91. Your Security Strategy Needs Graphs – Tech Field Day Podcast

In the modern security landscape, lists are not the way to get things done. You have to start thinking like the people that you're defending against, and that means lots and lots of graphs in this episode of the Tech Field Day Podcast, security Needs Graphs. Welcome to the Tech Field Day podcast, where we bring together a group of it technical experts to discuss a single idea about key concepts in the industry.

This podcast features a variety of perspectives from members of the Tech Field Day delegate community, and is often associated with one of our events. Tech Field Day is a part of the future and group, and this podcast is also published on our sister company's website, tech Strong tv. On this episode, we're gonna be discussing security specifically around Microsoft, but before we jump into that, lemme introduce our guests for the day, starting with Gerard.

What's going on? Everyone? I hope everyone had a great holiday out there in the tech world and in the real one.

Gerard Kalina here. Uh, I'm a network and security engineer for Aqueduct Technologies, and I'm also the founder and creator of Tech House five seven. Oh, you can find me on LinkedIn, Twitter, TikTok, wherever books are sold.

Super excited to be here. Great to be here. So happy to, uh, be a part of this story.

I think at this point, uh, you would say I'm the Chief Product Officer at Nexus Tech. I write a lot of things. org and I'm looking forward to the discussion.

Alright, and my name is Tom Hollingsworth. I'm the event lead for all things related to security here at Tech Field Day. Let's jump into today's episode.

You probably had an opportunity to listen to our special tech Field Day exclusive event with Microsoft back in October where we talked about some of the exciting new updates that were made to Microsoft Sentinel. But as we get closer to the end of 2025 and now into 20 26, 1 of the things that we wanted to do is kind of revisit some of those ideas because I think that there's a lot bubbling under the surface that people really need to understand. And the premise for this episode is that security needs to think in graphs.

Before we jump into that though, I wanna talk a little bit about the data part of this, and you're, you're thinking to yourself, well, Tom, you just talked about why graphs are important. Hold that thought, because we have to build a foundation before we get there. One of the things that Sentinel has decided to do, I'm sorry, Microsoft has decided to do with Sentinel, is they want to build a giant data lake for, uh, products to be able to reference whenever they're pulling data and things like that.

And this was one of those things that I think was mentioned in the event that kind of flew under the radar, is they really are building what I consider to be a, an ocean sized data lake, a great data lake, if you will, because one of the things that Microsoft talked about was the fact that they wanted to be able to let users and, and organizations keep data for up to 14 years in this data lake. I wanna give everybody a frame of reference for that, if you will. 14 years ago, I got my first Mac, it was a MacBook error.

I finally convinced the boss to lemme buy one. If I could keep 14 years of data in a data lake, it would literally be everything I've ever created for this job and defend some, and I can search through all of that to find data. Like is this a thing that's important for people in the, the industry?

Like you guys are the ability to keep all of this data for relatively low cost? From what I understand, I mean, I wanna jump in because I think what's really cool about it is, right, as, as an IT professional and working at different layers, you know, from help desk network, you know, infrastructure, all, all nine yards, like when you have data, how you treat data is imperative. Especially like, for example, working in a healthcare organization that's kind of numero uno like data and reten retention re excuse me, re re pretending data, re data retention is, is key and imperative, especially when you have records, a lot of important documents, things of that nature.

So, you know, back then we didn't have a solution like Sentinel where it was clean, very streamlined. It's like, we're gonna build these, metaphorically speaking four major walls to compound and streamline data, give you much, you know, deeper granular visibility, long-term retention, and the analytics piece, which is really cool. So I think that working from different, you know, sectors of it, it's really cool to see how we're evolving with data.

I think it's certain in certain areas or certain sectors like healthcare and, and, and, you know, maybe, you know, f financial institutions, that's kind of imperative. Other ones maybe not so much, but I do think that this is a solution that's, that's, that's really key and, and kind of growing and doing that and how we, how we start storing it, how being able to process it and more so being able to sift through the logs and, and really check to see like what's important, what's not. I think of it like this.

You've, you've, if you go back that far in time, uh, you're spanning multiple jump points on my career timeline. And in the early days, I remember when a two terabyte hard drive was, you know, pretty amazing that you could do that, uh, that many years ago. Um, then, then this progression of like, well, that actually might become how much memory you have access to.

And, and then when you think about the retention part of it, I was thinking about, well, your retention policies, um, very common for people to want to discard, uh, you know, data. And sometimes it was, uh, just to make sure it's not part of a discovery process, you know, if it's something related to like, you know, objects, uh, related to documents, things like that. But in the world of security, uh, the ability, like when you say, uh, one of those advanced system threats, you know, slow and low, it's been cooking.

Maybe it's been in there for, you know, the better part of 10 years. Um, there, there are obviously some interesting stories in that, but I will always go back to what's the inherent cost of that storage over that period of time. Um, every CIO event I've ever been a part of, uh, certainly the precipitous drop in storage pricing has not been top of mind for CIOs.

If anything, it's the ongoing increasing amount of cost to keep track of storage, even with the wonders of cloud, uh, which makes it super easy. Um, which is another thing too. I I think an important part of this is that this is an inherently, you know, Microsoft story.

Um, but I'm very curious to see how the data lake part of it fits into the, oh, you don't need to move your data. You don't need to rehydrate, dehydrate, move your data, move being a four letter word. So how I actually, uh, how it's actually executed is gonna be interesting.

'cause I I I would also doubt that over that many years, the day would all be in one place. Yeah. And that's the problem that we've got right now with a lot of this information is the data lake looks a lot more like a bunch of little puddles where we hope that we've got the right data.

We think we do, but where is it? Well, it's in this shelf over here, it's in that drawer over there. It's like when I try to look for things in my office, and I think that the value of putting it all together is not just that you have a good base to draw from, but it provides opportunities for exploration and honestly for attribution.

'cause one of the things that we've seen with some of these, uh, modern in 2025, several of the CBEs that were released with like high severities, these were not new problems in a lot of cases. They were problems that had persisted in certain versions of packages for months or in some cases a couple of years. And you don't know how far back that goes if you don't have access to that data.

Because we really do live in a, in a, uh, an enterprise IT culture now where if it's not immediately useful to me, I need to chunk it out. Um, there was a news story at the end of last year. Somebody discovered a running version of Unix system four on a tape.

Why somebody put the tape in the wrong spot. And it wasn't erased like it normally was. Like, think about how many episodes of DR who got lost because the policy at the BBC at the time was to erase the tapes because why would we wanna keep this stuff?

We already aired it, nobody cares about it. Now, 50 plus years later, people want that data. Now I'm not saying that my, that, uh, Azure, the data lake that that Sentinel is using is going to be able to restore, you know, old DR.

WHO films. Although that would be cool if it could. Uh, I'm saying that being able to keep that data and know that you have that data and be able to go back and say, okay, how far back does this exploit run, gives people in an organization a much smaller risk profile and it makes DFIR and auditors very happy when you can say, no, we have not been running Aval an invalid version or a, a, an exploitable version up until this date.

So we know that everything beyond that is probably safe. So auditors don't have to spend a whole lot of time combing through that. Neither does ZFIR and it really makes people happy when they can narrow their focus and, and use their energy for something useful instead of like combing the desert, so to speak.

And I think one of the nice, I mean, at least from what I took away from it is having that, that that security, that comfort knowing that Microsoft is behind the platform in the sense that when you have such large amounts of data, it's like where you, you pretty much grow your, your entry point into where a threat could attack, right? Or a CV could attach and then you start having to pillage through, well, what's the clean data and what's the infected data? So being able to kind of sort through that have more visibility into it, I think is huge too.

I keep coming back to, if, if you tell me that that many logs of, uh, years logs worth of like firewalls, endpoints, things that we're feeding into some kinda a log sink, if you can tell me that the cost associated with that storage is now decoupled, uh, literally set apart from what would've been traditionally. Like, you know, and I'm not gonna pick on any particular vendors today, but I'll just pick on Splunk from it. Like everything just went into Splunk and Splunk team got their Splunk IT budget and the Splunk IT budget inexplicably compared like the compute and the storage and the network and everything else.

Kinda like wrap that into a one little bundle. But now if you can decouple storage over here, you know, which is where the logs go to live forever, if you can tell me that starts to look like what, you know, if your TBM person technology business management person, you know, you have your, your traditional IT spend, then you have your, your variable or opex part of it, and you have your, like your, your labor attach when the labor attach, you know, looks like the investment of what would've been that storage cost. But now you're saying instead of it being the one year cost or the one year and a half cost for like, what's the hot tier?

If you tell me that starts to look like I can actually do some real comparisons, um, to understand business value for that, you're probably gonna get my attention. Again, execution matters. Uh, but it does appear that this decoupling, this data lake, uh, concept coming from the, the analytics world and now rushing in to solve, uh, what you clearly call out is like a multi-year span problem in security may maybe that will work out.

Um, I'm also very curious to see how many people are going to adopt this, um, because it also sounds like because it's Microsoft, you'd be consolidating your Microsoft spend. I'm not sure if you get any cost advantages there as opposed to having third party players involved. But, uh, that was my other thinking about this, is you're, you're kind of decoupling the storage costs away from a compute or a query cost.

Uh, and, and I think it's important, um, to, to lay out, uh, especially if you are thinking about this at a very senior level. So let's talk about why we want all this data, because one of the things that we, we realize now is that the way that security operates is slightly different than we're used to. Um, and, and one of the other things that Microsoft highlighted with Microsoft, uh, Sentinel was this idea of graphs.

I'm not talking about, you know, X and y coordinates. I'm not talking about bars and charts and pies. I want Jay to explain what you explained when we were doing the pre-briefing here about how graphs work to you.

Because I think this is one of the more succinct ways to think about graphing. I I would think of it, everyone loves a list. I love a list.

I got list, I got list on the wall behind me. I did erase them before this podcast, but, um, the list is great, you know, because I, I can literally, I can, I can see progress. It feels it's very, very satisfying to mark things off my list.

1, 2, 3, 4, 5, 6. The challenge with that now is that, um, if our list for, you know, defense is I start, you know, I focus on one, then two, then three, uh, a lot of what's happening is, uh, the the bad player, the bad actor might start on item number 14, then jump to 23, come back to the ones and twos. And that's because they're attacking a graph.

Um, they're not attacking a list. And so I think, I think what's good for the goose is good for the gander, you wanna call it spy versus spy, whatever that motif or, or metaphor was. But by thinking about a graph, it's a, it's a lot of cognitive overhead, frankly, to think about it that way because it's like saying, take your, take your to-do list and then just make it like a word graph on a page.

Now where do you start? And so I think simplifying that's gonna be important, but that's how I think of this, this transition from like a, a crude list to now thinking about this. Like it's a, it's a web of things or concepts that are interconnected with, uh, mapped dependencies.

And, uh, that's unfortunately how the bad guys are gonna think about it. So how we approach graph and how we can simplify adoption of graph as an approach is, I think gonna be important for the new, uh, realm of defense that we have to enter into in these modern times. It almost sounds to me like a, a choose your own adventure book, right?

It's like, you know, if, if you want to attack the email server, go to page four. If you want to try to hack the password database, go to page 12. And if I am a system that's trying to process that serially from page one to page 50, it gets really outta hand real fast because it doesn't make sense as a story.

I wanted to say too, and I think it's imperative is like to kind of the flip side of that too, or maybe on the same, you know, discussion, is that how that really helps security teams and soc teams moving forward, right? Like when they're doing audits or when they're doing, you know, full reviews of an environment, number one, it can kind of focus on like some of the targeted like pain points or areas where, hey, this is where a breach may be susceptible, or this is where a CV can attach itself. Also, when you have that type of data looking at those graphs, it's gonna go, well, let's take a look here at, okay, we have a compromised user account, or that user account is tied to this VM or this container.

So again, it gives you full visibility into this is where they're gonna hit, this is how we have to, you know, kind of position ourselves and it helps overall better posture, I think from a security assessment, you know, and it gives those teams extra tools and, and, and guides to leverage Not just posture, but I think it actually helps direct the incident re response side of things. Yeah, and here's the reason why. 'cause Jay, you bring up a really good point.

People don't think in lists. Like, there's not a 10 point checklist of, okay, I I'm on a server, you know, check the user login file, check this, check that. What do people do?

They look at the server, okay, what am I on? I'm an e I'm on an email server, boom. That immediately shortcuts to this thing down here.

I wanna try to expose boxes, I wanna try to get user login information. I wanna see if I can get other information that's maybe held in like group folders and things like that. Well, but if I'm on a domain controller, that is a, yeah, and I'm dating myself by saying domain controller, um, like that's a completely different attack chain, right?

Because now I have a copy of AD and I can go in and I can do stuff. But more importantly, if I do happen to find myself on a peer domain controller and I've developed all these attacks, am I going to start looking for a Linux database server? No.

I'm gonna see what I'm connected to that is other domain controllers so that I can see if I can compromise them as well. So by understanding the way that graphs work and being able to hunt down those graphs, I can follow that movement, right? Like, 'cause that's the other thing too.

The, the hacking methodology of land and expand is by its very nature chaotic. Because if I find a good, like, hit over here, I'm gonna follow it because there's, I'm gonna have to put less work in to compromise the system, or maybe it's a better foothold, or you guys weren't dumb enough to leave a Windows 98 machine on the network, were you, were you like that kind of thing. Whereas in, in Jay's example, the traditional checklist of, you know, we checked the firewalls, there's no problem.

We check the IDs, there's no problem. Well, we didn't see anything. Let's go back to the top.

We checked the firewalls. There's no problem like that. The, the playbook is old.

You know, it's, it's every war movie you've ever seen. You got Top gun maverick, good example. This is the natops for the F 18, you know, it back and front funk.

So does your enemy. I'm gonna have to teach you how to do things that are not that, and that's graph theory, right? Push an airplane in ways that it wasn't designed to be used.

Like, and, and I think Microsoft has hit on something here because like, like we've said, hackers don't think in checklists. So what, what, what is the value of having graph for someone in an end user position? Because like, I know what the value is for people who are making these decisions at the boardroom, it's flashy and it, it, it, it is money that can reduce my risk support.

But for you guys, the people who are in the trenches who do this, what is the value of having a system that can do graph? Uh, you get much faster pattern recognition than if you're just trying to correlate, Hey, look an endpoint, talk to another endpoint. Fabulous.

Thank you so much, Einstein. Um, this is the year 2025. Tell, tell, tell me a little more detail than, hey, look, this thing pinged this other thing, or this pork, you know, got knocked.

I, I think, I think as the patterns, um, you know, we, we almost wanna see, um, the, the orchestrated playbook, uh, being executed against us. Oh, look, that's one of these, um, it's, it's this pattern. We are seeing this pattern as opposed to an alert occurred again, you know, going back to rubbing sticks together with your domain controller, right?

Yes. Okay. An alert happened, but in the totality, what was the pattern of all of the, you know, call 'em several hundred, several thousand alerts that occurred?

What's the emergent pattern? Uh, what does this tell us is happening? Um, do we, you know, you go back to like even like a, like a CrowdStrike or other type of metaphor, um, um, or, or, or, or, or people try to come up with cute names to describe, uh, the new threats that are out there.

Um, animals, insects, you know, where they came from regionally. Uh, that's a pattern. Um, that's a technique.

Um, right now, uh, we, we, we really can't deal with the individual alert. Uh, we we're just not allowed to, I mean, I, I think there was another call, um, or talk around where do you even find junior, uh, uh, security analysts. There's no junior security analyst jobs anymore, because a lot of that has been, you know, given up to machine learning.

Uh, that's doing those very intro level roles. So again, we're, we're thinking in terms of patterns. And so, uh, once we have the pattern, we may have a counter, uh, you know, to that pattern.

But in, if we're still dealing with alerts, we're just, we're so lost in the weeds. Um, so I think graph is about pulling ourselves out of that, you know, one to one only view of the world. I, I love that you brought up that you can sometimes do attribution, seeing how people behave.

I think my favorite piece of that was a couple of the groups, the, the fa the bear groups. Um, we knew that they were probably Eastern European or Russian and Origin because they had blacklisted certain places from being infected by their malware. Um, basically it was, do not poke the Russian bear kind of stuff.

And if you can pick up on those little subtleties, then you can figure out, okay, I think this is this group. Like, like if you know that a specific group is, is famous for buying insider access, if you're seeing insider access type patterns, then that narrows it down. It's probably not this group from the far east, it's this group from Brazil or something like that.

And that gives you a more, um, solid place to start from so you're not wasting resources. Gerard, what, what about you, what do you think is, is some of the value of having a a graph type solution available? Well, like I said, I mean, I a hundred percent agree with, agree with Jay.

You know, having, having, my biggest thing too is with this platform, it allows you to blow out what you see as far as from an analytics and from a log perspective. Because from someone, especially for years who, you know, on the network side, it's like, well, we know this packet went to here and this is where the attack, you know, started. This is where it originated.

This is where we need more, we need more to be able to expand upon that. And I think the cool part is being able to ingest such, such data, but then rip it apart at such a level. Like, I remember during that demo in October, I was really excited to see how it breaks down, not just the time of when the attack happened, this was the attack vector.

This was the point. It, it, it broke down an entire timeline. Um, that really like, gives you a visibility.

And especially too, when you're looking at it not just from the technical perspective, but you know, with a lot of CISOs or higher level, you know, individuals, they, they want that data because they wanna know, Hey, what happened here? Why did this breach occur? Where did it occur?

And then how did we stop it? How do we remediate it? So when you have that all under one package, um, you know, I think graphing and, and, and kind of again, more granular level detailed reporting is, is what we need.

So for me, that's how I, at least I, I look at it, it's just, it's, it's, it's a way to blow out the, the, the initial data point instead of just saying, Hey, this is, this is infected. We got hit here, that's all we got. No, there's gotta be more.

We gotta be able to pull more from the firewall logs. We gotta be able to pull more from, from, you know, the, the, the Ford manager UI or whatever it may be. We have to be able to get more data and this provides it.

And I think that the other thing that's valuable for me, kind of going to something that Jay said was, where are you gonna find a junior engineer? Now my question is, why aren't you making junior engineers? You, you should take someone, bring them in.

But rather than to quote master, you must unlearn what you have learned kind of things. Set them down in front of a graph solution and say, okay, I want you to figure out how this works. Look at the way it thinks when it's building through all of this.

Don't worry about the data lake. Don't worry about all those stuff. We'll get there.

Just watch it hits this, it goes here, it does this, it does these things. Because there you're taking someone who doesn't have a preconceived notion of bastion host firewalls, who doesn't have a preconceived notion of VPN concentrators. Again, I'm dating myself, but it's figuring, helping them figure out how modern stuff works.

Like we always talk about this in security because we are so focused on technology solutions, right? We need to microsegment the network and create zero trust boundaries. And I saw an article yesterday where someone on a help desk is selling insider access for like a hundred dollars for an hour to get on and just do whatever you wanted to do, create a foothold, whatever.

We still don't think like attackers, we still think like defenders graphs get us, at least to the offense defense side, people that play the, on both sides of the ball, so to speak. I think it's more valuable for someone to come in and look at that and start their learning there. Yeah, you can give 'em the fundamentals later.

Here's how our firewall works. Here's, here's why we do access lists the way that we do. But letting them kind of have freeform capabilities, almost like a mind map is a much better way to build the next generation of junior engineers because they have to listen to me.

For any of us that have been doing this for more than five years, they're gonna be like, but why do we do it that way? Well, back in 2010, when this was state of the art, that's how we built stuff. And, and I know we don't do that anymore, but we still think that way.

So I wanna kind of wrap this podcast up and I wanna ask you both, like for people going into 2026 that are examining how things should work and, and if they're ma looking to maybe make some additions to their things, what's one thing that people should be thinking about in regards to data lakes and graphs and that kind of technology, even if it's not Microsoft's solution, what's something that people should be thinking about as they maybe are putting their wishlist together for the boss to say, Hey, maybe it would be good if we investigated these ideas? Well, you know, I definitely would say one thing to keep in mind is, I know, I know the hot, the the cool word is AI still, it's, it's the thing I know and everybody's like, why? But it's the truth.

You know, I know a platform like this, or any, you know, graphing platform is gonna leverage generative ai, generative AI to really ingest and pull that data in. So, I mean, one thing I would say to keep in mind is we have to be on the other side. We have to start thinking more proactively and stop thinking less like a defender and more like an attacker, or at least find a happy medium or a balance to both.

If we could find a balance to both, it's gonna give us a distinct advantage and a leg up to, you know, being more proactive at CBEs, and then we're gonna start the, the smarter we get, we need to start leveraging these tools to prevent the attacks. You know, I think one thing I had mentioned, and I'm not even gonna go into it, but it was a whole thing that, you know, there's an entire type of threat out there now that could actually freeze an EDR and XDR solution. It could just kill the whole solution right from the inside.

So we need to be able to start leveraging that, and then that way it's, we're gonna see that blanket across the entire industry, and it, it's gonna help a whole portfolio of next generation tools and suites, especially graphing. Yeah, I would say there's a couple of spaces that you'd want to explore, uh, group, group wise, uh, you know, lunch and learns, um, getting folks that maybe cut their teeth, you know, uh, rub sticks together around like a CIM or a a more of a common information model. And then, you know, bringing them, maybe kicking and screaming into the more advanced security world.

Um, thinking in terms of, okay, well that was, that was great, but, uh, we, we only had a firewall back then. Now we literally are dealing with potentially hundreds if not thousands of individual baby firewalls living everywhere, all trying to log sync to somewhere. Um, so getting people out of that, uh, you know, um, uh, I, I feel like picking on Splunk again, but like, there's the Splunk mindset of how it started and then there's this new world of like, you know, KQL or Cresto, however you say it.

Um, we're, we're, we're literally going, we're deep diving. Um, we're going into the lake. We're gonna explore the, the, the, the bottom and look for surface formations and patterns and see like, ah, this kind of fish, that kind of fish, this kind of eel, that kind of thing.

Um, and so the analogies are gonna be, um, taxing for those that maybe grew up in it and were used to doing it a very specific or one way of doing it. Um, but I would say there's probably some parallels there to the old, uh, storage engineer versus what's now called a cloud engineer. And so that same kind of progression of skill sets and matriculation is important, you know, so to Tom's point earlier, just you should be creating these folks.

Um, you, you, if you have the talent, um, investing in that talent, um, is absolutely on the table of possibility. Um, these, these, these are, uh, folks that have done it once before and it might have been the one-to-one thing, but now it's gonna be the one to many or the many to many patterns that we have to think about in the future. Uh, for my part, I'll say this, if you are even considering any of these solutions, all you've gotta do is go to the current providers that you have right now and simply ask, what are your plans to support advanced threat detection?

And, and if you wanna say the G word graph, say it, one of two things is gonna happen, they're gonna look at you and go, what's that? Then you get to educate them, send them the episode of this podcast. We would love that.

But more importantly, um, if they do have plans in the pipeline, they can say, well, that's something we're looking at. Then that can give you an idea. Is this six to 12 months out?

Is this 12 to 18 months out? If this is something that you and your organization need to investigate, either because you fear you're about to be breached, or your, um, stakeholders really need some kind of insurances here, then that means do I, do I stick with who I've got or do I make an investigation? Um, and if you do make an investigation, I'm sure friends over at Microsoft would love to hear from you because they put a lot of effort into Microsoft Sentinel and they'd love to sell it to some people.

And, uh, you know, we, we appreciate them partnering with us on all the things that we've been doing. So, uh, thanks again to them for that. Um, before we go, I'd like to let our, uh, guests kind of plug their stuff.

Uh, Gerard, if people wanna check out what you're working on, where can they go to find that out? They can find me all over the internet at Tech House Five seven. Oh, I'm, as I said, I'm on LinkedIn, Twitter, TikTok.

I make short long form content. I'm all over the place. I'll probably have some new videos coming on Sentinel and uh, it's gonna be a great time.

So yeah, you can find me anywhere. Books are sold and I'm all over the internet Tech House. org.

I'm on LinkedIn, I'm on Blue Sky. I'm using Go to Social now 'cause I had tried using a standard Mastodon server, but my Fed Averse is now on Go to social. So, uh, look forward to, uh, anyone reaching out to me there as well.

Thanks, of Course. com. You can also check out more of our great security conversations over at our Security Boulevard podcast.

com for more information there. We wanna thank you for listening to this episode of the Tech Field Day podcast. And if, as always, if you enjoyed this discussion, do us a favor, subscribe on YouTube, use your favorite podcast application.

We don't want you to miss any of these episodes, especially around the holidays when things are coming out, um, on a maybe less than regular schedule. Well, not for us because Corey is amazing. Uh, but if you do enjoy the content that you hear, let us know.

Give us a rating, give us a review, leave a comment. We'd love to hear what you have to say. This podcast is brought to you by Tech Field A, which is the home of IT experts from across the enterprise.

We are part of the Futurum group. com. You can also check out the website, tech field day com slash podcast for all of our episodes.

Or listen to us over on Techstrong tv. Thanks for tuning in. We will see you all next week.