Learning Library

← Back to Library

Proactive Threat Hunting Before the Boom

Key Points

  • “Left of boom” refers to the pre‑attack reconnaissance phase, while “right of boom” covers post‑attack recovery, highlighting the need to consider both before and after an incident.
  • Current industry metrics show a mean time to identify (MTID) of ~200 days and a mean time to contain (MTTC) of ~70 days, meaning organizations often spend roughly 270 days from breach to full recovery.
  • Reducing the gap between the actual breach (“boom”) and detection requires proactive threat hunting that can spot malicious activity before alarms fire.
  • Threat hunting relies on forming hypotheses and leveraging indicators of compromise (IOCs), indicators of attack, and external threat intelligence feeds to identify suspicious behavior early.
  • Unlike reactive investigations, threat hunting is a forward‑looking, hypothesis‑driven process aimed at detecting and mitigating threats during the reconnaissance phase.

Full Transcript

# Proactive Threat Hunting Before the Boom **Source:** [https://www.youtube.com/watch?v=VNp35Uw_bSM](https://www.youtube.com/watch?v=VNp35Uw_bSM) **Duration:** 00:06:41 ## Summary - “Left of boom” refers to the pre‑attack reconnaissance phase, while “right of boom” covers post‑attack recovery, highlighting the need to consider both before and after an incident. - Current industry metrics show a mean time to identify (MTID) of ~200 days and a mean time to contain (MTTC) of ~70 days, meaning organizations often spend roughly 270 days from breach to full recovery. - Reducing the gap between the actual breach (“boom”) and detection requires proactive threat hunting that can spot malicious activity before alarms fire. - Threat hunting relies on forming hypotheses and leveraging indicators of compromise (IOCs), indicators of attack, and external threat intelligence feeds to identify suspicious behavior early. - Unlike reactive investigations, threat hunting is a forward‑looking, hypothesis‑driven process aimed at detecting and mitigating threats during the reconnaissance phase. ## Sections - [00:00:00](https://www.youtube.com/watch?v=VNp35Uw_bSM&t=0s) **Left and Right of Boom** - The speaker explains the pre‑attack reconnaissance ("left of boom") and post‑attack recovery ("right of boom"), highlighting the lengthy detection (≈200 days) and containment (≈70 days) periods that together total roughly 270 days. ## Full Transcript
0:00boom 0:03you've just been attacked the bad guys 0:05in your system and your hemorrhaging 0:07data now this was not an isolated 0:10incident there was a time that led up to 0:12this and there was a time that happened 0:14after the time that led up to it going 0:17back and preceding this we call left of 0:19boom and the time after it when we're 0:22trying to discover that's the time that 0:25we call right of Boom so in this first 0:28phase we're basically doing the bad guy 0:31is doing reconnaissance he's looking at 0:34your systems he's probing them he's 0:36trying to figure out where the weak 0:37spots are where the good data is and the 0:40time after boom is when we want to be 0:43doing recovery 0:46the problem is we don't always know when 0:49boom occurred at the time it occurred 0:51there's a lag between boom and alarm 0:55this is when we find out 0:59when the attack has occurred which is 1:02not the same as when it actually 1:03occurred 1:04now if we look at these kinds of of 1:06intervals what could we do about this 1:09well there's a time here you can see 1:11that is the mean time to identify what 1:16the problem is according to poneman's 1:19Institute survey in 2022 the mean time 1:23to identify is on the order of 200 days 1:27that's how long it takes to figure out 1:29that the bad guy is already in your 1:32system 1:33and further still there's this notion of 1:36mean time to contain and the mean time 1:40to contain is on the order of 70 days 1:43you put those two together you're at 1:45about 270 days total between when boom 1:49occurred and when now we finally got 1:51everything recovered 1:53there clearly would be some Advantage if 1:56we could go back in time and discover 1:58and boom and bring the interval between 2:01the alarm and the boom closer together 2:03or even go back before boom and start to 2:07realize during the reconnaissance phase 2:08that something is amiss something is 2:11happening in this phase and this 2:14timeline in this area is when we could 2:17be doing 2:18what we call threat hunting I could do a 2:21hunt and we'll talk about what that 2:23involves after the Hunt is when we start 2:25doing the investigation 2:27and the investigation is reactive 2:31The Hunt is proactive that's the big 2:35difference here so as much as possible 2:36it would be nice if we could go back and 2:38anticipate some of these things 2:40now how does this work well it turns out 2:43we're going to take if we're doing 2:44threat hunting we're basically going to 2:47develop a hypothesis you have someone 2:50that's acting as an investigator but 2:52they're investigating things that maybe 2:54have not fully happened yet or in the 2:56process of happening so no alarms have 2:58gone off what are we going to use well 3:01we're going to use things like 3:03indicators of compromise 3:05that's information that we can gather 3:08from our systems that tell us that 3:10someone has done something here that has 3:13breached maybe it's a small hole here a 3:15small hole there are other things like 3:18indicators of attack that may be in here 3:20as well that says someone is knocking on 3:22the door they might not have fully 3:24gotten in but they did some things that 3:26we should be concerned about 3:28other sources of information would be 3:31security intelligence threat feeds so 3:33Intel feeds that tell us that there are 3:37certain types of vulnerabilities that 3:39are being exploited on the internet 3:41these days and certain things that are 3:44happening maybe in a particular sector 3:45or in a particular geography this is 3:48security intelligence we'd like to use 3:50that kind of information and leverage it 3:51to our best ability as well then we 3:54could do some vulnerability scans on our 3:57environment and that would tell us what 3:59the bad guy is also seeing and that is 4:02that there are vulnerabilities here 4:04there are systems that are weak here 4:05this is the soft underbelly this is what 4:08the bad guys doing why wouldn't we do 4:10the same thing 4:12then an intelligent threat Hunter who 4:15has some experience and some insight 4:17into this is going to use basically 4:19their experience and a little bit of 4:21intuition to 4:23connect the dots and say these things 4:26that seemed isolated in fact are not 4:30they're all grouped together we connect 4:32the dots and we realize before the boom 4:35has occurred that in fact this is the 4:38setup that's about to result in a data 4:41breach or it could be in this range of 4:43time when some things maybe have we've 4:47already been attacked but the alarm 4:48hasn't gone off but we're going to see 4:50the telltale signs and those are the 4:52things that we're looking for now what 4:54would an experienced threat Hunter use 4:56in terms of the tools to gather all of 4:58this information well they would use 5:00some tools like an xdr 5:04extended detection and response 5:06capability and I've talked about this in 5:08a previous video another thing that we'd 5:11use is a security information and event 5:13management system something that goes 5:15out and gathers information from all of 5:17my sources of security Telemetry all of 5:20the logs all of the flow data all of the 5:23different sources that might give me 5:24security intelligence and these two in 5:27fact can be related in feeding to each 5:29other and then another related 5:30technology that often relates to these 5:33as well is a user Behavior analytics 5:35capability and the uba would look for 5:39anomalous user activities here's a user 5:41that's performing differently than what 5:43their peer group is doing and therefore 5:45they're drawing suspicion and then we 5:47would ultimately want to take all of 5:49those Technologies and Infuse them with 5:52artificial intelligence so that we can 5:54get to the point find the source get to 5:59our investigation faster and do the 6:01reporting and hopefully a uncover this 6:05issue before the boom and if not at 6:08least very soon after the boom because 6:10what we're trying to do is avoid two 6:13numbers 6:14there's one number 270 and another 6:17number 4 million what do those mean 270 6:21is roughly the number of days between 6:23boom and containment according to the 6:26ponymon Institute also according to that 6:29same survey 4 million is the average 6:31cost of a data breach that's what we're 6:34trying to avoid 6:38foreign