Developer Data Platform – Andrew Davidson, MongoDB

This is texturing TV. Hey everyone, welcome back to techstruck TV. I'm happy to welcome back Andrew Davidson.

He's SVP of products at mongodb. First of all, Andrew. Hey, man, welcome back to techstrong.

It's good to see you great to see you two out the wonderful to be back. And you know, I was just thinking how the last time I was on. I think we were Peak pandemic and now back in the office.

Look how far we've come it's yes we have I think I was just said the last time I saw you you hold up in the house. That's right, you know doing these things, but I remember how excited you were. It was the announcement of Atlas right the outlet service and Jesus it's got to be a year and a half ago.

Now at least maybe that's what yeah. Yeah. I think that was the five-year anniversary.

We're easily six and a half years in now. Yeah, so that's exactly what it was a year and a half and And look, yeah exactly. Look how far we've come you're back in the office.

We would just talking about being out in Vegas for AWS reinvent and older people who got sick out there. And so I guess we haven't come as fast. Yeah, but Andrew You mentioned already Atlas is about six and a half years old it but it's one of the kind of crown jewels in the Mongo DB.

Universe why don't you know for people out here in the audience or not familiar? Why don't you give them a Little recap primer on Atlas totally. Yeah.

So let's go up a level real quick, you know not everyone's familiar with Mod CB or necessarily where we are today and Atlas, of course, I'm gonna be Atlas is really the core part of our company strategy today. But if we take a big step back, I think it all starts with this realization that modern businesses in the digital economy. They live and die on how productively they can build and evolve with software the ability to respond to threats the ability to disrupt Industries the ability to you know, not be disrupted really fundamentally lives and dies on that ability to empower software developers essentially The Artisans of our time to build that software and if you you look at where developers spend most of their time traditionally, it's basically wrestling with state or wrestling with data and so mongodb's core thesis from the very beginning has always been that making developers more productive by addressing this core pain of how hard it is to wrestle with data really will unlock agility for operating as a software company, which is increasingly critical for every type of company.

Everyone's operating as a software company today. So we you know our core strategy and I'm gonna be Atlas is centered on this idea that we conceptualize ourselves as a developer data platform and we think a developer data platform is made up of three key attributes. The first is that it must be standard on a modern data model.

It must be it must be the model developers think in code with it. It's a document Model A. Json style model documents are special because they're the super set of data models.

But the second key attribute of Adele developerative platform is that it should be possible to drive the vast majority of types of workloads behind that developer interface. You should be able to operational transaction on Surge and after analytics time series multi-region and distributed Global applications device-centric applications serverless event driven applications, all of these different form factors for applications. You should be able to do all of them behind that same elegant interface.

Why does that matter it allows for standardization? You can build a vast majority of your modern applications with the same developer data platform and the third and final key attribute is that this should all be accessible on top of a secure elastic truly multi-cloud Foundation available in over a hundred regions around the world. That's what I'm going to be Atlas enables today.

And that's why we have over 39,000 customers building amazing things ranging from disruptive startup to some of the most amazing Enterprise in the world companies like Vodafone Volvo and two thirds of the company's mentioned in the AWS reinvent keynote just to show how intimately tied in we've always been to the AWS fabric. Absolutely. and you know Well, it's not new star audience.

There were 55,000 people out. There it reinvent this year and it reaffirmed. really just the breath and width of that of that community and and what a force it is, right?

I mean we were talking off camera. And you and and you know, look there's a lot of did a lot going on in the database world, right? We've certainly moved Beyond sequel though.

There's still a place for the post-gresses and all that right here, but today's world. functions at such speed and scale and I think one of the big stories when I look back at 2022 is the Primacy of data, right? We've spent so long.

I think we lived in an appcentric world. The app was the King right? We're developing apps with as an app for that.

There's an app for this but what we forgot it's the data that drives all this stuff, right? That's why we have the apps because the data and collecting that data and using that data and we and the way we want to connect data and you know display data and analyze data. It's I don't want to say it's infinite, but it's it's infinite, right?

I mean, that's just You know, there's a lot going on. I guess my question to you is okay. So now Atlas is six and a half years old.

You got AWS. So Google and Microsoft you have a whole bunch of different. There's the Oracle Cloud as IBM.

There's you know, there's all different stuff out there and you don't have to keep it in a public Cloud either for that matter. Sure. Seth you know as a service offerings are dominating Is atlas stay on top of the pack?

It's a great question. Well, and I'm gonna I'm gonna answer that by going back to something you were mentioning a moment ago, which is the sort of Primacy of data and what's interesting about that phenomenon is and because we have a technical, you know viewer base here. I think people will intimately understand this but I think the average person on the street when you ask them what does data mean to them?

They typically jump to a very specific subset of the data ecosystem. They think about what I think we and in sort of the data world might call analytics because they jump to this idea of hey, I need to derive insights from data almost like the glorified spreadsheet. We all know to spreadsheet is regular folks know people on the street know what a spreadsheet is fine.

And so this idea of driving insights from data is pretty intuitive to people what it turns out is counterintuitive to regular folks. Is that operational or what we sometimes called transactional data, which is really the lifeblood almost the lubricant of the digital economy. That's what's so hidden away and yet it is where the center of gravity for spend and criticality in the data ecosystem isn't how is it that this huge phenomenon and data something that's so hidden away that we don't understand it or that the average people don't know it exists.

I think it's because you hit the nail in the head. It's actually it's it's hidden by these wonderful applications that abstracted away right? I order when I order my you know lift.

I'm not thinking about the database by the scene. I'm thinking about getting a car to go to where I'm trying to go or similar. That's the same way with all the software we use you don't think about the state behind the scenes but that operational transactional state that makes it possible to do all this.

It is the critical part of the economy that sort of doing all this heavy lifting all the time and it is it's just it makes it all possible. And so You know going to your question on how does Atlas stay ahead of the pack? I mean It's fundamentally because Atlas first and foremost that is our core.

It is powering operational transactional workloads with that expansion that I mentioned into capabilities beyond that coming from the fact that we're the date is born and originally put to use that's also where you can make it searchable and make it, you know possible to do those after an analytics on top of it. But having the data where it's bored and put immediately to use being operational transactional at one's core put you in a position of being able to really talk to your customers in our case developers Builders software Architects Etc about what it is that they need to build the software of tomorrow and we're just so obsessed with that and constantly talking to our customers about what is it going to take? You know, what does it really gonna look like to build, you know to build that next great idea from something small finding one's own product Market fit with one's own software with one's own application, whether within consumers or for internal business processes validating and growing from there and evolving and it turns out this key part of it is this ability to Evolved traditional ways of building software where you kind of figured it all out up front and then he would just build it and you know kind of more waterfall approach.

We all know that we've moved away from that now two Pizza teams need to be autonomous and building incredible pieces of software from day one that are can always being invested in there's a roadmap. It's based on their own product Market fit and their own customers needs. The ability to modify software is so important and that's why you need a flexible scalable developer data platform that sits behind it.

It allows you to keep up with the needs of the application as its needs evolved. You know the alternative Paradigm and what we've seen folks do over the years is to start with maybe more of a legacy model starting with like let's say a relational database and then for each feature in their application that kind of burst beyond what that could service whether it required flexibility or scalability or any number of other important criteria, you'd bolt on these other engines. Then you kind of find yourself in the spaghetti code challenge of lots of ETL system.

Lots of bolt on engines search engine over here caching layer over there to bring it all together and make it cohesive. If you look at that alternative Paradigm, what we've come to realize is basically the customer is sprawling the developer sprawling across lots of different apis lots of different management paradigms, which leads to all kinds of governance challenges and all this brittle ETL to breaks and it's costly and we've realized we can just flip this whole concept on its head and go back to that vision of single interface for the vast majority of your application stateful needs the developer data platform vision. That really has helped inform our strategy and it allows us to keep executing an atlas.

If you if you were to ask me what's our top priority right now. I would probably say building the awareness and and going bigger and bigger in terms of use for Atlas search because searches everywhere and not all of us recognize or even imagine that it's possible then our main data service that powers our applications. We would just get search out of the box for free but that's what's there in Atlas And so this this realization that we could really execute in a way that truly transform the idea of how to build software.

I think that's what's thus far allowed us to execute now, we can't rest on our Laurels. That's for sure. It's dependent Market.

Oh, absolutely not. I absolutely I I get that for sure. You know.

sitting here listening To Angie do you explain they say I think back? 2008 2009 no sequel kind of be burst on the screen. Maybe it was 2007 and a no sequel burst on the on the scene, right?

I interviewed I forget who it was. It was like the first CEO over there at mango. And you know, what was the reason?

Well, it allowed us to just store data and access data and make connections with data without being tied into the the schema. If you will of the whole sequel database thing and take tables and all of that crap. And here we are.

Toss on top of it today. It's it's still at the core. That's what it's about.

I yeah, no, it's a great point, you know. we some of my colleagues would say, you know, it's wonderful to conceptualize ourselves as developer data platform as if that's a new phenomenon for us and I would almost say We've always in my view been a developerative platform if you think about that moment. You were just mentioning 2009.

Let's say 2010, you know AWS had just been born mongodb had just been born. And another key Trend the the iPhone had just been born and what this kind of this combination of factors, you know, the iPhone was leading to this proliferation of new levels of scalable application requirements, including geospatial Primitives, like never before uws had enabled commodity Hardware as a service which you know, everyone at that moment. I'm like cool I get to operate like Amazon or like Google this is awesome.

Right, but if you think about it, Having a bunch of commodity Hardware is available at your fingertips is super empowering. However, if you're a developer, what you need is something that enables you to turn a distributed system a cluster of that Hardware into something relevant to you and mongodb kind of rode that way but the perfect time wasn't just that it was this incredible document data model at its core with flexible scheme and all that. It was also that it was a distributed system that could make relevant these early commodity Hardware building blocks AWS democratizing in a way.

That's why I think of mongodb in AWS is two peas in a pod. developers needed something that would translate this document interface into A bunch of compute on the back end at AWS is making it available and you know, we we rode that wave. They rode it with us the demand of the applications in that moment that the iPhone was driving pulled us into the direction of geospatial Primitives.

Then over time we were able to keep reinvesting multi-document asset transactions Mission criticality, you know, and then basically eventually we realized we could be the best database in the world delivered as pure software, but guess what? People will end up using other databases that aren't as good that are managed as is fully automated service, right? If that became a cool.

That's the way the world went. Yeah, so we went all and we realized we didn't disrupt ourselves. We can't be worried about this.

We have to go all in and go up a level of abstraction. We were pulled by our customers in that direction and go all in for fully automated service. I'm going to be Atlas and of course that opens up.

So many changes in how you can conceptualize you no longer managing low-level Plumbing now, it's all declarative. And if you think about it that declarative model for the infrastructure that Atlas brings it's very analogous to the declarative model that we brought for the data plan the document model and so marrying these two enabling it two beats of team to move fast. That's not just great for the two beats of tea the game changer for large Enterprises.

If you've got dozens or hundreds of those two beats of tapes for them to all be able to you know, build their active assembly line on top and whip it all up automated infrastructures code with great devops culture and to do it in a standard developer data platform that bill you know, that they can build on top of for the vast majority of their applications everybody wins, and they're not locked into a particular hyperscale. A great, you know, I am listening to you. So I think another story around Dev Summit.

I'm on year-end stuff, right? We've got our predict show coming up in early January. About what's big in 2023.

And so I think another story around devops for 2022 is I I think for the most part the world got the two Pizza team concept down pat, right? The problem is is when you go to large Enterprises. There's a lot of two Pizza teams in the average company right you It's Not Unusual to have 50 a hundred hundreds of two Pizza teams.

and I think some of the like Mega success stories in in Cloud native and devops and you know at SAS kind of stuff is the ability to recognize that in an organization and make it easy for oh for the organization a standardized on a particular solution. And still allow the dozens of two Pizza teams. Independence autonomy if you will, but still using the same solution the rest of the Company uses right?

I wondered like I don't know if you tracked that Atlas, but that's this is ABS. I mean this is you music to my ears where you're describing. I mean, this is totally how we have seen this play out, you know in the early days of Atlas before, you know, we were battle tested and trusted in the enterprise.

We would have smaller companies and startups adopting us primarily and in the years to come subsequent to that time, when we started earning that trust earning that those chops to be understood to be able to run in an operational sets. You asked about sass earlier the pace of operating it true Mission critical sass. It's something he doesn't happen overnight.

You have to build and earn that right. So once you get to the point where you're running some of the largest applications in the world banking government use cases, you know Mission critical software for Consumer great consumer scale applications Etc. You know, once we earn that, right and our customers learn how to build on top of us we started seeing this play out where our customers would utilize our control plan primitive to essentially enabling each of their two feet to teams to experience.

their own level of autonomy and you're hitting the nail on the head that allows those teams to move fast and never feel you know held back but the alternative you could do that. Let's say in just directly on a hyperscaler. So the question might be well, what does Atlas do that's different than that, and I think we talked about it before that.

If you go and use the hyperscaler what ends up happening is you're sprawling across yourself across hundreds of services hundreds of Primitives. And so, you know, I jokingly say imagine that you took a bunch of two beats of teams and ask them to each going to a Home Depot, which is Empower who doesn't love Home Depot, but if you ask each of those two beats of teams go build a house. You're gonna see all kinds of chaos.

I mean the average house built by some random group of people. It's gonna be just chaos. So what if instead you could you know, instead of it being hundreds of services, you can kind of collapse it into the best in breed best in category abstractions that you can compose with.

You're not feeling held back by them. They're just understood to be instead of hundreds of offerings you're building with these really good solid building blocks. And everybody then wins the standardized standardization governance Folks at the Enterprise win because they don't they don't have to be doctored.

No, you know and the builders win because they don't have to try and make sense of building with rebar as well as sinks and everything else in between. So, you know, I think we've definitely seen that Trend and it's funny. I remember speaking to Enterprise developers and I remember them telling me things like, you know, I'd love to use Atlas but my security team would never allow me to use it.

And I always tell them. It's funny you say that because they want it's totally the opposite. You think they won't allow you to use something.

But actually what they want is you to have standardization and governance and and to be building on the ability. Yeah, right. I'm saying that they just don't want to go down the road to even having to worry about the complexity of doing their own internal governance.

But we've seen it time and time again in every major Enterprises we go through the approval process and they kind of understand how we operate and you know, it makes us stronger each step the way or it's just look to be fair to the developer and I'm a security dude. So I'm not banging on security people but to be fair in the developers, it's not it wouldn't be the first time that security people said. No, you can't do that.

All right. I love empowering security people to be like, okay. Absolutely.

Once the security really people really understand this this model of Better abstractions scar rails here in this grooves and it's an environment, right? You can't turn off smooth Network encryption authentication everywhere in Christian everywhere those absolutely out of the gate. You know, it's amazing how You just make it so you can't you just pull away the idea of even having to worry about TLS certificate management.

Yeah, boom encryption over the wire everywhere. Not negotially 2022. That's just normal.

It should be normal. But sadly it's not yet in many contexts. No, but I think that's another thing that's coming in 2023, even at the consumer level.

We're going to see encryption everywhere. Love it. Anyway.

Hey man, where we're running low on time for people who want to get more information and John. On Atlas how weird I mean, I know it's mango DB is it mango DB / Atlas or where do they go just come to our website? I'm gonna be calm you'll find Alice from there.

You know, we have a free forever free tier. It's fantastic for building that next idea. We have 150,000 people signing up every single month doing exactly that it's massive Community.

We have Mom gonna be University with you know, large-scale courses that you can do with your own pace. We've got amazing documentation. We're billed for developers to learn and quickly start building so very community Andrew let's not wait a year and a half to have you back on.

Okay, man. Look forward to it Alan good to see you again. All right, Andrew Davidson Mongo DB Atlas go.

Check it out. We're gonna take a break here on Tech strung TV. We'll be back in a minute.