Watch episode three of Mission First, our Department of Defense (DoD) and National Security podcast series focusing on the mission, not products.
During this session, Keith Nakasone, former deputy assistant commissioner of acquisition management within the information technology office at the General Services Administration’s (GSA) Federal Acquisition Service and Federal Strategist, VMware and Craig Bowman, Sr. Director - Federal, VMware, will discuss:
Craig Bowman: Hi, welcome to Mission First, I'm Craig Bowman, your host. Before we get started with the podcast today, two things. Number one, we're back in studio, which is fantastic. So, we're gonna be able to do in person interviews. And I wanted to call out Carahsoft, for letting us host the podcast right here in their building in their studio for all the support staff at Carahsoft. Craig A bud. Thank you very much now on with the show. So welcome to the third edition of mission first, just to recap in the first two podcasts that we had, we had Joe betta from Kubernetes and one of the founders of Kubernetes on with Paul Puckett, who is the head of the army Software Factory. And then in the last podcast, we had one of the founders of Kessel run on talking with an engineer about continuous ATO. And so today, I've got the privilege of talking with Keith Nakasone. And we're gonna play buzzword bingo it buzzword bingo today. But I want you really quickly Keith to do an introduction for those of you that have been buried under a rock for the last 10 years. Maybe you will learn who Keith is, for those of you that have been in the industry, you probably already know. But I thought I'd start off with just doing like, do you mind just giving a quick background on who you are and why I have you on mission first?
Keith Nakasone: Sure. Thanks, Craig. So, I'm Keith Nakasone. I spent 32 years of my service in government. And I have 20, some years of experience in the DoD space, as well as a few more in the civilian space. And so, all of my concentration effort has been in the acquisition and procurement realm around IT and telecom. And so, one of the things that we've learned very early on in the process of how acquisitions are supporting technology is really that that shifted, that's occurring. We're looking at problems and solving Mission Operation, solutions better with acquisitions and procurement. And this is the fascination of how do we leverage the procurement process supporting the technology efforts that are going on right now.
Craig Bowman: And that's the reason why you're on Mission First, right?
Keith Nakasone: Absolutely.
Craig Bowman: So, Keith, and I had an opportunity to actually head out to Vegas for CES a couple of weeks ago, we spent some time out there. So, I'm gonna structure the buzzword bingo, we're gonna make sure we cover Zero Trust multicloud DevSecOps, you name AI/ML, we're gonna get to all those things, because it's like everything that people are talking about. But while we were at CES, you and I were both sitting at the same table. Bunch of CIOs and CTOs onstage said something that was really interesting. Keith, I don't think anybody questions. The NIST controls the NIST 800 controls. But it was surprising to see from a stage from executive saying that they should do away with FedRAMP that FedRAMP I think the actual quote was, was ill conceived, and needs to be done away with. So, what are your thoughts about that comment?
Keith Nakasone: So when we look at the whole of government approach, right, when we look at the NIST standards and controls, we have to look at how is it implemented, and FedRAMP was a from a thought leadership perspective of looking at the whole of government, so that we would be able to build out a framework that would be able to provide some kind of governance over the entire process. Now, can there be improvement with the FedRAMP process with technology, leveraging automation to help accelerate that process? Absolutely. I think there's a lot more private and public engagement that needs to happen to see where and how fast we can accelerate this process. But from a foundation, looking at things from a whole of government, and in looking at vulnerabilities and sharing that information with all the agencies as well as the private industry can only improve the posture of addressing cyber-attacks vulnerabilities and building out a risk management framework that works for all.
Craig Bowman: Yeah, it's interesting, because I think one of the criticisms to play devil's advocate is that it's not so much that you know, you're right, an automation and things need to be fixed and FedRAMP. But it seems as if FedRAMP is an inhibitor to innovation, right? That it maybe it's not the framework, maybe it's the process to get through the framework. So, I'm gonna ask you that, which is it? But is it is it stifling innovation? Are companies either not able to get their stuff to the government fast enough? Or some companies just simply saying we're not going to do it? Right, first of all, is too costly. You know, we're just going to get out of the government altogether. So, first of all, what do you think about the criticism? Is it fair? And then the second is, is it the framework or the process, the framework is holding up? If you agree with the statement about innovation?
Keith Nakasone: I love this question, because when we look at the private industry, and we look at government from a public sector Our data is our data, intellectual property, we all have to be concerned about that. So, if we build innovation with software, hardware, etc., we need to ensure that that is done at the foundation layer, the security framework is not just a government responsibility, it is a private and public engagement, where we should be building out a security framework that addresses and minimize the cyber incidents and events that take place, right. So, if we build an ecosystem where we actually educate, train and deliver capability with that in mind, that will only put us in a better posture from not only from a government perspective, but from a private industry perspective as well, that meaningful collaboration is going to pay huge dividends moving forward. Why? Because now we're building that security with the mindset of delivering capability. And that's what that's what we're trying to change with DevSecOps.
Craig Bowman: Oh, interesting. So here we go. First buzzword bingo, we got it. DevSecOps. I knew it was going to come and I didn't even have to, like ask the question. It was going to come out anyway. So DevSecOps, you know, what, when you look across the spectrum of contracting, is procurement setup, to take advantage of the speed at which DevOps DevSecOps promises to deliver, you know, if we if we, you know, it seems like procurement is still stuck in the Waterfall, you know, methodology where we've moved on to continuous release, continuous release, but the contracts aren't set up that way. So, what needs to change to take advantage of DevSecOps? And then do all the things you just talked about with private, public partnership, and, and that kind of what's your thoughts on that?
Keith Nakasone: Absolutely. Well, when we look at the procurement process, you know, it starts with the people, the people are driving some of the behaviors that are happening. So, when we look at educating and training our workforce, our acquisition workforce, they have to understand how industry is providing technology solutions, it's no longer just about a transactional type of action. It is we're moving towards more of that best value, more of that IT solutions based to solve mission problems, mission challenges, right. So, it's really bringing that whole that whole ecosystem together, understanding the problems, understanding the challenges from a government perspective, but also building out that capability, in a way from a proof of concept to a pilot to then scale.
Craig Bowman: It's kind of like SBIR.
Keith Nakasone: Absolutely. If we try to build the big bang theory and build everything into a single procurement, it is almost doomed to fail. So, if we strategically build procurements that align with the Agile development of software, or IT solutions, it will put us in a better position moving forward. Why? Because now we have some scope that we can do around proof of concepts. We have a scope that we can do around the pilots, and then move to scale, because we're building as we build that, or as we inject innovation over time, it's actually looking at all of those security requirements and the technical capabilities that are being delivered downstream.
Craig Bowman: So, it's interesting you the way you're describing it, so SBIR is coming to mind right away, right. That's that. So, what's great, I think about what you just said was the SBIR framework is there, which means small businesses can come in phase one, phase two, phase three. Yep. And it's pretty fast, because the framework is there. And all they do is post the hey, we want this solution tested, and then delivered OTAs come to mind, that'd be another one. So, but then you have other contract vehicles, like EIS, which is interesting, because right now, if you try to procure, you ask for technology, but the moment you attach a person to it, whole different process kicks off, right? Gotta go out to bid. It takes months. And so, the procurement people don't like to do a lot of these big things, because it's very hard. So how do you merge? Or is ESD? Example? How do you merge delivering solution outcomes that require people with technology when the procurement process is set up to be eight months, 16 months, anytime a person is attached to it?
Keith Nakasone: And this is a good question because government wide acquisition contracts can, is a good process to accelerate the procurement process. at the order level, it should have the different instances where you're building out based on maturity of where you are in the process. Let's take cloud for example. If you're building out a cloud solution, and you're at the infancy stage, and you're just trying to get into that market of making some change in the multi hybrid cloud environments, you may want to start with at the order level, being very specific with a proof of concept. Some agencies may be more mature in the process, and they may take on a bigger challenge and put a, let's just say an indefinite delivery, indefinite quantity contract in place themselves. So, it really is matching your agency's skills, as well as the workforce as well as how well can they deliver capability and aligning the procurements and not just saying, hey, this worked for one agency, I'm going to take it and apply it somewhere else. If the similarities of what they need is the same, then they should use that same procurement process. But really, it's having that active collaboration and communication between private and public to see where they are, can solve the mission challenges that they have. First, customer experience and user experience has to be at the forefront in delivering technology solutions moving forward.
Craig Bowman: Uh-oh, user experience, you know, listen, listen as like you say user experience, right? So, you know, I'm with you. How many times have we seen things put in, and then it doesn't get adopted? And you know, because users permission users are simply not consulted. And, you know, I think that it's gonna be different people that are on this podcast, we're gonna have government people listening in, and then we're going to people that are trying to sell to the government that are listening in. And I think that, you know, as we sort of just have this discussion, I'd be interested to hear what you say is that we go forward, and we sort of sell things to the government. We focus on meetings, meetings, meetings, and CIOs and CTOs and relationships. And I think it's a good thing. Because it allows us to hear kind of like, what are the problems? Can we do? Can we address these problems with technology? But ultimately, the users are the ones that are the inheritors of whatever's decided. I have yet to see procurements come out where it says, you know, there's a heavy user experience element to it. It's always based on features and price, can it kind of deliver this capability for the lowest price period? And then Oh, you want people now you've got a different procurement process? So, should it change? I mean?
Keith Nakasone: Absolutely. So, when we look at the way, I'll go back in time, go for a lot of the procurements used to be done either by name brand, by a software or by hardware, etc. And then they say, well, this is everything is IBM, Xerox, whatever, right, right. It's just a price run off. In today's environment, where we're working towards how to solve mission outcomes, how to be more successful, we have to change almost flip that around. Now we have to look at get with the customers get with the users, and see what their challenges are, and how do we solve that problem moving forward. And the way we do that is through the DevSecOps process, because as we go through that process, and as we build upon technology, it may not just be one product solution, it may be a mix of products that come together to deliver a capability to solve mission challenges moving forward, we can't we no longer we're no longer looking at just lowest price technically acceptable, we're looking at best value, where you're injecting value add requirements, such as if you look at the multicloud deployment. And if there's ways to build cross services that can govern over multiple clouds in an environment. That is kind of the things that we need to do moving forward. Because workloads may exist in one area today, but it may be ported over to another. So, we have to be concerned about not only the aspects of workload movements, but we also have to be concerned about the technology. One technology may evolve over time, and may accelerate the process in solving mission challenges moving forward. So really getting with the customer getting with the users may change how we deliver procurements in the future. And so, we need to absolutely, we're seeing that right now. We're looking at services more upfront. And then you look at product solutions moving forward. Because there's potential of a product. There's a potential of multiple products, but you have to look at lifecycle management cost, how much is that going to cost me in the future? So, as we build out procurements and as we look at the cloud strategies in the hybrid in order to meet the customers, pandemic, it was a perfect example of how the workforce when remote Yeah, and now we have to look at how can we deliver capability faster, better, more effective.
Craig Bowman: In remote locations? Yeah,
Keith Nakasone: Absolutely.
Craig Bowman: Yeah. It's funny you are talking about you Your experience I think all of us remember when Microsoft released the new ribbon in Microsoft Word, I think productivity in the US government went down 87%, the day that we launched the ribbon in the US government, because nobody could find spellcheck, we couldn't find. It was like, why did they make this? I mean, now we look at it, and it's like, oh, we all know where everything is. But the day they made that change, the user experience was so different that nobody knew how to actually use the product. And I think you know, we're seeing that just in general, across the board, but you said something interesting. You said, multicloud. It's another buzzword bingo word multicloud. So, I'm gonna ask you this question. Everybody talks about multicloud. AI, and we're using workday. And then we're using Amazon or Google, whatever, Oracle, and Adobe, you name it. But is when we talk about hybrid cloud and multi cloud? Is anybody in the government actually doing? What multi cloud buzzword is supposed to mean? I don't even know what it means. I mean, I think I know what it means. But everybody uses the term multi cloud hybrid cloud, but is anybody actually doing it in your opinion?
Keith Nakasone: So, what we see is a lot of the, they're using multiple clouds, but they're built-in silos. And that's where it's sometimes it brings some, some problems, right, as well as challenges. So, if we say multiple clouds are being built within government, is that true multi cloud?
Craig Bowman: Are you asked me, no? So right? To me, the answer is no.
Keith Nakasone: So my design perspective, we continue to if the required technical requirements and specifications keep coming over to the acquisition workforce, with this, the silo build in the mindset bit, building multiple clouds, and if the procurements come out that way, the contracting officers are going to be challenged to really get to that multicloud, because all those underlying requirements have been built for a single stack, not multi. So, when we look at that we really have really interesting, we have to take a look at how things are being built from the technology side, as well as the acquisition, they need to come together to say, how can I deliver true multicloud environments?
Craig Bowman: That's interesting. So, if I, if I look, if I hear what you're saying, we're then. So, two things come to mind. One is huge cloud procurements just came out, got protested, got rewarded, right. In those cases, whenever there's a protest, usually one of the vendors is claiming that the testing rules were designed to favor one particular vendor. First of all, do you agree that that happens? And is that is that how do we fix that? Or do we fix it? What's your thought? And then I'll go to another question.
Keith Nakasone: Again, it's an education and training perspective with the people, right people know what they know. But if they if they're trained and understand the full spectrum of multicloud delivery, if they understand the different stacks, if they understand how the enterprise works together, and as well as understanding how to write requirements so that you get true to a true multicloud solution, your experience in the delivery will change over time. Again, it's a culture shift. So, it's the people, they know what they know, it, which they need to be upskilled rescaled, to understand how to deliver that capability, then we have to look at the processes that are in place. Right? If the if there's regulations that have been, or I should say policies and procedures that have been put in place to put some guard were guard rails around how to procure, but yet that is still in the old mindset, or legacy mindset, I should say. It's going to create some challenges, right? Yes, industry is already moving towards the adoption of delivering capability to solve it solutions. So, if industry is moving in that direction, now we have to say how do we change up both? It's not just the technology side, but it's also on the acquisition side. So, both side education and train? Also how do we revamp policies and procedures so we can execute cloud solutions?
Craig Bowman: Right? Yeah. So, you when you were asking that question, I think about the procurements that came out. And you know, most of the testing for these cloud contracts, basically said, hey, we're gonna rate you based on these kinds of performance metrics. And but nowhere in there did it say, hey, how easy is going to be for people to operate these clouds. And so, the protests, were all based on metrics, use that user experience support, and then the lifecycle is important. And he had the procurements coming out. Don't do that. That's right. They go, hey, if you've put this payload and you do XYZ, and ABC, and it cost this, you win the award, and then the users go, oh, I've been I've been trained in this version of technology now. Gotta switch into a different version of technology. Okay, so what you're saying. And what we're actually doing are two entirely different things.
Keith Nakasone: Absolutely. So when we look at, and we'll go back to DevSecOps, and when you build in stage your procurements with proof of concepts, pilots, and then scale, you're learning along the way you're building and you're bringing in those users, so that you can build upon those user experience, then you'll be able to see what is truly more effective and efficient for those users. If you don't have that build strategy, procurement strategy, alongside with the Agile it framework, you're gonna run into potentially some problems. And so, if you stage it, it's actually educating the workforce yet at the same time you're building in flight. So that that's why when we look at Dev SEC ops, it's truly a journey. It's not, it's not like buying a product upfront, and then figuring it out. It's actually you're bringing in those users to experience the process early on in the process to say, this works, this doesn't work, what can be improved, so you can work towards that process and release to production a lot quicker for things.
Craig Bowman: Wouldn't it be cool if we had a DevSecOps procurement process, like
Keith Nakasone: Absolutely.
Craig Bowman: You know, I mean, imagine if we did microservices, in procurement, there will be set up a con construct where people could come in and say, hey, I'm gonna deliver this one capability for X amount of dollars. And they just deliver it within the confines without having to put it out for bid because it's already pre-negotiated. Right? I mean, that'd be an interesting. So, we've covered multicloud, we've covered. Let's go to another buzzword bingo. Let's do Zero Trust. Zero Trust is the new buzzword is coming out in procurements everybody's looking to do it. So, what are your feelings on Zero Trust? Is anybody actually doing it? And what is the future for Zero Trust in your mind?
Keith Nakasone: So, from an ecosystem, you're not going to have this one single black box that's going to govern Zero Trust?
Craig Bowman: What?
Keith Nakasone: Right?
Craig Bowman: What I'm out of here, you know, what, what are you talking about? That's what they probably promised. Zero Trust is going to be like Skynet, it's going to like monitor and score everything in the world and make sure that all the security is perfect. Are you telling me it's not the case?
Keith Nakasone: No. It's going to be an ecosystem of collapse of many partners that are coming together to help develop and to deliver that capability. It's really miserable, minimizing that that threat and the risks in your entire IT infrastructure. So as capability is pushed out, as we look at IT solutions moving forward, you know, this black box doesn't exist, it's going to be an ecosystem of different capabilities that are put together to help us manage risk within an IT infrastructure. So, Zero Trust is a great framework in something to work towards. But it's going to be a lot of partnering a lot of collaboration, and looking at the entire product suites that each agency is using in their within their IT infrastructure.
Craig Bowman: So, if you know, I think that if we had a infrastructure, Keith, where it was locked down, right, if we had, if it was a lab, if it was an a, an organization with a very rigid and well-defined borders, and we could probably get to a unified Zero Trust model. But the likelihood of the US government, any agency, let alone the whole US government, being able to put anything in place that is division? I think Zero Trust is what I'm hearing you say, and I think I agree with, it's changing constantly, the environments changing. I mean, the literally IoT devices, how many IoT devices?
Keith Nakasone: There's going to be two digit billion IoT devices.
Craig Bowman: How do you Zero Trust something that's in flux constantly. And so, CIOs and CTOs that are on the podcast, need people selling that of selling this vision? You know, and of, hey, we got all these components to do Zero Trust? Is it like a little bit of a farce to think that that's going to happen? Are we really just saying improving security within layers to meet the guidelines of the government? And we just call that Zero Trust because it's a buzzword.
Keith Nakasone: Right. We have to start somewhere, right? Yes. And it's going to evolve over time. And with the with implementing, for example, 5g, right? It's going to enable more devices coming onto the network, we're using the low bandwidth speeds, the high bandwidth speeds, and we're going to be able to solve some of the rural area issues, right? For sure. Bandwidth perspective. So when we look at that capability, and we when we look at how we deliver again, we talk about mission first getting down to the user experience, and customer experience is going to be key So we're going to have to look at how the devices work at the edge going back to the core, where the how the data is secure moving both ways. So, at rest and in and in transit. So, when we look at those any capability that we put out, we definitely are going to have to take a look at the entire ecosystem, which is Zero Trust. However, with the implementation of another buzzword, artificial intelligence and machine.
Craig Bowman: Don't start, don't start. Don't start, you know, when we were out at sea. Don't even start you know what you went and brought up AI/ML, there is the last buzzword bingo. So, we've got multicloud. We've got Zero Trust. Now we got Am I am AI/ML. Alright, so AI ml, we were at CES, one of the CIOs stood up and said he didn't believe in artificial intelligence. So, I can understand why he would say that right? Because, you know, I'm not sure what prompted it. But I think that there's so much effort put into artificial intelligence, big contracts like Jake and programs like the Jake, are we emphasizing too much artificial intelligence and not really machine learning and automation? What's your thoughts?
Keith Nakasone: So, I think, you know, as we look at other countries, and we're looking at what they're doing with AI and ML, we see them on an accelerated journey of implementing AI and ML within their ecosystem. And we as from a United States of America perspective, we should private and public needs to be engaged and needs to see what's going on with on how it's being used within the ecosystem. And so as we look at our internal processes, as we look at how we can leverage it to one, when we talk about business intelligence, we have to look at how we can use the power of AI and ML to help us take huge amounts of data to make any kind of sense out of it, like other countries are doing. So, as we move forward, we have to start with the workforce. And we look at the people, we look at how we how we have to look at education and training, the best way to do it is have them injected into the process. Because if we bury our heads in the sand, and we only look at the what's going around what's happening in the other countries? Yeah, we from a workforce perspective, have to start educating and training upskilling and rescaling our people to learn how to use it, when to use it. And, you know, we talked about I think in the session, it was also brought up of how it can be injected, right where it can be injected. And so, some people think that we're going to turn the workforce into robots. No, we're not. We're not, we're not saying that we're going to turn the public sector into a robots. Right, what we're saying is that we have to figure out where we should be using artificial intelligence to help us.
Craig Bowman: Where do you think that is in your mind?
Keith Nakasone: So, I think, you know, when we look at progression of time, and we look at all the things that have happened, for example, if we're looking at, let's just say, environmental impacts, yeah. And you have hundreds of years of data. How can we use that to make better decisions on what's actually happening in the environment? Right, yeah, we look at our infrastructure system. We talk about our rails, we talk about our roadways, and if we see certain parts based on the data and the information that's gathered, so FEMA would be a big user and probably a good place to use CBP on the border could probably be absolutely transportation, right. CSA. Sure. Yeah. So, there's huge areas where we can take us those opportunities.
Craig Bowman: That's. Yeah. And that's AI ml for analyzing data. Absolutely. And then what about as we tie into automation? What are some of the areas where we could use artificial intelligence machine learning, i.e., either or right? Because they're not the same? To help with automating things. Because that's, I think, what everybody's thinking of right, you think AI? Is humans out of the loop? Right? You just discussed AI assisting humans and making decisions, supervised, supervised? Yeah, or even just AI that we learn from to make decisions ourselves? Right. What about machines learning and making their own decisions? Where do you see that helping?
Keith Nakasone: So, I think, you know, like, Call Centers?
Craig Bowman: That's good one, right? Yeah. Call Centers are good and basic.
Keith Nakasone: If we have.
Craig Bowman: That's a really good example.
Keith Nakasone: Hundreds of 1000s of calls that come in that answer the same question, and we're doing some work right now. Yeah, we should be leveraging artificial intelligence to do some of that to take the burden off.
Craig Bowman: And that can reduce cost. And in that case, we said we're not turning people into robots, but we could do more with fewer people.
Keith Nakasone: Absolutely. Yeah. And we can use them for more of the complex. So, we can handle those surges that come in.
Craig Bowman: Tier three support versus tier one support. So that's good example. Yeah, that's a really good example. So, on this podcast, like we've talked now about multicloud, we've talked about FedRAMP. We've talked about Zero Trust. And we talked about AI/ML. Did we miss any buzzwords or any of the DevSecOps? We got upset got us, right. So if I, if I wrap this up, like if I think about what I've heard, procurement, if I'm hearing you correctly needs to continue to shift to like the EIS, SBIR OTA models, you define a construct of a contract, where many contractors can come in and deliver micro deliveries that are mission focused, that those contracts, and the delivery of those things involve users, right, so that they are more of a part so that we don't end up releasing stuff that people just put in the corner and not use because that's never happened in the government bought something that nobody wanted to use, it's never happened. That DevSecOps, that we need to have our mission capabilities, more and more aligned to DevSecOps, so that we can release more often. Zero Trust is, is a is a goal, or it's a lofty idea that will not come to reality as a Skynet orbital black box, but rather, there's gonna be areas of Zero Trust applied with access given only when needed by who needs it, for what they need. And that could adjust many vendors working towards that. Let's see. And then the last one was AI/ML and an AI/ML, you're saying we aren't spending too much money in that area. But we need to focus that on areas that are gonna have the biggest impact, while we maybe put some things into, into competing with a foreign countries, right.
Keith Nakasone: And let's look at the data governance side first, right? So, we have to stage it in order for AI ML to be effective, we have to start from the very beginning of how is the data being used? And is there any kind of data governance within each of the respective.
Craig Bowman: I mean, one of the biggest problems with AI is that it creates bias the algorithms are easily create bias, groupthink, or you know, so I was working for CTO one time, and that was one of his biggest challenges. And in that area, you know, when you have AI making decisions on targeting, or you know, wherever we're going to put resources that affect people's lives, that becomes much more critical, make sure it's not creating biases within the AI algorithms. And I think we see that we've heard about it all over the place, right on all the social media platforms. So, it's clearly happening. So, I think that's why there always be some form of supervision, even within AI to make sure that it's grabbed. And so, we've covered a lot on this podcast. And so, if you haven't seen our last two podcasts, I'll say it again. We had Joe Beda, founder of Kubernetes, and Paul Puckett, from the army Software Factory. On the first one, you go back and listen to that. The last podcast we had was with Jeremiah Sanders and an engineer talking about Kessel Run the Air Force, DevSecOps program, and now we have Mr. Keith Nakasone, talked about what all that means from a contracting perspective, including getting his views on where the government is going. I think on our next podcast, I'm gonna I was gonna give you the name, but last time I got the name wrong. So, this time, I'm just going to say we're going to have another senior executive from the government come on, and he's going to talk to us about cybersecurity. So, thank you for joining mission first. Until next time, I'm Craig Bowman, your interviewer and thank you very much, Mr. Keith. Nakasone,
Keith Nakasone: Alright.
Craig Bowman: For coming on the podcast.
Keith Nakasone: Thanks a lot.
Craig Bowman: Take care everybody.
Keith Nakasone: Bye.