>> Narrator: From around the globe, it's theCUBE with digital coverage of AWS Public Sector Partner Awards brought to you by Amazon web services. >> Hi, and welcome back, I'm Stu Miniman and we're here at the AWS Public Sector. Their Partner Awards, really enjoying this. We get to talk to some of the diverse ecosystem as well as they've all brought on their customers, some really phenomenal case studies. Happy to welcome to the program two first time guests. First of all, we have Jared Bell, he's the Chief Engineer of self response, operational readiness at T-Rex Solutions and T-Rex is the award winner for the most customer obsessed mission-based in Fed Civ. So Jared, congratulations to you and the T-Rex team and also joining him, his customer Michael Thieme, he's the Assistant Director for the Decennial Census Program systems and contracts for the US Census Bureau, thank you so much both for joining us. >> Good to be here. >> All right, Jared, if we could start with you, as I said, you're an award winner, you sit in the Fed Civ space, you've brought us to the Census Bureau, which most people understand the importance of that government program coming up on that, you know, every 10 year we've been hearing, you know, TV and radio ads talking about it, but Jared, if you could just give us a thumbnail of T-Rex and what you do in the AWS ecosystem. >> So yeah again, my name's Jared Bell and I work for T-Rex Solutions. T-Rex is a mid tier IT federal contracting company in Southern Maryland, recently graduated from hubs on status, and so T-Rex really focuses on four key areas, infrastructure in Cloud modernization, cybersecurity, and active cyber defense, big data management and analytics, and then overall enterprise system integration. And so we've been, you know, AWS partner for quite some time now and with decennial, you know, we got to really exercise a lot of the bells and whistles that are out there and really put it all to the test. >> All right, well, Michael, you know, so many people in IT, we talk about the peaks and valleys that we have, not too many companies in our organization say, well, we know exactly, you know, that 10 year spike of activity that we're going to have, I know there's lots of work that goes on beyond that, but it tells a little bit , your role inside the Census Bureau and what's under your purview. >> Yes, the Census Bureau, is actually does hundreds of surveys every year, but the decennial census is our sort of our main flagship activity. And I am the Assistant Director under our Associate Director for the IT and for the contracts for the decennial census. >> Wonderful and if you could tell us a little bit the project that you're working on, that eventually pulled T-Rex in. >> Sure. This is the 2020 census and the challenge of the 2020 census is we've done the census since 1790 in the United States. It's a pillar, a foundation of our democracy, and this was the most technologically advanced census we've ever done. Actually up until 2020, we have done our censuses mostly by pen, paper, and pencil. And this is a census where we opened up the internet for people to respond from home. We can have people respond on the phone, people can respond with an iPhone or an Android device. We tried to make it as easy as possible and as secure as possible for people to respond to the census where they were and we wanted to meet the respondent where they were. >> All right. So Jared, I'd love you to chime in here 'cause I'm here and talking about, you know, the technology adoption, you know, how much was already in plans there, where did T-Rex intersect with this census activity? >> Yeah. So, you know, census deserves a lot of credit for their kind of innovative approach with this technical integrator contract, which T-Rex was fortunate enough to win. When we came in, you know, we were just wrapping up the 2018 test. we really only had 18 months to go from start to, you know, a live operational tests to prepare for 2020. And it was really exciting to be brought in on such a large mission critical project and this is one of the largest federal IT products in the Cloud to date. And so, you know, when we came in, we had to really, you know, bring together a whole lot of solutions. I mean, the internet self response, which is what we're going to to talk about today was one of the major components. But we really had a lot of other activities that we had to engage in. You know, we had to design and prepare an IT solution to support 260 field offices, 16,000 field staff, 400,000 mobile devices and users that were going to go out and knock on doors for a numeration. So it was real6ly a big effort that we were honored to be a part of, you know, and on top of that, T-Rex actually brought to the table, a lot of its past experience with cybersecurity and active cyber defense, also, you know, because of the importance of all this data, you know, we had the role in security all throughout, and I think T-Rex was prepared for that and did a great job. And then, you know, overall I think that, not necessarily directly to your question, but I think, y6ou know, one of the things that we were able to do to make ourselves successful and to really engage with the census Bureau and be effective with our stakeholders was that we really build a culture of decennial within the technical integrator, you know, we had brown bags and working sessions to really teach the team the importance of the decennial, you know, not just as a career move, but also as a important activity for our country. And so I think that that really helped the team, you know, internalize that mission and really drove kind of our dedication to the census mission and really made us effective and again, a lot of the T-Rex leadership had a lot of experience there from past decennials and so they really brought that mindset to the team and I think it really paid off. >> Michael, if you could bring us inside a little a bit the project, you know, 18 months, obviously you have a specific deadline you need to hit, for that help us understand kind of the architectural considerations that you had there, any concerns that you had and I have to imagine that just the global activities, the impacts of COVID-19 has impacted some of the end stage, if you will, activities here in 2020. >> Absolutely. Yeah. The decennial census is, I believe a very unique IT problem. We have essentially 10 months out of the decade that we have to scale up to gigantic and then scale back down to run the rest of the Census Bureau's activities. But our project, you know, every year ending in zero, April 1st is census day. Now April 1st continued to be census day in 2020, but we also had COVID essentially taking over virtually everything in this country and in fact in the world. So, the way that we set up to do the census with the Cloud and with the IT approach and modernization that we took, actually, frankly, very luckily enabled us to kind of get through this whole thing. Now, we haven't had, Jared discussed a little bit the fact that we're here to talk about our internet self response, we haven't had one second of downtime for our response. We've taken 77 million. I think even more than 78 million responses from households, out of the 140 million households in the United States, we've gotten 77 million people to respond on our internet site without one second of downtime, a good user experience, a good supportability, but the project has always been the same. It's just this time, we're actually doing it with much more technology and hopefully the way that the Cloud has supported us will prove to be really effective for the COVID-19 situation. Because we've had changes in our plans, difference in timeframes, we are actually not even going into the field, or we're just starting to go into the field these next few weeks where we would have almost been coming out of the field at this time. So that flexibility, that expandability, that elasticity, that being in the Cloud gives all of our IT capabilities was really valuable this time. >> Well, Jared, I'm wondering if you can comment on that. All of the things that Michael just said, you know, seem like, you know, they are just the spotlight pieces that I looked at Cloud for. You know, being able to scale on demand, being able to use what I need when I need it, and then dial things down when I don't, and especially, you know, I don't want to have to, you know, I want to limit how much people actually need to get involved. So help understand a little bit, you know, what AWS services underneath, we're supporting this and anything else around the Cloud deployment. >> Sure, yeah. Michael is spot on. I mean, the cloud is tailor made for our operation and activity here. You know, I think all told, we use over 30 of the AWS FedRAMP solutions in standing up our environment across all those 52 system of systems that we were working with. You know, just to name a few, I mean, internet self response alone, you're relying heavily on auto scaling groups, elastic load balancers, you know, we relied a lot on Lambda Functions, DynamoDB. We're one of the first adopters through DynamoDB global tables, which we use for a session persistence across regions. And then on top of that, you know, the data was all flowing down into RDS databases and then from there to, you know, the census data Lake, which was built on EMR and Elasticsearch capabilities, and that's just to name a couple. I mean, you know, we had, we ran the gamut of AWS services to make all this work and they really helped us accelerate. And as Michael said, you know, we stood this up expecting to be working together in a war room, watching everything hand in hand, and because of the way we, were able to architect it in partnership with AWS, we all had to go out and stay at home, you know, the infrastructure remain rock solid. We can have to worry about, you know, being hands on with the equipment and, you know, again, the ability to automate and integrate with those solutions Cloud formation and things like that really let us keep a small agile team of, you know, DevSecOps there to handle the deployments. And we were doing full scale deployments with, you know, one or two people in the middle of the night without any problems. So it really streamlined things for us and helped us keep a tight natural, sure. >> Michael, I'm curious about what kind of training your team need to go through to take advantage of this solution. So from bringing it up to the ripple effect, as you said, you're only now starting to look at who would go into the field who uses devices and the like, so help us understand really the human aspect of undergoing this technology. >> Sure. Now, the census always has to ramp up this sort of immediate workforce. We hire, we actually process over 3 million people through, I think, 3.9 million people applied to work for the Census Bureau. And each decade we have to come up with a training program and actually training sites all over the country and the IT to support those. Now, again, modernization for the 2020 census, didn't only involve the things like our internet self response, it also involves our training. We have all online training now, we used to have what we called verbatim training, where we had individual teachers all over the country in places like libraries, essentially reading text exactly the same way to exactly over and over again to our, to the people that we trained. But now it's all electronic, it allows us to, and this goes to the COVID situation as well, it allows us to bring only three people in at a time to do training. Essentially get them started with our device that we have them use when they're knocking on doors and then go home and do the training, and then come back to work with us all with a minimal contact, human contact, sort of a model. And that, even though we designed it differently, the way that we set the technology of this time allowed us to change that design very quickly, get people trained, not essentially stop the census. We essentially had to slow it down because we weren't sure exactly when it was going to be safe to go knocking on door to door, but we were able to do the training and all of that worked and continues to work phenomenally. >> Wonderful. Jared, I wonder if you've got any lessons learned from working with the census group that might be applicable to kind of, the broader customers out there? >> Oh, sure. Well, working with the census, you know, it was really a great group to work with. I mean, one of the few groups I worked with who have such a clear vision and understanding of what they want their final outcome to be, I think again, you know, for us the internalization of the decennial mission, right? It's so big, it's so important. I think that because we adopted it early on we felt that we were true partners with census, we had a lot of credibility with our counterparts and I think that they understood that we were in it with them together and that was really important. I would also say that, you know, because we're talking about the go Cloud solutions that we worked, you know, we also engage heavily with the AWS engineering group and in partnership with them, you know, we relied on the infrastructure event management services they offer and was able to give us a lot of great insight into our architecture and our systems and monitoring to really make us feel like we were ready for the big show when the time came. So, you know, I think for me, another lesson learned there was that, you know, the Cloud providers like AWS, they're not just a vendor, they're a partner and I think that now going forward, we'll continue to engage with those partners early and often. >> Michael the question I have for you is, you know, what would you say to your peers? What lessons did you have learned and how much of what you've done for the census, do you think it will be applicable to all those other surveys that you do in between the big 10 year surveys? >> All right. I think we have actually set a good milestone for the rest of the Census Bureau, that the modernization that the 2020 census has allowed since it is our flagship really is something that we hope we can continue through the decade and into the next census, as a matter of fact. But I think one of the big lessons learned I wanted to talk about was we have always struggled with disaster recovery. And one of the things that having the Cloud and our partners in the Cloud has helped us do is essentially take advantage of the resilience of the Cloud. So there are data centers all over the country. If ever had a downtime somewhere, we knew that we were going to be able to stay up. For the decennial census, we've never had the budget to pay for a persistent disaster recovery. And the Cloud essentially gives us that kind of capability. Jared talked a lot about security. I think we have taken our security posture to a whole different level, something that allowed us to essentially, as I said before, keep our internet self response free of hacks and breaches through this whole process and through a much longer process than we even intended to keep it open. So, there's a lot here that I think we want to bring into the next decade, a lot that we want to continue, and we want the census to essentially stay as modern as it has become for 2020. >> Well, I will tell you personally Michael, I did take the census online, it was really easy to do, and I'll definitely recommend if they haven't already, everybody listening out there so important that you participate in the census so that they have complete data. So, Michael, Jared, thank you so much. Jared, congratulations to your team for winning the award and you know, such a great customer. Michael, thank you so much for what you and your team are doing. We Appreciate all that's being done, especially in these challenging times. >> Thank you and thanks for doing the census. >> All right and stay tuned for more coverage of the AWS public sector partner award I'm Stu Miniman and thank you for watching theCUBE. (upbeat music)

(light music) >> Welcome back, everybody. Jeff Frick here with theCUBE. We're having a CUBE Conversations at our Palo Alto studio, getting off the road, getting ready for the holidays. A little bit of break in the conference action, and we're excited to have our next guest, David Greene. He is the CEO of ZeroStack. David, great to see you. >> Thank you. Good to be here, Jeff. >> Again, for those who aren't that familiar with the company, give us a quick and dirty on ZeroStack. >> ZeroStack is a software company based here in Mountain View. We're building a new kind of private cloud infrastructure. The idea is to use automation to simplify operations while still keeping IT in control of that infrastructure. We're really trying to deliver what a public cloud-like experience to users while keeping IT in charge and in control. >> It's funny. your website, I've been doing my background work, self-driving private clouds like the autonomous private cloud. >> That's kind of where we're trying to get to, right? The idea is that too much of the work that IT has to do is bogged down in day-to-day administrative task and manual operations and working with boxes. Instead, we can start to bring automation and machine learning and intelligence there, IT can move forward on the things that are more important and move faster more importantly and support the rest of the business. >> And Andy Jassy might argue with me, but I don't think he would necessarily, but part of the genesis of public cloud was this friction between the dev and the ops because I'm on the hook, I have to develop a new application, I don't have time to wait for the IT guy to provision me a new box, I don't exactly know what kind of box is that. Various questions, you know? >> It's a ticket you submit, and then in two weeks, we'll get back to you, >> Then the operating system Then you got to order it from Michael Bell and then it's coming in the mail. So really it's that tension that probably created that demand for a quickly provision easy to provision, swipe my credit card and it just appears on my desktop. So that's the piece of cloud you are trying to emulate. >> Exactly, I think that's a good analogy right, it's interesting when you go back to the origin of Dev Ops, the idea was that the developers would take care of operations as part of building the application, as part of the application's life cycle. >> Right. In reality, I have yet to meet an application developer who has any any interest in operations. So really, Dev Ops today is about how does the IT organization, better support the development organization and the application teams. In context, keep in mind that every organization today is becoming a software company because every customer interaction, every business process, every service delivery is somehow being instantiated in a piece of software that the organization is rolling out to them, right? So when the business is driven by software, the developers need to move at the speed of the business, now how does IT keep up? And that's where this idea of IT being able to provide the kind of that experience that you talked about, the swipe and go, becomes so critical. >> But at the same time for all the reasons that have been well documented, there's just certain stuff that's not appropriate for a public cloud, but what we are talking about has nothing to do with the appropriateness of whether it is or isn't. It's really trying to deliver the benefits of that type of a working model to whatever your infrastructure is, and in your case it's private cloud. It's my own data, >> Exactly. It's my own infrastructure. >> Exactly. >> I think it's important to acknowledge that there have been people in the industry who have said that the whole world was going to become apart of a cloud, right? And I think our view is that that's not that case, and as you said, for a variety of good reasons. There are some really important external factors in the world right now that say that's not the case. As you start to see the current political climate, the current geopolitical climate, you got more and more barriers going up around the world that says my data shall be mine, shall remain in my country, it's not going anywhere else. Every time something rash and unexpected happens I have another set of customers in some European countries saying that my data will never leave this country. So that's one external factor that says you got to keep control of your data and your work code, right? There's also a set of internal factors that says people are discovering that as is always the case, it's much more expensive over time to rent than to own. You have houses and you have hotels. You don't live in a hotel, you use a hotel when you need it, you go live in your house. As public cloud is spreading out more mature, people are realizing that there's a need to bring that home to better control the cost around that. I think there's also a human dimension to this too, which is that you have an entire ecosystem of IT professionals with deep expertise, deep knowledge that is only relevant and only applicable in a world that still has a notion of on premise's private cloud infrastructure. And you can be sure that those people are going to do their best to make sure that their livelihood, their careers, that it all stays relevant. We see all those dimensions playing out as kind of motivators for organizations to want a private cloud. The flip side's been, it's typically been hard. And I would argue that the appeal of public clouds and, the users like it, but it's easy. So by trying to bring what we do, with that self service view and add an ease of operations around it, now IT can participate fully in this ecosystem. >> Now it's interesting, obviously the incumbents aren't not just taking this lying down. All these big infrastructure providers like, Dell EMC and HP have been pitching hybrid cloud. They accept that some stuff is going to be in the public cloud, so they're also trying to put in place to make their infrastructure more cloud like. So what are you guys doing differently than say, what might be coming down the line from Dell EMC, or coming down the line from HP, in terms of your customers point of view? What I think essentially is that we're going to work with a Dell EMC or a HP or a Lenovo, whoever, as part of that infrastructure right. Every cloud at the end of the day needs a set of computer resources, a set of storage resources, a set of networking resources, and those companies you've listed make excellent products in those areas, and we are going to use those to apply our software topfit. Where we see the bigger gaps around past cloud solutions has been around on the software layer. So look at some of the generations that have existed. You have VMWare, which is kind of the point of reference. A VMWare cloud is complicated, it's multiple products that are acquired over time, different architectures, different code bases, they don't integrate together. Hard to hire people, they're expensive, they're hard to keep, those challenges. What have we tried to do to make them better? You've had an open source alternative that came with OpenStack, okay? Better software, lower cost software, but even more difficult to operate. At least that's the feedback we get from our customers. I love the idea of Openstack, it's too hard to keep it running. You got a solution, like Nutanix that says, I'm going to restrict your options, by restricting your options to just my world, I'm going to make it simple to operate. But people don't want that restrictions, people still want access, particularly developers want access to a very rich set of tools that are available out there, that are only available in kind of more of an open world. And then of course you have ease of operations that the public hot guys have done. What we've tried to do is to take that same excellent base of infrastructure that, The HP's, and the Dell's, and the Lenovo's and whoever else provide, take that great foundation, then add software on to it that says, let's try and drive for the better software stack, like you would've got with OpenStack, let's try for software to find infrastructure like you would've gotten with Nutanix, let's try for your own automated operations, like you would've got from a public cloud, let's wrap machine learning around it to make sure we are continuously monitoring the behavior of this cloud such that it can more effectively what is required of it. >> So what is an engagement look like with a customer? Because obviously they got this infrastructure, they want to get more cloud like in the deployment of that, the accessibility really. Do they carve out a piece, is it a greenfield project? Is it some percentage of allocation of the infrastructure, how do they go about it? Because clearly, stuff's up, and it's running, there's still the IT piece of keeping the lights on. How do they carve it out, kind of what is there, I don't want to say go to market, but their internal project plan to start to bring this type of capability in-house? >> It can take a variety of us. The driver of it typically is Dev Ops, right? There's typically is a pain point that where IT isn't keeping up with it's application outbursts. That's usually the catalyst. >> And what's the screaming, bloody, I need help right now >> The screaming bloody I need help right now is, if I don't get my developers working any more quickly they are all going to Amazon. >> They just go, right. And they aren't allowed to do that, and I'm out of a job. >> I'm trying to stop the flame. >> I'm trying to stop this, stop that knee jerk reaction that says Amazon is the answer. But, I can't, because my current infrastructure is too hard and I can't keep up with it. So, that's simply the catalyst on how we bridge that gap. What we'll see, kind of probably two huge cases, to the examples that you gave, now one in maybe, in the context of a new application being deployed, I'm going to apply a new application. It is a cloud based application that needs a more flexible infrastructure. I don't want to put it on the stuff I have which doesn't work. Help me set up a new environment, that's a new use case. Similar, we also see, the I have a set of applications running and ready, the infrastructure is on, it works, but it's expensive, it's cumbersome, it's complicated, let me move some of those applications that is your stack as a a better place in which to live and operate and be managed. And we are operating both those models. In some cases, new infrastructure, in some cases using what our customers have. >> You've mentioned it a few times, the machine learning piece, a really important piece, not only the easy access and the easy interface with the infrastructure, but now you got a different level of intelligence around the use of that. So I wonder, are you seeing, do you guys flag them like, you not only have the cloud attributes of vis-a-vis but now you need a cloud attribute of big explosion, you better get some PO's in, with Michael and Meg. >> David: Don't call her Meg anymore. >> Don't call her Meg. Antonio, we love Antonio. >> I think that is the answer, right? So machine learning, that's a great use case for machine learning in a cloud, that says hey, given your current usage trends, this is when your resources are going to be consumed, let us help you get more. But machine learning is also helpful in how to get the most out of your infrastructure. Here are the resources the people have said they needed, versus what they're actually using. How do we better match for people actually using, to what's available and what's on demand. And over time you start to watch the behaviors in the system, these are the patterns we see advance. The whole idea here is that there's too many tasks that IT has had to do manually. And we want to be able to automate those tasks. We are not trying to eliminate jobs with automation, we are trying to eliminate tasks with automation. And machine learning is really the key that allows us to do that intelligently. >> It's funny this whole jobs discussion because on one hand all you hear about is the machines are taking all of our jobs, then you just go to the newspaper, whatever's your favorite, LinkedIn, and there's no shortage of jobs, right, there's plenty of IT jobs so, they aren't eliminating jobs, they're shifting jobs. They're looking for truck drivers still, even though we are going to wipe out all the truck drivers in a couple years. >> That's a different discussion. >> That's a whole different level of automation. But it is interesting, and it is about getting people to do higher order work and as you said, IT is no longer about keeping the lights on. It is the business, it's not support. >> It's about how do we grow with the business, how do we flex with the business, what's the right policy to support the business? That's not about configuring network addresses. Let the machines do that, let the cloud do that. Let's figure out what our strategy should be for connecting with our users and how IT's can handle that. >> So I'm curious, so you've had some deployment, you've got some early customers, kind of unexpected results, or second order impacts that you didn't necessarily expect or weren't that obvious that customers are starting to utilize by taking this approach to their hardware? >> David: Well, there's a couple of things. One is, part of what we do is we provide this idea of a workbench, we call the Dev Ops workbench, which takes a set of leading Dev Ops tools, you know, Jenkins, Ansible, Dooptroop, make your choice, and makes those available in one click, down to the users. What we've seen is people go very far, in terms of linking together those tools to fully automate their deployment. So being able to literally, software drive, software provisions the infrastructure, configures the application, deploys the application, spins it up, gets it going, I have service fires actually allowing users to go to a web portal, use a credit card, to order an application they want to use, which then creates the DM, installs the application, runs the application, and makes it available to users. So people just running with this idea of fully automated operations this way. I think the second thing is -- >> Literally IT guys are loving that, making them heroes. >> For a very large pharmaceutical company, an IT guy sat in the room with his Dev Ops peer, and said, hey, the more he can do without me, the better. And that's what we are trying to do. The second thing that's been, there's quite a few companies, that just said they're tired. You've got people that's been struggling with these cloud infrastructure requests for years, and at this point, they're like, you know what, I don't want to deal with it. We've had quite a bit of demand, and some of our big projects right now are actually in partnership with cloud service providers, with manage service providers who have been asked as a trusted advisor by their customers to come and say, build the next cloud for me. As an enterprise, I'm going to focus on the software, I'm going to focus on the applications, I'm going to focus on kind of managing my resources, you run it. And again, I think that opens up a new set of possibilities, in terms of how IT can evolve, and where they can focus going forward. >> That's a really interesting kind of under reported subset of probably new infrastructure providers, where it is, it's kind of a private cloud managed by a service provider. So I get the benefits of it, but I'm not having to run it. It's still undifferentiated heavy lifting, in terms of how my core business. >> Yes, and to add to that, all those things, and it's with someone you trust. Most of the time we see, is this is a long trusted relationship. If you're going to go to the cloud and you're not going to run it, you want to be able to look someone in the eye, and know they're taking care of your data, and they're securing your information, and they're taking care of your workloads, and that you can count on that partnership that you have. >> Well, I definitely think it supports the, it's a multi cloud world, right. >> It's a multi cloud world, yes. >> But the cloud benefits are still there. It's about being agile, it's about being fast, and like you said, it's about freeing up the Dev's to do dev, and not to do ops. >> Exactly. Let Ops do ops, and do it better, and faster, and easier than ever before. Let the developers focus on applications. >> Alright, David thanks for taking a few minutes to tell us all about ZeroStack, appreciate it. He's David, I'm Jeff, you're watching TheCUBE. We're in our Palo Alto studios for a Cube Conversation, we will see you next time. Thanks for watching. (light music)