In this episode we speak to 3 of the principles of Raft (Bhaarat Sharma, Barak Stout, and Paul Otto) talk about their approach to Govtech! They discuss how they work with government agencies (Such as Platform One and many others) to adopt open source projects like Kubernetes, Apache Kafka, Flux, and others. They also provide some additional detail around Raft's Diversity and Engineering Principles.
Today’s episode sponsor(s) are:
Lightstep - Lightstep’s observability platform is the easiest way for developers and SREs to understand changes in the health of their applications. By automatically analyzing application and infrastructure metrics and connecting dashboards to the underlying changes that matter, Lightstep helps you quickly mitigate outages and triage other issues. Download “The Ultimate Guide to Cardinality for Observability” at https://lightstep.com/popcast to learn more about how having the right data makes all the difference.
Redhat - https://www.redhat.com/ world’s leading provider of enterprise open source solutions, using a community-powered approach to deliver high-performing Linux, cloud, container, and Kubernetes technologies. We help you standardize across environments, develop cloud-native applications, and integrate, automate, secure, and manage complex environments with award-winning support, training, and consulting services.
In this episode we speak to 3 of the principles of Raft talk about their approach to Govtech! They discuss how they work with government agencies (Such as Platform One and many others) to adopt open source projects like Kubernetes, Apache Kafka, Flux, and others. They also provide some additional detail around Raft's Diversity and Engineering Principles.
Who is Raft
A new breed of digital consulting firm, that’s part full stack consultancy, part think tank and part band of creative folks. We’re problem solvers and innovators with a focus on Open Source. Our name is actually inspired from the , which revolutionized solving consensus problems by being more reliable and easier to understand. It is also the algorithm implemented by etcd (brain of the Kubernetes). We took on this name because we share a similar goal: replace any overly complex and outdated system with a new, efficient, and secure one that just works.
00:00 - Sponsor - Lightstep
00:47 - Sponsor - Redhat
00:53 - Popcast Opening
01:03 - Intro to Raft Team; Barak, Paul, and Bhaaraat
01:30 - Barak, Paul, and Bhaaraat Journey
05:40 - Raft Protocol? What is with the Raft name?
09:07 - Open Source Adoption in the government
12:26 - Why Raft over other providers?
19:19 - Raft's Diversity as a Super Power
22:55 - Engineering Principles
26:08 - Contributing to Open source projects
27:59 - How to get in touch with goraft.tech
29:25 - Work you are most proud of
Episode Links:
POPCAST SHOW DETAILS
Watch (YouTube): https://www.youtube.com/c/thepopcastpop
Listen (Apple PODCAST and others): https://popcast-d9f7b6dc.simplecast.com
Follow us on (Twitter): https://twitter.com/PopcastPop
Follow us on (Linkedin): https://www.linkedin.com/company/the-popcast-with-danpop
Dan Papandrea (00:00):
This month's sponsor's Lightstep. Lightstep's observability platform is the easiest way for developers and SREs to understand changes in the health of their applications. By automatically analyzing application and infrastructure metrics and connecting dashboards to the underlying changes that matter. Lightstep helps you quickly mitigate outages and triage other issues. Download The Ultimate Guide to Cardinality for Observability at https://Lightstep.com/popcast. That's L-I-G-H-T-S-T-E-P dot com slash P-O-P-C-A-S-T. To learn more about how having the right data makes all the difference, Lightstep. Our other sponsor for this month who's sponsoring captions and transcripts are Red Hat. Thank you very much, Red Hat.
Dan Papandrea (01:01):
Hello everyone. Welcome to the Popcast. I got some friends here. They're sponsors of the show. They're fantastic people. I have Barak Stout, Paul Otto, and Bhaarat Sharma. They're from Raft. Hello guys.
Paul Otto (01:13):
Hi.
Bhaarat Sharma (01:15):
Hi, Dan. Happy to be here.
Barak Stout (01:17):
Hey.
Dan Papandrea (01:18):
Let's talk. There's three of you. This is a first for the Popcast. I never have more than two people, all right? But you are a force. This is like a triad of awesome. So let's talk about... I want to talk about your history. Each of you talk about how you got started in tech and then we'll talk about just starting Raft and what you all do there. So Paul, do you want to go first maybe?
Paul Otto (01:42):
Sure. Why not? So, yeah tech is definitely in my DNA. I think that definitely goes around the table here. I've been doing tech since before I could write English. I learned on a Trash80 on my dad's lap when he was learning how to do machine code. And then my first computer was a Commodore 64, and I've just been in and out of tech ever since. Had a stint in the military, but came back-
Dan Papandrea (02:20):
Hold on, Paul. Everybody who's listening and watching this right now, please have a shot. He said, "Commodore 64." And those of you who don't drink, don't have a shot, okay?
Paul Otto (02:29):
Everybody has their start in either the Commodore 64 or what, the Apple 2, I think it is, right? Yeah. In more recent times, I definitely got really plugged into orchestration and cluster computing. Got in through Apache Mesos, and then as things started to shift towards Kubernetes, I really got drawn into Kubernetes. I've been in the Kubernetes space since I think it was 2014 or so and have been doing it ever since. So I'm just excited to always be able to be working close to or around Kubernetes and big data using orchestration.
Dan Papandrea (03:22):
Nice. All right, Barak, how about you? How'd you get your start?
Barak Stout (03:26):
Oh man. I think I started programming really and getting into tech around high school. I got a first computer. I got into high school. I don't think that thing was ever screwed on right. It was always taken apart with pieces sticking out of it. My uncle would come over and have arguments with me, "That's not safe," and I can start a fire or some shit. It didn't work well. But yeah, I mean, it's been in the DNA. Tried to avoid it. Went to school. After my first class in comp science school, I was like, "Yeah. No, I need to be in computer science." Sort of continued that path. I was a teacher for a bit, so I did more of the academic side. Got started with Raft and got introduced to Kubernetes, and I just fell in love with what it can do. That's pretty much it in a nutshell.
Dan Papandrea (04:29):
Bhaarat, how about you? How'd you get started?
Bhaarat Sharma (04:32):
Yeah. Sort of the same story like Paul said, goes around the table. I started back in high school as well. I just love the fact that you can do something on the computer and make something out of it. It's a very short feedback loop and that's what resonates with me really well. You write something, it gets a response back, you can iterate on it and go faster and make something happen. So I think from there just doing that in high school, and then, hey, like I said, if you can do something for work that you love, then that's the best of both worlds. From there, it's really once I got into quote unquote being paid for development, it got more of what can you use the tools to make a difference in people's lives. And that's really what I've sort of tried to build my career on, and that's what Raft is built upon as well. Tools, Kubernetes, whether it'd be Pivotal Cloud Foundry, programming, any language, those are all foundations to make something happen and make it click for the end user in the end, right? That's what clicks for me as well.
Dan Papandrea (05:40):
Got it. So everybody who's watching this is probably wondering the same thing as me, right? The name Raft. It's like, "Oh, is it the Raft protocol?" Are you guys working on the Raft protocol? What exactly is Raft? One of you take it. Whoever wants to take this one.
Bhaarat Sharma (05:55):
Right. I'll tee it off. So the name, yes, it definitely comes from etcd. We're into the Kubernetes space. We are into distributed computing. So, when we were searching for a name, we were like, "Why not? Why not Raft?" It floats multiple avenues. And then of course, domain name is a problem, but that's why the Go Raft is there. So the whole thing sort of falls into place that you have the Raft distributed computing algorithm, and then you add Go to it and you can go to Raft and be on a journey with us.
Bhaarat Sharma (06:27):
The journey that we're really on is to be the non-traditional disruptors in the GovTech space. So in the GovTech space, there's really two buckets of where companies fall. Either you build a product, enterprise product, that you're selling to the federal government, or what is a dirty word in the government that I at least feel is "system integrator" that you're taking something and just integrating it. I feel like there's a third niche that hasn't been carved out yet is open source accelerators. So there's so much work that's happening in the open source community. Why shouldn't we leverage that in the federal sector? And we've found some great partners along the way. There are some agencies out there who are forward-leaning, and if a government agency is building a product, why shouldn't that product be open source, right? I mean, it's taxpayers' money at the end of the day. The source code should be openly vocal.
Bhaarat Sharma (07:28):
So we are working with a couple agencies, I can talk more about them, and all the source code that we build for them is open source. And while, at the same time, being able to contribute back to open source as well. So we have some ideas we're floating around where you get X amount of hours to work on open source and contribute back as well so that we're not just taking open source, but we're actively contributing back. So we wanted to be the company that makes open source happen for the government. There's multiple ways how we make that happen, but that's sort of a nutshell of what Go Raft is.
Dan Papandrea (08:09):
Again, government is usually not... It's not adopting as quickly and we had Nick Chaillan on as well and we talked about that doesn't make [inaudible 00:08:20]. But there's been always outpourings. And I did an interview at some... Last year in terms of like government kind of symposium, the Fed government symposium. We were talking about like the FDA was one that basically, they were the first ones to kind of have this, kind of adopting a lot of the open source paradigms and using them within agencies. I've seen that more and more. If you're going to... Again, what if you think about what we do at Sysdig with Falco is this is like internal, external actors. If you have the ability to create rules from an open source perspective, and somebody has already done that, and you can address those things quickly from an open source perspective, that's amazing.
Dan Papandrea (08:55):
Or doing something like Kafka and Kubernetes, you can easily create topics and things at a much more quicker and getting adoption through the government. So that's amazing. Those boundaries that happen from a government perspective, how do you all traverse those as a service that Raft gives?
Bhaarat Sharma (09:17):
Yeah. That's an interesting question. I mean, one of the basic things that in the government... Sure, government might not be the first to adopt things because quality matters a lot. And some of the things that we build for the public consumption has a lot of security aspects associated with it as well. And security not being... I mean, I think open source community is adopting security now going forward, but it hasn't been the first thing they think of, right? When we bring in these open source libraries, whether it be Kafka, whether it be any products, that's sort of our first focus, "Hey, does it have CVEs? Can we fixed those CVEs? Can we bring it into a federally FedRAMP way?" And then have it complied. And you mentioned Nick Chaillan. So Iron Bank is a great example there, right? Can we build a repository where a bunch of these open source repositories live, open source products live, that already comply with everything that the federal government is trying to do? And then everybody can leverage it from there. So sort of a well that everybody can come to to pick stuff up and build upon it.
Bhaarat Sharma (10:27):
Some of the other things that... I mean, we've been fortunate to work with some of the clients. One of our clients is also CFPB, Consumer Financial Protection Bureau. So for CFPB, we built a platform called the HMDA platform. So any government agency, any sort of... Not government agency, but any mortgage lender that provides mortgage across US has to submit their mortgage data to CFPB on a yearly quarterly basis. We built a platform back in '17 that went all in on Kubernetes, right? We've been all in on Kubernetes, Kafka, Keycloak and all these open source libraries that weren't fully developed, but we knew they were on a path to get there and we were able to go deep in and start adopting them because the needs of the platform needed them.
Bhaarat Sharma (11:20):
The HMDA platform, there's a mandate that all the data we collect, about 20 to 25 million records every year, comes in between Jan 1 and March 1. Right? So 60 day period, and then everybody hits submit on the last two days. So we see normal growth and then hockey stick growth all the way in the end. So we knew what we had to sort of architect, and that's why we went all in with these technologies that we knew would fruition into the labor.
Dan Papandrea (11:51):
And you think about this amount of scale. Right? Why wouldn't you use [inaudible 00:11:56] cloud technology? Why wouldn't you do distributed computing? The taxpayers like... It used to be you have to build this Cadillac and you had to cross your fingers and hope you had all these records versus you have distributed computing and those types of things. And you can expand as you grow and grow and grow as these agencies grow and grow and grow. And if you built a platform that somebody can basically just like cookie cutter this, I mean, why wouldn't you want to work with Raft? Right? That's phenomenal.
Dan Papandrea (12:25):
And one of the things too... So again, I'm going to ask this question. The question is, anybody listening to this, why raft? Why would I use... There's so many service providers out there. There's so many companies out there that provide this to the governments through the [inaudible 00:12:40] platforms. Why would I use you?
Speaker 6 (12:43):
Paul, Barak, y'all want to tee that off?
Paul Otto (12:46):
Yeah. I'll let Barak start this time. And then I'll follow.
Barak Stout (12:53):
I think at the heart of it, we like to solve hard problems. We like to take hard problems that have a human that has a need and figure out a solution that fits. That does take from the open source. We don't want to reinvent the wheel necessarily, but there is some tailoring that's needed, right? You can't just take Kafka and run it and everything is fine. Every solution needs to be customized to what that product needs. But it's really about solving a problem that's presented to us and going through the motions and following the human centered design, following discovery, finding out what the MVP should look like and acknowledging that it's a stepping stone to something bigger. It's foundation layer [crosstalk 00:13:39].
Dan Papandrea (13:39):
Can we double click on that? Let me double click on that. Let me ask you this. I'm a government agency. I could take Kubernetes and spin it off, do Kub admin, and whatever. Or I can take Kafka and take it off the box. Why would I use Raft?
Barak Stout (13:54):
Well, we're one of the first, if not the first that ran Kubernetes. So we have the years of experience. We've been through a few filing seasons with this hockey stick growth. We've been through a few cycles of Kubernetes and updating and patching and trying to stay current with all their versioning. I mean, that by itself is some learned lessons that other people don't have to go through.
Dan Papandrea (14:21):
It's two things. Again, two impressive things. Kubernetes in itself is very difficult to operationalize, but once you have it up and running, everybody knows the advantages, but we don't have to sell Kubernetes anymore. Kafka, amazing technology for being able to have all this distributed data in all the things you can do with this. So you all kind of have this methodology for making it in a place for a government to be able to adopt it quicker. Would you say that's the value of Raft?
Paul Otto (14:51):
Oh yeah. Yeah, definitely. I think that, going off of where Barak started there, it's like, all right, there's already an industry. There's best practices. You can read all sorts of different variations of it, what some people say. But it's all following infrastructure's code, and things are starting to move more and more towards like getopts and whatnot. But it's one thing for places to say that they're doing it and it's a whole nother thing when you go and sit back and look behind the curtain and see what's really going on and how many bash scripts there might be and how many tickets are being filed to get things deployed.
Paul Otto (15:43):
So bringing those best practices and the things that each of us have done in industry. Because I came from working at commercial outfits at all my previous jobs. We're used to that in a lot of these commercial operations. Being able to take that and adopt that towards putting that into place in these government organizations where they're kind of used to red tape in some cases. So it's easy to kind of fall into that old IT mindset. So us being able to come in and really push that envelope and move the needle forward, but then also really being interested in hard problems. Because there's only so many times you can really sanely just build and deploy web apps. I mean, everybody's done web apps. Now, I mean, you can just run a couple of commands and you can have a whole Kubernetes cluster and Istio come up and have a web app running. Wonderful, awesome, cool. I've got a make-believe shop that goes down and little versions go up.
Paul Otto (17:08):
But what happens when you have to have something that's completely disconnected from the internet, but I need to be able to have a virtual registry still available and be able to push hardened images that are cryptographically signed? And it all still has to work as if it's a regular cloud native application. I mean, those are difficult problems that many people just might not even be interested in trying to approach and we're like, "Bring it to us."
Bhaarat Sharma (17:41):
Yeah.
Dan Papandrea (17:43):
That's phenomenal. I'm sorry, Bhaarat, were you going to say something?
Bhaarat Sharma (17:47):
I'll add to that. I think Paul And Barak put it well, but I'll add to that, and why Raft? Because if you are an agency that, sure, you can get Kubernetes and Kafka running quite easily, but if you want to get to production faster with those technologies, then you've got to go with somebody like Raft where we've taken those lessons learned and we can bring them to practice. Then second, we can build PowerPoint presentations better than any other company out there, but that's not our thing. We don't put pieces together and be like, "All right, this is what we're going to build. Now we're going to go build it." If you want to get shit done, you go with Raft. So if you want to turn that PowerPoint presentation into actual MVP that you can touch and feel, you go with us.
Dan Papandrea (18:37):
I've interacted with a lot of agencies out there. So I've seen a lot of it in my space. What I love about you guys is, again, a lot of folks will put Kubernetes on their websites and they'll put Kafka on their websites and you actually live and breathe it. That's what's great about you all. And folks, look, I don't have people on my show that I don't believe in. And I absolutely believe in Raft. I was talking to Bhaarat before. I'm a fan of you all, so it's definitely good stuff. And I'm really excited to have you on the show.
Dan Papandrea (19:08):
So one of the things too, that again, government agencies are all supposed to be this stoic... The things that you go into. But I love one of the things that you have on your website. It says, "Our people and culture set us apart." I want to elaborate more that you talk about diversity in tech in your engineering principles. Let's talk a little bit more about that quote and diversity in tech and engineering principles.
Bhaarat Sharma (19:32):
Yeah. I can tee that off. So diversity is something that is very near and dear to our heart. So I mean, our CEO is Shubhi Mishra. She's a woman. She says it the best. She gets referred to as Mr. Mishra all the time, right? So if somebody emails the CEO of Raft, they inherently assume that it must be a guy and they don't know her, but they get Mr. Mishra associated with it. So I think that just shows you how much of a cover there is on diversity within the federal government. Nobody expects people to be out there and doing stuff on their own. And that's really what we're trying to build there. From culture perspective, Barak and I have been working for about three years. There's people within Raft that I've known since, I don't know, 15 years. We've tried to make a pact that is out here trying to do stuff that is not traditional, not same. We are trying to make something that is totally different and be in a space where we're not doing things in a traditional way.
Bhaarat Sharma (20:49):
And that goes hand in hand with our culture, right? In our culture, we've been remote first, even before COVID. So in the federal government, that's hard to do, but we've taken that conscious choice that, "Hey, we don't want the work that's going to have us sit in [inaudible 00:21:04] or sit in a desk somewhere and do nine to five. We don't want to do that work." That's a conscious choice that we've taken as our culture. As well as there's no time-tracking within Raft that we do. It's really, what is the output? Nobody cares whether you're working 80 hours a week or you're working eight hours a week. What is the output? That's really what matters.
Bhaarat Sharma (21:30):
So I think those two things really set us apart as far as culture and diversity go. In terms of diversity, I think we... We're going to have a blog post on this. It's in the pipeline, but we're contributing to [inaudible 00:21:46], we're contributing to other initiatives that are from not fully represented communities, and we want to bring those into the government. We can't disrupt what's being done in the federal government if we keep bringing on people who are from other companies who are already in the federal space. We want to bring people on who haven't been in this federal space. I love this quote from... I forget who was the quote from, but "Perspective is worth 80 IQ points." Right? So we want people coming in looking at things from different perspectives. And that's how we are going to get smarter. If we keep looking at the same thing from same perspective, we're going to come up with the same solutions.
Dan Papandrea (22:30):
Fantastic. Again, it's what sets you all apart from the other people that put Kubernetes and Kafka on their thing and put diversity, but you live and breathe it. I saw multiple blog posts about it from... It wasn't just the marketing people. It was literally like engineers within the organization, which was really, really cool and I like to see that. Talk to me, and maybe Barack or Paul, talk to me a little about engineering principles. So if I was maybe a developer or somebody who was, "Hey, I want to get in this space," or whatever. If I wanted to work for Raft or something like that, talk to me about the engineering principles. Hello?
Barak Stout (23:15):
Paul, you want to start this? Or should I take it?
Paul Otto (23:16):
Yeah. Sure. I'll try, and then you just cut me off. Yeah. So with the engineering principles, first and foremost, it's about moving fast at the very beginning, like Bhaarat had talked about. We're really focused on being able to deliver value early, rather than showing amazing glitzy looking slide shows or just some UX kind of drawings. And that experimentation is king. You don't learn if you're not making mistakes early. We don't want to make the mistakes when we're delivering. We want to make the mistakes early and break stuff before it matters. So that way, we already know. We already have the runbooks. we have the documentation. We have hopefully some self healing scripts or what have you so that's not a problem later on.
Paul Otto (24:33):
I mean, there's definitely that. I know we do a lot of... It's a lot of collaboration. If somebody's getting stuck on something, we're working in open source, so we're already used to the community. You reach out to the community to get some synergy and to really discuss things that other people have already thought about and hopefully solved. We do the same thing internally as well. You get on, in our case matter most, and you bring up, "Hey, I've got an issue. Hey, is somebody available to do some screen-sharing? Let's figure this out." And everybody gets raised up at the same so that the next time there's an issue, if I'm sick, somebody else can pick it up, hopefully. I mean, those are just a couple of things that come to mind for me. So I'll let you take over, Barak.
Barak Stout (25:33):
I don't know if you missed anything. I think you hit a lot of them on the spot. A lot of striving for learning and sharing, stand on the shoulders of giants. If I've learned the lesson, there's no reason for somebody else at Raft to come over and learn the same lesson the hard way. But definitely a lot of verification, checking each other's code, doing screen share, making sure we're sane. Sometimes we're not. Yeah, those are some of our key principles. Doing as much as we can in the open, whether it's open to the team, but hopefully open source.
Dan Papandrea (26:08):
Are you contributing to projects like Kubernetes and Kafka as well? I'm just curious, is there any open source projects you all are contributing to?
Bhaarat Sharma (26:18):
We are in terms of Kubernetes. We've gotten people who are actively trying to join calls and trying to get in. I mean, Kubernetes is [inaudible 00:26:30], so maybe that's not a great example if you're trying to get in there. But in Kafka, we've looked at getting FIPS 140-2 encryption getting done inside Kafka. We haven't taken that on yet. It's a big thing to take on, but we found an itch that we need scratched and we can go in there and do it, but it's just about time.
Dan Papandrea (26:51):
Got it.
Paul Otto (26:51):
Yeah. I know I'm working on some ideas around improvement on Flux v2. Flux is a really promising project. At the Air Force side, the Platform One group has moved towards that instead of using just Argo CD primarily. I've done a lot of work around the Flux, getopts stuff, and there's definitely some paper cuts that could use some addressing. And there's only so much that Weaveworks can really commit to just solely that project. So I know there's enough times where I've complained about certain things and I've thought I definitely need to get my hands dirty. So I'm working on some things around the UX for deployment.
Dan Papandrea (27:49):
Nice. Alrighty. And I'll put this in the line notes for the episode, but for folks that are listening, how could they get in touch with you to use your services?
Bhaarat Sharma (28:02):
Email. Email is the best way. Send us an email. We reply fast. That's one of our qualities as well. We are in the community as well. From time to time, if you're attending an event, we'll open our Mattermost channel to the public. So come in and join. And if you just want to learn more about Raft, that might be just the best way for you to come in to learn more about Raft.
Dan Papandrea (28:31):
Do you have any events up and coming, Bhaarat? Any events up and coming that maybe we can... Again, this is probably going to air in like the middle of February, end of February, March timeframe. Maybe in the first quarter sometime.
Bhaarat Sharma (28:44):
There is an event, but it's not a Raft event. But yeah, we don't have any that come to mind. We were in the Cube Con. For Cube Con, we had opened up a room, but for now, I think right now we are focused on some of the initiatives that we're working on the projects for. But yeah, for now, we don't have anything yet.
Dan Papandrea (29:07):
So everybody, if you want to get in touch with them, I will have it in the line notes of the episode. Here's the contact us page. Again, really useful, really amazing site. They talk about their principles and again, if you've haven't gathered from this episode, I'm really impressed with them. So definitely check them out. Last question for you all. Each of you are going to have to answer this one. What work are you all most proud of? And maybe we'll go down the horn again. We'll go Paul, Barak, and then Bhaarat.
Paul Otto (29:33):
Oh yeah, of course. I have to be first. Let's see. I think the work that I'm most proud of most recently would be the work that's around getting... Let's see. It's with Kubernetes and Kafka working with Edge over at Platform One Air Force. It's a challenging project because we have all these different moving parts. We have things that are running in more of a traditional gov cloud type of environment. So it actually has all the services that you'd expect available to it, but then that has to be able to reach out to other environments that are a lot more secure. So some of the handshaking that has to happen has been pretty tricky.
Paul Otto (30:48):
Then on the whole other side, there's these edges and it's different kinds of edge. I think Barak would probably like to speak more to that, but they really go off this notion of Edge being something that is not on the internet, and it could be something really big in a big truck where it's like a mini cluster, or it could be some really small device. And being able to run in somewhat unpredictable kinds of environments so that you can have a cluster come up, be able to phone home, sync some data and then be able to also be a hub for activity going on in an unknown area, capture information that's going on between devices, and then later on be able to transmit it back or what have you. It's challenging because you don't know ahead of time what the different kinds of specifications are going to be like for those kinds of environments.
Dan Papandrea (32:00):
Barak?
Barak Stout (32:02):
Oh boy. I've worked on a lot of things that I'm proud of. Yeah. I would say hardware in the loop. Because it's an active problem that has been haunting me at night and every weekend and every time that I can get a few days away, I kind of go away to it. It's idea of bringing some of our Kubernetes principles, some of this infrastructure as code and starting to include hardware as part of it. So you want to continuously deliver to like an Arduino board and you want to update on the fly, but you also want to run some tests on it. So you want to have somewhere in the loop that you can either emulate it as a digital twin. And Azure has been doing some work with that. There's a bunch of different services that are starting to do some work with it.
Barak Stout (32:54):
But at the scale of the Air Force, it's much bigger. If you think about the amount of devices that they have and the amount of digital twins that you need, or you might also want like a physical lab, like somewhere that I can deploy code as a developer, see the code run on actual hardware before it's even close to the field. Right? The last thing you want to do is deliver some code to a drone mid-flight that's buggy. That can have some unforeseen circumstances. But working on that as a big problem has been something that I've started and I'm proud of, and we're continuing to actively work on that. And it connects to some of the other work that we've been doing with submitting data from Edge, from these small, tiny devices and bringing them into a cluster that can be used by other application and other services in real time.
Dan Papandrea (33:43):
Bhaarat?
Bhaarat Sharma (33:46):
Yeah. So for me, it would be the project I already mentioned. Home Mortgage Disclosure Act for CFPB. I mean, in that project, we are using Kubernetes. We are using Kafka. We are using Cassandra. We're using the cutting edge, bleeding edge technology, but what's most cool about that project is how we're using that technology to make an impact in a bigger mission, right? HMDA is all about collecting data from these mortgage institutions to see whether they're serving the needs of their communities. Right? So once the data comes in, the policymakers can assess the data to see whether the lenders are doing any discriminatory lending practices, whether it be race, sex, gender, right? Then what are the patterns within the mortgage industry? This data wasn't there before the Dodd-Frank Act, when CFPB got started. This data was collected over literally faxes and emails.
Bhaarat Sharma (34:53):
And now, this data is available for collecting first and then in real time making it available for the public so that if you want to buy a house, hey, you can see the data that's available for public use, and you can see whether you're buying the house in the right area. What is the mortgage industry like, what rates might look like, and which lender might be best for you? So I think making an impact from that perspective has been really good from my perspective. And combining that with technology, it's really a project where technology, policy and data are really combined. And it's a project at the intersection of all those three.
Dan Papandrea (35:35):
Well, gentlemen, we did it all. All right, we got Raft. If you're an agency or anybody that's interested in exactly what they do in terms of Kubernetes and Kafka, this is it. This is the apex predator. These are the people right here. Right? I'm a little upset with them that I didn't get a hoodie and I'm not part of the cult, but it's okay. Hopefully I'll earn a hoodie. But thank you all for being on the show. I really appreciate it. Thank you all for just being awesome.
Bhaarat Sharma (36:03):
Thanks, Dan. You'll get a hoodie soon enough.
Paul Otto (36:06):
Hey, thank you. Yeah. Ask Barak for one of those cool lab hoodies because I don't even have one of those.
Dan Papandrea (36:15):
Thanks, everyone.
Barak Stout (36:24):
They were hard to find. Thank you.