subscribe for Updates

Platform Hacks: Getting the Right Tools in the Right Environment


Duong Hang, Deputy Director of Platform One, shares an inspiring journey of transforming how the Department of Defense (DoD) delivers warfighting capabilities through innovative platforms and a culture focused on agility and collaboration. With an emphasis on both technical tools and organizational frameworks, Duong presents how Platform One is shaping the future of GovTech and the DoD.


Duong Hang (00:14):

My name is Duong Hang. I am the Deputy Director for Platform One. Even though we put the word DOD in front of it. We support the DOD, but we support more than the DOD apparently, because I'll get into that a little bit. But we are actually an Air Force organization, surprisingly. I know Ms. Knausenberger yesterday mentioned Platform One. So we've been around for about three years now and we're actually getting stronger and better every day.


I've been with the program for about two and a half years now. I was not a founder of the organization, but I certainly inherited a lot from the team. And so the things I'm presenting to you today, I can't claim credit for everything that has happened in the past. A lot of this stuff is from the people that started the program, and we've been carrying it forward and we're building upon it. So I just want to recognize those people that started it today, and some of them are here in this audience.


So I'm here to talk about platforms and what they are, why do we need them, the tools that support that, and how do we actually do that in the government, at least the way we do it in Platform One. So not to say that there's one way of doing business, but this is just one perspective from where I sit, presenting to you how we do it today.


So anyone, has anyone been to this, the Sphere? Have you heard of, you guys have seen this, right? You've been there? I have not had the pleasure of seeing it. I've seen many pictures. But reading about it, it just opened up back in September, and right now it's reported to be costing about $2.3 billion. They had some delays, they had a lot of construction delays. Supply chain costs have gone up. It holds about 18,000 people sitting and 20,000 people if they're all standing. That's the platform. This is a venue for entertainment. So right now U2 is out there doing their performances. I heard that they do movies out there as well. So really cool venue. I'd love to go down someday. Maybe you can too.


But the tools are the things that this place provides. You see the physical dimensions there. It's the largest in the world, 875,000 square feet, and they have 16K resolution wrapped around screens across, and that's all screens, right? That's all giant LCD screens. And each of the screens, interestingly enough, have speakers that are mounted behind those. And then you also have speakers underneath the floor. So it's like this, call it this four dimensional experience.


And so this is like the tools that they need to actually put together in order to have this amazing experience. And so I'm using this as kind of a shoot off of what does that mean to be a platform in our world? So Evan Bottcher threw out this definition. I thought it was actually pretty good. Digital platform, foundations for a self-service, APIs, or application programming interfaces, tools, services, knowledge, and support, which are arranged as a compelling internal product. So that means that you can have autonomous delivery teams that are using the platform and they can actually deliver products at a higher pace and reduced coordination.


So obviously I highlighted some of the terms that I thought were important of what a platform should do. And so some of these are maybe not new to you. A lot of you guys are here because you guys are doing this work. But I think it's important to understand we have the same vision here. That's what we want to do with Platform One.


So why do we need them? And so the domain complexity, the applications that exist, the software that we build, have to live in the modern world. They have to live in the real world. They have to actually provide value to people. And so people are complicated. The world's complicated. There's complexity there. Then you've got the business that we're in, most of us are in, is the technical side. It's obviously complex, much more complex than when I started, much more complex than when our fathers and grandfathers started, or grandmothers.


And of course there's also for the government space, there's a regulatory complexity, policies and process and things that we have to follow that unlike many other private companies necessarily don't have to follow as much. But it's a lot more complicated for us in the government to do things that we need to do.


And of course, individual cognitive load. Now I made sure I had two different icons, the left hemisphere and the right hemisphere. Both sides are important in what we do, and I know sometimes we emphasize the left side or the right side more. What's important for us as a platform delivery organization is that we appeal to both sides. And that's very complicated, both in terms of the people that do the work within the platform and the people that can consume the platform and use it. It can be very complicated. So that's why we need to have something to help manage that complexity.


So this is our logo. You might've seen it before. I'm wearing the T-shirt, obviously. It's not Grogu. It's a little green alien. That's what we say. And our mission is to transform how the DOD delivers the war fighting capabilities that it needs through innovation. But innovation is not just the tools, it's not just the specific technical tools that we are very comfortable with in the technology space, but we focus on the frameworks and the culture, the collaboration culture part, and the process technologies that then support that.


And it's actually, we spent many, many months working through how do we update this mission? And this is what we came to as a group. And the order in which these are presented is what's important, because you can have all the great tools and technologies at your disposal, but if you have nothing above that, everything above that, if you don't have that, you're either going to overspend or underutilize what you have and a lot of waste will happen. So that's why it's important in this order to focus on. So all these are tools in some ways, but the bottom tool, the technology tool, is actually the least important.


And so talking about frameworks, starting with that tooling, what does that mean? So for us, Agile, we're all familiar with Agile. I've spent many, many years thinking about this. And I grew up in the DOD, as an Air Force officer starting off and then becoming an Army civilian, then a Navy civilian, and now an Air Force civilian again. And being in both tactical units as well as program offices, a lot of times where agile was always a fancy word growing up, it was something that only the tech rows did. It was only something the commercial side did. It was never something that the government would do. And I had to think about what does that really mean as a practitioner in the government? What does that mean?


And so reading a lot of books and working in the space, agility being at the very top, you can't be agile unless you're learning. If you're just randomly doing stuff and you're not learning from it, you can't be really agile and be effective. But in order to learn, you have to do it quickly, because if you're learning slowly, then what you're learning is irrelevant. So you have to learn quickly.


And then in order to be fast, you have to have autonomy. So if you don't have the ability to do things on your own, then you can't go quickly because a lot of dependencies will occur. Then in order to have true autonomy, you have to have real choices. If you don't have choice and you only have one way of doing business or you only have one tool choice, you have only one process, you really don't have autonomy. And then finally, in order to have choice, you have to have modularity, meaning you have things that you can replace if you need to.


I have this as sort of a hierarchy because this is what builds on top of each other. And I realized this is not just from a tooling perspective, but it's also from an organizational perspective, that you need to have both in order to actually align that. So it's people first cultures and processes on the left and technology and tools on the right. That's what you need at every part of this level to do so. So I'll explain how we do that in Platform One.


So starting with this, we talked about the frameworks and now the collaboration culture. So this is a double infinity symbol. And really what it's showing is that we are a very small organization. We're about 250 people, both government and contractors. We don't plan on it being much bigger than that. And one of the reasons is because we feel like we only provide one solution that supports many use cases, but not all use cases. What we really depend on is partnerships with the defense industrial base on your left, the open source community, and the commercial to provide the talent and the software. And we bring that in and we actually adjudicate that. We check it out, make sure it's good to go for DOD use. And then we provide enhancements to that and we bring it over to the Department of Air Force, the joint program offices, and we call it airman coders. It's the Soldiers or the Sailors or Airmen that are actually at the field units actually doing the hard work.


And so now it's not just a one-way transfer because those innovations that are happening at the tactical level can be brought in, and we think that we want to bring that in back to the community for sharing and then bring it back. And then sometimes some of that can be pushed up back to the commercial open source. So this is a model that we feel this is how we as a community can scale the innovations that we all provide because we are often talking about innovations as if the goal is a product. We believe in Platform One is the innovation is just a byproduct of the process and the community and the orchestration of that ecosystem that produces innovations. So innovation is a byproduct, and that's what we focus on, is the actual environment to provide innovative opportunities for people. And so what this is is a mission focused, secure, tailorable, government orchestrated vendor neutral ecosystem that we're shooting for.


So our four product lines in Platform One, and I try to summarize it in various bumper sticker words here. We have four of them. Cloud Native Access Point, or CNAP, is the one on the upper left, and that's our zero trust platform that securely connects everything that we do because security is baked in, and we do that from the very start.


We also have Iron Bank, which is on the lower left. And traditionally, back in a few, maybe about a year ago, you've heard me talking about this, it was like, hey, we hardened containers and we provide a place for you to download. We've broadened that mission. So Iron Bank is now really a place for secure supply chain. And so we're doing a lot of things that are more focused on not just the software containers, but where they come from, who are the actors that update them, especially open source, and then how do we make that available to other people in other classification levels. So those are the things that we're working on there. It's really about supply chain and not necessarily just hardening containers.


But in order to have to build that platform of those containers, those pieces and parts of tools, you need something to orchestrate all that. And that's where Big Bang comes in. And that's where we say deploy platforms. You can take a lot of this stuff from Big Bang. It's a lot of the infrastructures code, configurations code, it's all the recipe books, I call it that, for the computer to take the certain parts and pieces in Iron Bank and then basically lay out your platform that you need to actually deliver your DevSecOps platform. That's what they do.


And then finally for Party Bus, for those that don't have the ability to really say, hey, I want to run my own platform, I can't do all that, I just want to code and develop applications, we have the Party Bus and that's where you can develop and operate. And so what this is showing you is there's different flavors of DevSecOps capabilities that you can use and deploy in order to be on your journey no matter what type of organization you are, whether you're just someone that the tactical level needs to develop an application or you're someone that needs to run a very bespoke platform and you want to still use safe and secure containers and capabilities, you can do so as well.


And so to put it together in perspective, so this is the different tools for different use cases. And if you look at the very tops, secure access, getting hard containers components, build templates, build pipelines, run your apps, those are the big rock kind of things that we see if you're going to actually deliver capabilities. And what we're trying to explain to a lot of organizations, especially ones that don't really understand the technology, is like, look man, you could do it all yourself. That's the very first thing up there. You can do it all yourself like you've done before and you have a lot of control over that, but you're going to spend a lot of money on something that someone else has already done. It's really not your mission, but you felt like you had to do it, so you've got to do it. But you can do it all yourself.


You have the option of adopting our CNAP as a service, and that's the next line there, where then you can start saying, hey, I've secured my infrastructure and I can actually focus on the things I need to build the platform. Or if you want to go next step down and say, hey, you know what, I don't want to focus on those components. I want to actually just take the components out there in Iron Bank and now I can actually focus on just taking those ingredients from the shelf in Iron Bank, and now I want to design my own platform. You can do that.


And if you want to go further, you say, hey, you know what, I just need to run a platform. I don't need to figure all that out. You can adopt Big Bang for free. And then finally, if you're like, hey, look, I don't have to do all that stuff, I just want to build my applications, you have Party Bus. Again, the thing is as you step down in this area, you have less control because someone else is giving you the stuff that's already there. It's already been hardened, it's already been secured, but it's also been very, it's highly regulated, but you don't have to pay your own way to do that.


You can just inherit what we've developed and you can then focus on your mission. You can dedicate not just your money, but really the people that work for you, the talent, that's really the resources that are most scarce in GovTech, is the talent, and focusing them on the hard problems, actually developing an application, a software application that actually is the game changer, difference maker in your organization and for the war fighter, that's what matters. And that's what we want to do for everyone.


Of course, I can't talk about tools without actually talking about tools. So for Big Bang, we have curated several different pieces of tools out there in all these areas. It might be hard to read, but there's the three roles there with the core add-ons and community. And so how we've done this in terms of trying to make sure that we build something that is useful for most people, but tailorable for others, is to build a set of core pieces of what a platform should look like on that first row, that green row. And you can see where there's service meshing, there's monitoring, logging, tracing, security, compliance, and exit control. All these things are available to you today in terms of being able to download and have Big Bang orchestrate that for you. This is what you need to run a Big Bang platform. And whether you're running the platform as a development only platform, a staging or an operations, you need the core piece.


You can see that in some of these areas we have multiple tools. So again, vendor neutral is a key principle and having the ability to swap out what you need. And one of the cool things is about four or five months ago, we actually are able to deploy or anyone can deploy Big Bang Core with only open source components. So you can actually do it for free, which is really awesome. So that's one thing that we push for.


Now, obviously Core is not enough. So if you're a developer, obviously you're going to need some stuff beyond what Core provides. That's where the add-ons come in and you can see there's a bunch of stuff there. Some of it is actually you need it for when you're doing your DevSecOps pipelines. The add-ons themselves, we manage as well. And though we have some partnerships, you can see there's a lot of companies out there that we work with quite closely. And then finally, there's community, and this is just a portion of it where the community then can also add to Big Bang and provide their management of capabilities. This is a snapshot of what that is. And so you can go to the, and you can read more about what Big Bang is and what it can do for you. But this is the main area that I want to focus on or talk about for Platform One.


So what does that all mean? What's the net effect on this? And so this is actually a slide that I can't claim credit for. This is actually from Lieutenant General Jack Shanahan. He was the ex-JAIC director, or now CDAO, and also the CTO for CIA. And they have presented this for many, many senior executives, saying why does platforms matter? Why do we have that in enterprise capability? So on the left side you can see that a stove piped capability, when you have applications of platforms and infrastructure that is stove piped, not only are you dealing with multiple costs, but you have multiple ATOs and multiple AOs that have to approve everything. So your total costs are higher, and every time you add a new platform or a new capability, it's just more expensive.


On the right side, you can see there's benefits of that enterprise scaling. And the focus is that having an ATO that covers the whole thing, and for us in Platform One with Party Bus, we have a continuous ATO, just like you've heard that before, where we allow, as long as you build an application and use our tools and pass all the checks within Party Bus, you can basically operate your applications within 26 days of development. So very, very fast, compared to the nine to 18 months of ATO work that many of you in GovTech are familiar with. But all platforms are wrong but some are useful. So obviously a play on the idea of models.


And so I talked about Platform One, it's not the only platform out there. And some of those shortcomings and things that, and this is why partnership matters, we still have some challenges for continuous delivery. Network classification air gaps. That's a challenge today. We're working on it. I know some organizations have figured out their path, but it's still a challenge for many people of having to go from unclassed to classified is a problem. If you have an air gap environment going to obviously a platform that doesn't connect to the internet, it's a problem.


Legacy software integration support, that's a challenge as well. Embedded hardware, you have the F35 little logo there. So a lot of embedded systems can't do what we do today, which is something that we're tackling, or partnering with others to tackle. Offering system requirements, software acceptance procedures, and then federal acquisition culture. All these things are still hurdles that we need help, we need you guys to help us with, we need to help each other with, and we'd love to partner with you if you're able to do so.


And then from an acquisition culture, this is a book, I love this book, and if anybody has read this book, it's just an amazing book. One of the things that I want to touch upon from yesterday's discussion with Adam talking about the workforce is one of the reasons why, at least according to Jennifer Pahlka doing some research, that we have this sort of culture of technologists and people that do technical work are not important is because it was something that since the fifties, there was a change in view, that the work that you do as a technologist or things they do is basically not as important as people that do policy and processes.


And so how many of you looking at USA jobs see GS-15 jobs, GS-14 jobs here in DC, they're all policy guys and people that just know how to talk, but they don't actually deliver. So this is one of the reasons why we have the challenges we have today. We've hollowed out our technical talent in our program offices, our headquarters. And so people that are in charge today don't really understand technology. And so what happens is they're very risk averse, they focus on process over outcomes. And then what happens is in a program office, a lot of times they're like, "Hey, look, I just want to outsource to one big prime." Nothing against the big primes, but it's a challenge. One big prime, and we're going to have multiple years for a performance, have long lead times, we have less competitions, incremental improvements and innovations. Those are some things that she talks about that it's a challenge in the government.


And fortunately we have tools in the federal acquisition tool set now. 18F is an awesome organization in GSA. These are some of the things that they looked at, outcomes over products, talent over products, government orchestrated multiple part primes. So having a government organization that has many, many different companies working together and integrated teams. You have contract vehicles are pre-negotiated. So you can go quickly when you get these new contracts. And you have short pops, short periods of performance. So you make sure that people are just getting their job done. If they're not able to do their job or do it well as a company, then you can pivot to another company. So those are some things that they've looked at experiment with. We've actually implemented within P1 as well.


And of course, our talent, I don't want to say, let me go back one, I don't want to say too much here because I know there's others presenters that talk about the talent piece, but I would say that this is one of my most passionate things about technology and government today, is our people in the government are completely underappreciated. And we need more people like you to join us because we can spend money on everything from missiles and bombs and all kinds of stuff, but the people that actually do the innovation, that have the great ideas to make the lasting changes in the organization, that's the people. We need people like that to join the government.


So where do we go from here? Some couple things I want to cover here. HCD, human centered design, a lot of these you know, but I think this is where it starts. It obviously starts with empathy. Focus on our strengths, not our weaknesses. And that goes to people that are both practitioners as well as the people that are four stars, because they have a limited time and capacity and experience. So be empathetic to that. We want to lower the cognitive barriers, whether you're doing it from technology or you're doing it from the acquisition side of things. That means psychological safety and reducing the complexity on that. And we want to design to incentives and not to penalize. So you want to design for positive behavior.


From an organizer's perspective, which I'll touch upon a little bit later, there's that. And then, of course, tooling. That includes holistically your technical acquisitions, processes, policies, they're all tools. You want to balance the safety and security over effectiveness. That's always a challenge for us. And the thing is that time is the only tool or resource that no one can control or renew. If you've grown up in old school project management, they would say it's the iron triangle of cost, schedule, performance. And you can do two out of three. That's a false statement. Time is something you cannot trade off. So that's always going to be a constant. And you have to make sure that everything, those two are balanced against that.


And so just to touch upon organizations, how many of you, you've seen this, very bespoke, funny, showing the models of different organizations. I can't take credit for it. There's where I got it from, on the link there. But what this talks about is organizations. This is another book I highly recommend. This book, Team Topology, really talks about how, it really focuses on Conway's Law, which is that an organization, the way it's designed and is organized within its foundations, basically builds the system that mimics the communication strategies of that organization.


And so for basically how your organizational culture is really vendors reflecting the systems you build. And so if you look at the DOD and all the bespoke stove piped systems, they're very much a product of a program office that has their own funding. They don't talk to anyone else. They deliver it the way they deliver it. And that's why it is what it is.


And so they offer ideas about a reverse Conway maneuver, which is basically in order to change the system, change your organization first and your system will be a byproduct of that. And they talk about interaction styles, which this is where your tools come in to help support collaboration styles, acts as a service, style of communication between teams, and then facilitating teams.


So just in closing, I'm a little over time, but I want to say a couple things just to take away. A bad workman blames his tools. A bad leader blames others. Your environment shapes your imagination. Your tools are limited by your imagination. And leaders change your environment. So if you think about it and if you connect all the dots, it's up to us as leaders, whether you're down here or up here, we're the ones that are going to be the catalyst for change. And we change the environment, we change how we talk, we change how we organize. That would change the tools that we use and that will change and transform the systems and outcomes that we're actually looking for. So thank you so much. I really appreciate your time.