CarahCast: Podcasts on Technology in the Public Sector

Lattes & Learning: Transforming Public Sector Networking Systems

Episode Summary

Listen to the Juniper Networks podcast to hear from data center experts discuss the benefits of AI-native networking and how leveraging Juniper’s Mist AI platform enhances end-user visibility and experience. Discover how the world’s first AI-driven wired and wireless network optimizes Public Sector workflows and data center systems.

Episode Transcription

Jon Jumento

So the next part of this segment, we welcome Loralee Wood. She's from mitre. She's a principal engineer over at MITRE, that 20 years in the industry, mainly on the telco side. But if you know she's come down and kind of talk a little bit about her experience with using this. So Loralee, thank you for coming down. I really do appreciate it. Let me start with a basic question. So. So tell me a little bit about you know, some of the pain points, what made you gravitate towards looking at Mr. Initially, and, you know, what are your thoughts afterwards?

 

Loralee Wood

First of all, thanks, everybody for having me. So he says 20 years it's actually over 30 But I didn't want to say that. And typically as an IT engineer, you know, we're the we're the people behind the scenes. I like to say that we're more The Wizard of Oz. We don't usually come out to events like this, but I will compel p because we didn't know about any of that and I couldn't As we first started looking at the poll, we're hearing half as well. And I hope I'm louder than a lot today. But that's all I have to copy. We started looking about a year and a half ago, because we're at that point of having the lifecycle our infrastructure, and I challenged our team to not think of it as just a lifecycle, but to think of it as a new way of tackling some of our major problems. So as we did the RFP, we thought of a few of our pain points, one of them was our old architecture was controller based, and the controller based architecture was just not going to grow with the amount of users that we had using our wireless system. We couldn't backhaul the entire country and all of our offices back to two data centers on the East Coast. And secondly, we knew that we wanted to architect something towards Zero Trust. So we're looking to kick our users off the network. I know that statement sounds kind of bold, but not the researchers, not the folks in the labs. But the folks that just are in doing day to day business, having to reach applications on the internet, we wanted to build a completely separate infrastructure, and focus on those users having to use Zero Trust VPN in order to gain access back end, or use the the access to gain access to the internet. And then lastly, we were focused on a couple of the bigger initiatives for is a cloud based in is it AI? Those were some of the directives that came from our leadership. And so that's how you approach this problem. Great,

 

Jon Jumento

fantastic. So when you think about integration, so integration is tough enterprise wide, with any type of system you try to introduce

 

 

into a heterogeneous environment. Tell us a little bit about your experience with the integration of spot. Well,

 

Loralee Wood

building a completely separate architecture, really, you didn't have to integrate it as much, because again, we wanted to kick the users off the network, but we decided to do a full package. So it wasn't just Agena, proneness. And wireless technology, it was looking at switches. And looking at security. We did, we did need to make sure though, that all of the data that we needed, could be looked at in the same manner that we looked at the data before. So using the API's was very important for our developers, they would they did say and they constantly note to me that it's much easier for them to program to the missed API's than they had in the past. They're very appreciative of the ease of use and how much is open. The other thing was, and this was a funny point, I would call it and probably not as important anybody outside of IT ops. But the fact that when I call us pop and swap, the fact that I could send my engineers out, take our old access points off the ceiling, put in these without having to change the brackets, made a deployment of I think was almost 880 access points happen in less than 10 hours for our campus. So that was just a major win for us, because we wanted it to be there when our users weren't to change this infrastructure.

 

Jon Jumento

It talks about scale. Judson talked a little bit about it in his presentation. So that's a great example of, you know, kind of the power but also, you know, a practical example behind operations when you kind of look at the solution as a whole. So what about benefits, advantages? Disadvantages, right, you know, give us an idea of that.

 

Loralee Wood

So I'll start with the advantages. So I was compelled to come here, because the first time that we deployed this system, I call it a system of systems. But we deploy this on our campus in Bedford, it was about 540 access points. We did it on a Friday afternoon. And when we were done, we kind of looked at each other and, and said, Well, wait a minute, it's work. And was it really that easy? Now we have to go back and test. And then you know, we kept testing kept retesting and kept looking at all the data. And come Monday morning, I still walked in with trepidation wondering what was going to happen with us people I've been, I've lived it for that long, right? You had that maybe eight to 10% of the users come up to the service desk, they ask questions, mostly. I didn't read by now. I forgot this is happening. How do I access? And it was probably I think we had five or six issues that we had to fall back on. So I think, you know, that was a great, great way to that was a great selling point for us was that we didn't have all those problems. You know, when you have a rollout that doesn't have a lot I've talked about it afterwards, kind of a successful rollout, right? Nobody's saying and we also saw a lot of benefit on the troubleshooting or the alcoholic proactive troubleshooting. We were alerted that that day, you know, we had some engineer that was on the back end, looking at this going, alright, you got a cable, that's bad, you got another cable, that's bad. You know, we had about six or seven cables that were bad, which normally it would have been on Monday or Tuesday, or whatever day the user happened to be in that area, they would have called the service desk and said, network doesn't work. And we know the network doesn't work, right. But how long did it take? Would it normally take for service desk to try and troubleshoot that problem who didn't have access to all of the data, so they would escalate to the network engineering team, who then had access to the data and would have to go try and replicate it and then find out? Oh, it's a bad cable. I mean, so we save so much time with that, we consider that very proactive. I think the pain point was that we were a little bit of an early adopter to the full solution. So we worked very hand in hand with you guys, on the SRX, inclusion and missed, and now, we templatized, everything with the SRX has become a lot easier. So we can manage everything through the platform.

 

Jon Jumento

That's fantastic. So would you say response time is faster, especially with the size of team you have to deal with with the example you gave?

 

Loralee Wood

Well, not. So we heard that one of your customers just wanted zero calls from the customers and I tried to be a little bit more realistic. I said, Alright, let's, let's reduce the calls to the service desk by 80%. Let's shoot for that. And we did. So not only are the calls not coming in the users happy, we look at the performance statistics. And we realize that they're getting faster performance, not just because they're 60 available, a lot of our clients don't have 60 year, but because they're reaching their applications faster, even on their existing wireless cards. So the calls aren't coming in. And when they do come in, we were able to open up the missed platform to our service desk. So sure they can go in, they can look at all those details. But they can also ask a question. And they'll ask a question that says troubleshoot. And they'll just give the laptop name. We also are testing the integration with Office 365. Which I don't know if you know that but in office 365. In a team chat, you can actually query missed. And so because we are in office 365, the doctor, our service desk can use that chat feature right from the application they already have open, we don't have a seven by 24 not staring at a screen anymore. And when anybody does that. So they just they just query through the application they already have open.

 

Jon Jumento

Fantastic. So does anyone have any questions thus far? Before we continue? Now? I don't see anything on the screen. Okay, so let's keep going. So let's flip it on its let's flip it on its head a little bit, what were some of the hurdles that you came across in terms of, you know, trying to get things done? It doesn't sound like there's any, but figured I'd ask.

 

Loralee Wood

I wouldn't say most of the hurdles were probably our own internal controls, they were just there was there was self created hurdles. So you know, kicking the the this network out of our infrastructure made this making our CUI data and our compliant data. Architecture smaller. So it's reducing the threat surface of that. And so, but we didn't want to put controls around it. So we're still doing all the same scanning, we're still doing all the same controls. So I would say they were more self inflicted, to be honest. And then being an early adopter with the SRX invest. You know, we did cross some hurdles with that. But now that, you know, all the releases come out, pretty close to when they released in Juno's in the midst. We're doing the same client.

 

Jon Jumento

Great. Yeah. Now that's great. That's really important, especially in our space. So what about future plants? Right. I mean, you've seen the capabilities, right, you've, you know, obviously MITRE is, you know, seeing the impact. It's having been so early on or deployment. What are you guys thinking about for the future?

 

Loralee Wood

So I like I like that question. Because now my engineers are realizing that they're not running around service desk is realizing they're not running around answering all of these questions. You're just complaining. I probably have to get rid of a phrase I used to say to new hires. I used to say, just remember, nobody comes up to you and says, Hey, thanks, networks working today, because we only hear the complaints, right? We're hearing less and less complaints. We have less and less churn in that space. So now they can focus on And the fun stuff. And the fun stuff is around Wayfinding. So mitre has adopted wayfinding for our internal users and our guests, it'll say you are here. And you can tell our app on your phone, you know where you want to go to, and they'll tell you how to get there. And that's kind of cool, because I rarely come down to McLean. I'm not a flyer, I drive. So I'm used to our Bedford campus. So when I come down to our McLean campus, and they need to go to a space, I can enter it in our Wayfinding app. So we're doing a lot more with Wayfinding. We're also using the wayfinding in a manner where we're going to try and set up security parameters. So I think all of you may be familiar with like secure spaces, and you can't bring any device at all, except maybe a pacemaker, right with you into those secure spaces, but people still forget. Right? So how do we stop them from forgetting? So if we set up a security perimeter using BLE, you know, you can get off an audible alarm and say, he crossed the barrier, you still have your phone with you right outside the barrier, obviously not inside. And then we're, we're looking to do more and more with that conversational aspect of of the API. Now, how can we? We haven't seen it yet. But I, you know, I like to challenge like, Alright, how do we open this up to our users? Right? So we can integrate with Office 365? And you can chat and a team with it? Why can't I use a chat with the album? Why can't a user say, you know, I'm having trouble what's going on?

 

Jon Jumento

Yeah, that would be really cool. Especially in teams. I mean, I'm sure everybody uses teams here. But I think that visibility is key. Right? Because there's, I guess there's a lot of assumptions you can make, but usually the network's the first one to get blamed. Right. So, okay. So you talked a little bit about your users, right, in terms of, you know, them being able to come up ask questions, you know, some of the stuff is probably more on the lesser complicated things, right? What about benefits to your customers? Right, you know, the customers that you serve, right has missed having the benefits there, or, you know, maybe possibly in the future

 

Loralee Wood

mean that, like my customers, or my integrity, my customers, or anybody that's at the minor campus, talking about our sponsors, sponsors? Yes. So, I mean, I would say that, you know, we don't ever recommend products. So, you know, obviously, we'd help customers go through RFIs RFPs. But the ones that I personally have helped with a lot of them, were dealing with similar problems that we dealt with, you know, how do we get how do we get out of having a controller based system? How do we make things more efficient for our IT department? Because we have to do more with less? How do we get out of you know, that hour long troubleshooting that says there's a network problem when it really maybe wasn't, I'm seeing things show up in the midst AI that says, this D DHCP problem on the user's laptop, or there's an issue on a user's laptop, and that's that used to take us a good hour, and sometimes escalations to my team. So I think just any RFI RFP, as long as it meets the requirements, you guys happen to meet us.

 

Jon Jumento

I appreciate that feedback. Because, you know, it's interesting that even our government customers, they see applicability of this, and some of the most complex environments, even the ones that are gapped right now, you know, some of the requirements and logistics are a little different, right? As we talk a little bit about how, you know, this particular solution involves kind of the way we look at networks. But you know, it's really fascinating. Well,

 

Loralee Wood

it's not just the wireless piece, right? So because we, we rolled out the the separate architecture, that's all off network, so we can give wire drops that are off network, and those folks have to VPN back in as well. And we can get insights into those wire jobs, we can get insights into the switches. Also, you know, one of the things front and center is compliance, right? So I have a team of three engineers just in the network in my immediate networking team that you know, focused on writing all things widgets, let's make something easier for somebody, let's do this for somebody, oh, zero touch deployment doesn't work. Let's write all the code around it, so that we can make it easier to deploy that code upgrade to oh, this set of routers has a specific program, they wrote, the set of switches has another program, they wrote this will now through mist, my switches, my Sr, x's and all of the APS, I just set to Auto Deploy. And we ought to deploy in a manner where we roll out to our little testbed, make sure everything works. And then we just schedule the rollout for the rest of the enterprise. And the enterprise right now happens To be five remote sites in the two campuses, and we're continuing to roll out this year, so why don't have engineers at those sites? And it's been very busy. Yeah,

 

Jon Jumento

I think that's key. Right? You know, today kind of do more with less, I guess less is bodies, right? Because I know when I started off, when I started off on the services side of things, if you needed something done, you just threw bodies at it. I were definitely not there anymore. Right. So I think this is a great example of figuring out how not only do more with less, but now that you have the time, what else can your team focus on? Right?

 

Loralee Wood

Yeah, they're, they're really excited to adopt more. And I'm like, well, slow down, you know, we have to do we have to do our due diligence anytime we do an RFP, you know, but we're looking at our access layer. And we're saying, Alright, well, we'll do an RFP, and we'll see what happens here. But, you know, then my developers and our service desk could have access into the nest for the rest of the rest of the required job feed company that are on the other network. It's

 

Jon Jumento

fantastic. Fantastic.

 

Loralee Wood

All right.

 

Jon Jumento

Any additional questions before we start kind of Q&A? Oh, how is Miss provided a better experience for your customers? I

 

 

think this was directed to you.

 

Loralee Wood

Probably. So I wouldn't say it's missed, per se. So I think the system is better. So because we don't have the controller based system, our users are getting to the internet faster, right? They're getting direct to the internet, or they're getting direct to the resources they need. But we require everything to go through our VPN, which happens to be on the internet. So let me try to describe one of the scenarios that we were surprised at. So because we use Office 365. Well, these coasts is service right here in this area, right brought us Microsoft Office 365 for certain level of, of compliance of certain level. So I assumed that my folks from McLean might not realize it a great benefit, or you know, we might not measure anything better for them, you know, their experience might be the same. And what we noticed is just pure network performance wise, getting the office 365 was 10%. better in terms of speed, network speed, not not maybe application. So if you use that for fourteens, and for video and video conferencing, they're getting less jitter less interruptions. And then we measured it to our Zero Trust VPN. And we saw better performance there. So I was really surprised by that, because it wasn't that I went from Wi Fi five, five or six years ago, right? It was the same Wi Fi users, the same app, same day, same experience, and we're realizing the benefits. So that was really surprising. And when they don't get the latency and jitter, they don't get the interrupt interruptions talking to our sponsor, the don't they feel happy, right? And then we look a little better, I'll say, right, like, Hey, wait, you're gonna advise us on but yet your infrastructure can't support having a video call? Right. So

 

Jon Jumento

yeah, that's really big. You know, I think that's worth mentioning. So that's a really great example. Andrew, I can't read the

 

Loralee Wood

second one here is house myths help improve your daily activities based on your current role

 

 

and responsibility.

 

Loralee Wood

So I'm a group leader, I have a team of I think it's 1111 engineers off the top of my head, we just had a little bit of a move. So yeah, it's 11. It's interesting, where, once we prove that this works, I kind of don't have to argue for it. Right? So you always have to prove, you know, does it measure is it going to be better? And now that we are able to show all of the data, you know, it's a lot easier to say, hey, this, this AI stuff or mist is working for us? You know, let's consider it when we look at, you know, the next RFP that we have, right, let's really open it up and see if it's applicable to ICU AI network when we look at the access.

 

Jon Jumento

Fantastic. So any questions for any of us up here? At this point?

 

Loralee Wood

One more question. Here comes air. Yeah.

 

Judson Walker

Yeah, making some very great thing about the positivity, you've seen posts. There are certain trends that hey, look, this is what I'm experienced in my campus post deployment. Has that spearheaded other technical conversations with your peers from an architecture perspective that may be more datacenter centric, or more Wan centric of maybe moving some of those capabilities potentially into those? Well,

 

Loralee Wood

we're looking at more and more cloud based applications. So I think that you know, the more users that I can that I can get to go internet first You know, their experience is going to be better. But now that my engineers aren't running around, they're now spending time saying, alright, what can we do? What can we code? Where else does this apply? Right? So without naming any products, I will say that we do have a data center product as a beast. And my program is just, they don't like it. Right? We're unable to realize its full potential, but yet we pay for its full potential. The licensing model is something that I don't like to navigate as an engineer, I never thought I would have to have a, an engineering degree in licensing. So So now they're saying, hey, you know, in a year and a half, and we look at data set and stuff, can we relate? And I'm like, well, whatever we look at, we have to do our due diligence, right? Let's not just do lifecycle upgrades. Let's see what's out in the industry, how we feel are benefiting by AI. And look at what's out there that, so yeah, we're gonna do our due diligence with that, too.

 

Jon Jumento

Fantastic. Hey, I think I see a question up there for for Judson, actually. So yeah. He doesn't even like Trump, what makes mist unique from other AI ops solutions.

 

Judson Walker

One is, longevity and development. So again, you look at what we've done with this is seven years of data modelling has taken place today. So there's a lot of issues that have been identified in the past, that have now been addressed inherently within the modeling that we've already done. So those classic experiences, you're just going to get day one. Now, that doesn't mean we're not going to run into the problem we are, as networks evolve and shape, and then we will adjust the model will adjust accordingly. And then so you're going to see a build out of capabilities in this as an insertion point that I don't think you can find in a lot of places today in industry, just because of the proactiveness of the past that sort of dictates successes in the present and in in the future. I think some of those are something that Laura Lee had mentioned that she's experienced at MITRE. But more importantly, where I think where things are starting to manifest itself in this is the ability to associate what's proactively seen with corrective action, especially when it comes to trending. Because I think we don't do a very good job as operational personnel to see trends easily. We see things when they're broken. So what miss being able to see trends and like to identify those based on a set of SLAs, I think is huge. And then VA will then say, Okay, I want to take action on that early. So at the end of the day, we solve problems for your consumer see problems. Cool, I want that good. I'll pay for that every day, all day. Because again, it allows you to get in front of a lot of things that that you know, to your point around love your common around the user and the laptop and the issues and they, it can really take time, right? Whatever we can proactively say, hey, look, I am having a major zoom call with a sponsor. This is a level of performance I expect coming from a particular switchboard. Please identify those ports that don't meet that SLA early before I even start. And then let me get those off the table quickly versus now I'm real time problem happens. And in option we're in you know pharma, depending on who's got the problem, right. So I think that's huge. I think that really sets us apart from a lot of stuff that's going on in India right now.

 

Jon Jumento

Fantastic. Okay. Well, we are here. So have teams seen an improvement with regards to application orchestration or service operations? Although it sounds probably you orally.

 

Loralee Wood

So alright, so that's not capital T teams. So that's to have internal team seeing with service? Art. Yeah, well, I service operations. Absolutely. So we're getting less escalated ticket. Now. I mean, that was that was key for for my engineering team, right. And I want the service desk to be able to handle 100% of the calls, and I want them to have less calls to handle. So I'm realizing they're a little happier because they're not getting the escalating call. And you know, what happens when that call gets escalated? Right? So they had a bad experience? They call the Service Desk. But now the experience is over. How do you replicate that experience? By the time it gets to my engineers? How do they replicate it? But one of the things that you mentioned I forgot to say is Well, now sometimes we don't have to replicate it because if the issue was on the network or in jitter, and latency or anywhere seen in our infrastructure, that that kept that PCAP was launched. So I could go back in a month, whereas before I couldn't, because everybody I think if you've lived it and loved it operations, you know, the intermittent issues, probably the worst. Replicated what happened when, when, when it was happening to you? So yeah, that's that PCAP has helped us. But yeah, my team's just really excited to just work on on the exciting stuff. Nobody likes to run around in their network closet and find a cable or, or go out and stand next to an angry person.

 

Judson Walker

I'll even add to that when it comes to newer capabilities that you want to operationalize. Like, for example, there's a lot of conversation today about maybe moving away from that traditional core distribution edge in your campus. And now more focus on campus fabrics. We're not talking about VX LAN e VPN, and everybody's heard those buzzwords, there's a lot of people to this day. But what does that really mean for me? Because when you hear it, you think, well, that's BGP, we'll do BGP, it's so much more than that. Right. And, yeah, there's a lot of No, you know, kind of tingya challenges as far as configuration is concerned. But what they miss is done is obfuscate that, and just think about, hey, I want to create a fabric, that fabric has a spine, it has a leak. And I now have created that relationship. Now let the system build the syntax on the back of it to elect to enable that fabric for you. Or if I want to create a BNI or a layer two extension, you specify that and allow the machine to dictate the syntax on the back. So once the architects know what needs to be done, and the underlying technology and why they want to deliver it, as far as how its implemented, I got multiple vendors on the backend, they have different ways of doing that's the operational activity, tactical action. But at that tempo actually is predicated on my architectural thought. And so to me, that's huge. For an architect to say, I can build this out, I can model it appropriately. Now I can let the machine do the deployment work versus handing over to a whole second tactical deployment team to actually now put hands on I think that's huge. And I think that's hard for people to say, Okay, I've got a UI, that I'm going to specify these capabilities, and now I'm leaving it up to that machine to now implement it on my behalf and to give me a green light. That's like, Okay, how much can I trust that? And do I'm just going to default back to my old ways, and like art, let me SSH into that platform and go through every show command I can to be a validator do it myself, because I don't trust it. Again, that's a time thing. That's a maturation.

 

Loralee Wood

That's not just not trust, right? I don't know, you haven't seen problems where you are. But I have a small section of my team that worries Well, we will, that's my job. It's gonna happen my job. And then they realize, wait a minute, that really isn't my job. Because we were seeing the writing on the wall, we can't keep doing the same, the same things that we were doing before. There's no way that with a team that I had, if I weren't going to get more people that we were going to be able to sustain going out and trying to solve user problems. The more people that we have on campus, I mean, we we have typically every day and we haven't reached our full deployment over 8000 devices. After that one point. You said five devices for for students. Everyone comes in mind we average three, but that's their own devices, right, their personal phone, work, phone, your laptop that excludes any other devices that may be on the IoT network already.

 

Jon Jumento

All right, thanks for coming up on the end, we got time for one more question. So thanks for putting that up. Are the underlying AI models generative models developed in house by Juniper or something available from another vendor? Can it be swapped?

 

Judson Walker

So the short answer is we do development in house. So we do have data scientists that are that are on on our clock to help with this modeling perspective. So it is not saying we don't provide some level of integration from an API perspective. But if you think about the core modeling elements that we've built in house process,

 

Loralee Wood

aspect.

 

Jon Jumento

So we're at the end, it last call for any questions. There's one more in the chat. One more. Okay. So getting one more, how is the underlying model fine tuned in order to address problems like drift, adversarial defense? explainability exploitability? Thank you.

 

Judson Walker

Oh, that's a great question. I don't think I have a definitive answer for to be candid with you. Um, I think there's, as I think about it, architecturally, I think we have done some things especially about making forwarding decisions outside of the world of MAC addresses and IP. We are now on the precipice of start looking at the core essence of metadata and now Not only what applications are being used, but who's using them, and where they're at and where they need to go, and how to prioritize that accordingly. I think to me, that's the first step into building models around individual people, not products or data flows, and then be able to prioritize those and provide different security measures per flow associated with that propagation. I don't think we're there yet. But I think that's a great springboard into a new round of models that are TBD, in my mind. So I know it's not a definitive answer to the question, but thinking more of where we are going versus where we are, so that's okay. I didn't understand. Because, especially when it comes to out of adversarial defense and how you're talking about offensive measure, oh, different ballgame. Right. It goes down a pretty nice, cool rabbit hole there.

 

Jon Jumento

Yeah, just just one thing to add to that. So I was talking to Dan, on the break, and I was attending a government session, I won't go into those specifics. But, you know, one of the things kind of left me kind of hopeful, but then also a little bit concerned. And, you know, the, the concept was basically, you know, traditionally, knowledge comes from humans, right. I mean, that's, that's where it comes from. But we're kind of on the cusp, we're knowledgeable, we won't be the only ones providing it. So if you think about everything we're talking about in terms of, you know, models, capability, automation. But now we're talking about adversarial defense, we're talking about policy to try and put up a fence around it, all that will come into play. So when you think about kind of what we talked about here today, you know, think about how things are going to change as a whole. Right? As you begin to implement this stuff across your enterprise and really realize some of the benefits, but also understanding how to police it. And it's Johnson said, All right. Well, I think that's all the time we have today, unless you had some nice

 

Loralee Wood

wine that makes you I thank everybody for having here. I wasn't I was compelled to come and share the story. Because, you know, this was the first implementation I've had in over 30 years. But that went as smoothly as it did. It did. It had us questioning, oh, my gosh, we missed something. And we didn't. And I thought I needed to share this with others. Because being in it that long. I just wasn't used to it. And I wanted to make sure I told others. No, I implemented it, it really did happen. Thanks for having me.

 

Jon Jumento

Yep. Now I definitely want to thank you, Laura Lee, for coming down from Boston. I like to think it's just for us. So I'm just gonna, I'm gonna think that but, you know, thank you for sharing all of your experiences and your feedback. And I'd like to thank density as well. You know, Johnson is great at this, you know, spreading the word talking a little bit about, you know, some of the benefits we've had, and some of the benefits our customers have had. So

 

Judson Walker

I will need to say, as I close again, I just want everybody to think differently. So if there's one takeaway when you go back to your home shops, have we thought about it in this way? Have we thought about focusing more on proactive versus management versus reactive measures? If the answer is yes, maybe. But we're not sure what kind of word we start, what does that collective vision you want to get to, and you need insights, please call us I've got cards here I was, I don't mind talking to anybody got a very big map, no problem sharing in various ways, various levels of depth. But when everyone starts to collectively throw ideas around and start thinking about the next step, we can put some meat on that bone. And then we can figure out how to get that first single, and then that double, we're not going to solve all the answers today, because existing environments must keep running, they must be maintained. But how does augment and how you augment with definitive outcomes in mind. So let us to be a callus for that technical exchange, because I think we could be a huge benefit. And we again, the last thing is we don't want to start from we want to give you models, validated designs, we like to have a starting point, at least as some type of candle on a something that can easily copied and limit it if we can do that slight augmentations. I think we can get you there. But again, we're not going to networks business anymore. We're in a networking business, which is above blocks conversation, and I'm so looking forward to having it because that last question talks about networking is breadth, its depth. Its intricacy that we need to start talking about and it takes our engineers is Laura Lee's organization and what she's doing and she's doing something different. And now everyone's seeing it. And now that's now turning it to the next level of questions. Again, like much like I said about the Jetsons imagined to be able to be remember, there was a there was an AI engine in the Jetsons called Rudy that spoke to George Jetson to say hey, what do you want to do and How's your wife?

 

Loralee Wood

Your wife What's your wife wants Rosie Rosie she did everything

 

Judson Walker

so again thank you guys I really appreciate your time