go in. Here's all the context. You guys go and determine if you need to take action. You can't do that with firewall logs in real time. Right. You can't feed a bunch of firewall logs to a SIEM and try and do analytics and run queries, you can't do it at the speed of forty eight minutes. Completely agree. Alright. Looks like we've got one last one here. And I don't know that we'll be able to answer this one on the call, but, the the question is, any use of tagging or advanced alerting notifications for in house AI services and tools? I guess what I'll all I'll say is the the way, that NDR is being implemented at Thrive, we will have, alerts that create cases in our our Thrive client portal, when there's a Vectra AI detection. That's something our SOC will work, and you'll have access to those tickets. We do have some data, sharing tools that we can make available. So that's something we could we could speak with you maybe offline to get a little bit more information about what you wanna accomplish by by feeding those into an AI. And we could talk about how much of that data is could be made available to you and how, but, we could absolutely have that conversation. Another one came just came through, Kevin, Oh, and I. think that'll. be our last one. Alright. Sounds good. So how is NDR solution packaged commercially, identity based detections or agents devices. So, the way it's packaged is really around, the size of the network, when it comes to active concurrent IPs. So what we wanna see is, how many how many devices are on the network that are generating traffic over a time period, and then there's, you know, there's some some averaging and calculation to determine that. But the the general thought here is the larger the network, the more signals we're gonna be taking in, whether those are identity or. traffic based metadata. So that's really gonna drive how much, we detect, how much our SOC is responding to, and that's kinda how we we we size things. So hopefully that that answers your question. Alright. Great. That was going to be our last question. If there were any more that we missed, we will, like I said, follow-up via email. So I do wanna thank everyone that joined today. We do appreciate you spending this time with us and hope that you've taken away some really useful information. And I hope you have a great rest of your day. Thank you again, Kevin and Mark, for the great presentation. Thanks, Awesome. everyone, Thanks, for. joining. Thanks, Tori..