Greg Irwin  0:18

Hi, everybody, I'm Greg Irwin. I'm one of the moderators here, one of the co-hosts. I'm one of the partners of BWG. And we're partnered up with, with EDB, to talk about Oracle alternatives, you know, enterprise Postgres, and to talk through some real stories about how people are managing through some of those migrations. So, let me do a couple rules of the road. And hey, Peter, and Hardik. Nice to Nice to see you guys. Thanks for Thanks for joining us first off video, if you're able to put it on, it makes it a more, you know, engaging experience, if you can understand, you know, that's, we certainly we certainly know how all this works. The way we run this is a true open group discussion, which means Tom and I can talk, but it's so much better when we have a full round-robin conversation, which means I'm not asking all the questions. Tom's not telling all the stories. And basically people share, like, hey, look, I tried this, it was a disaster, or I tried this and it was great. Or, Hey, I'm really working on and I've got this one issue. And so the way we do that couple things to kind of priming the pump. This chat window here, works really well in this format. So please use it throughout sidebars are good, right, you can have a side conversation, that's that, please do it to be, you know, lean in, you know, jump in with a question. You don't have to raise a hand. This is an open zoom, anybody can jump in and say wait, hold on, can you repeat that or hold on? I got I got, you know, a curveball for you. It makes it so much more fun. So please take advantage of this time that that we have together. And then lastly, you know, we brought together an outstanding group of people use this, I've always found it like, use this to expand your own network. Make one goal very, very simple goal, make one new contact, one new contact, do this. Right? LinkedIn is brilliant. You're you're you're talking with people who went so far as to join this group conversation. So my guess is everybody's open to it. Just you know, take the initiative and reach out if you want come to me, or Corbin at BWG, and we'll be happy to make connections. Lastly, we're doing this with with EDB to drive awareness for what they do. And they are the world's experts in Postgres, among other you know, data architectures, so, you know, you're going to be hearing from them, they are the best we do get to pick the partners we do these with. So if you have opportunities to work with them to do follow up, of course, that's not going to bury the lead. That's a big part of what they're what they're doing here. Okay, enough of my setup. A Tom, would you take a moment here give a little intro on EDB. And on Tom Rieger.

Tom Rieger  3:39

Sure. So this is Tom Rieger, here in the Twin Cities of Minnesota halfway between Portland, Maine and Portland, Oregon, not a place to come get fresh sushi. By the way, if you're going to, you know, think about what you want to do in the Twin Cities in Minnesota. I've been in and around enterprise computing for 35 years, I was 12 when I started if you're doing the math, and I, you know, started my career, the first half as an IT professional doing a lot of what you do today. But back when the mainframe was cool, and Unix was new, and I feel very lucky when I first got out of college 35 years ago at 12, that the first thing I did was this thing called the Unix non-relational database. This was very leading edge. If I fast forward to what we've gone through sense between open systems and all the different buzzword bingos of how we got to 2023. I think there's two fundamental things we're trying to move to next. The database is still this expensive, complicated thing. Secondarily, there's this thing called the cloud out there. And I think if I ask each of you what you think of the cloud, I'll get, you know, more answers and I care to master so whatever you define your next platform decision is as the world is moving from capex to OPEX, because we're also all having to become accountants in our job, but we're also having to come attorneys because of software audits and licensing and what I can do in it. But I can't do. So I've been here as a client engineer, work with customers, your EDB, as he tried to figure out how to get to the cloud, how to get to something that is more compelling from a financial flexibility innovation perspective. So over to you back to you, Greg.

Greg Irwin  5:17

Well, give us a little bit on EDB, now. Oh, sure if he's sorry. So you know, we don't we don't really promote sales pitches. So they'll tell us kind of what it is. And so thinking about that I want to give give a task here for everybody dropped one question into the chat. What's one thing you want to talk about either from Tom or from others on the call, and you'll find I ask questions across the whole group. So don't be surprised if I come to you and kind of ask you ask it to participate. There you go. Yeah.

Tom Rieger  5:51

So great. Thank you for the setup. It's it's weird. It's Thursday. So I made it all about me. It's all about Tom Thursday. So EDB were the largest purveyors of Postgres double. So Postgres came out of Berkeley in the 90s. Dr. Michael Stonebreaker, you know, started off with Postgres always being an open source out there as his iteration beyond Ingress. For those you who remember those back in the day. Well, EDB has always been a Postgres-driven company started almost 20 years ago. And they also have created this version of Postgres called advanced Postgres, that looks and smells like Oracle. So we help make it so much easier for clients to get off of, you know, paying for Larry's Island, dare I say, and onto something that is moving you quickly, you know, into wherever you want that next platform to be with Postgres is kind of the, the repository, and also Postgres, if you go look at like DBengines.com. And they do this great job at looking over time, kind of the trajectory of a database technology, not by revenue receipts, but they have a big algorithm. And Postgres is now the fourth most used database in the world period. And it's been awarded by dB engines and everything else. So it's really come a long way, since it left Berkeley was Stonebreaker, back in the day, and has become very pervasive across all the clouds. And we are the largest contributors to innovation back to Postgres. Because there's a lot of posers out there in the open source world who just want to go download code and go make money off of it. We're always giving back on that front. So back to you, Greg.

Greg Irwin  7:28

Let's go to a customer story. So one, one environment story. You're trying not to give us, the Babe Ruth of both Postgres stories, try and give us, you know, how it normally works when somebody comes from Oracle. And they've got all of the baggage, all of the custom integrations, all of the contracts, and they want to go to cloud. Tell us a little bit about how one customer how it's actually playing out.

Tom Rieger  7:57

I had a customer who on-site, opened up a drawer, just to give us some perspective of how long they'd be doing Oracle. And he says, Here, here's our first Oracle software. To give you perspective, that's how long they've been doing Oracle. For those you don't know what those are, those are called floppy disks. They weren't you

Greg Irwin  8:15

have them in your drawer?

Tom Rieger  8:16

Is that a problem? Because it is? Well, it's in the same drawer with all those old cell phones I've had all the way back to that original Nokia. I mean, to be honest, in terms of entertainment. Alright, so in terms of entertainment value, how many folks on this call, have a drawer in their desk full of old phones? Come on, everybody either either. Somehow raise your hand with Zoom or just raised? I don't know if everybody does. So it, I do. I do. You know, I sometimes like to keep these here just to kind of lighten the conversation. I also tend to you know, I play with these a little bit, too, if you don't know what this is, it's a Raspberry Pi. And the moral of the story, though, is, you know, what you could buy for $30. Today online, it's this small, in which you get for that this whole idea of Moore's law is really pivoting all the ways in which we can do new things today. But the new things we want to do today and tomorrow are hampered based on the technologies and the contracts, and just what stuff costs. So I'll use, I'll use an example that is very near and dear to my heart, because I may have one of their vehicles in my garage. But a large auto manufacturers, US operations are out of Plano, Texas, had this desire to get off of Oracle, it's kind of an edict. And they also have a desire to move into a more flexible world and flexibility was both on premise data center and in the cloud. So their definition of cloud was hybrid cloud, if you define that hybrid cloud is that, you know, you want cloud like flexibility in your datacenter. And they you know, their first question was, where do we start? And also, you know, this this, this oracle stuff we're consuming Make this many cores as much storage. What do you need for Postgres because we got to size stuff. So it started as almost a sizing exercise. So through the sizing exercise, we also then help them realize where to start. And what does that and in the spirit of dating myself, too, because I love them, by the way, is what does that counter that Gantt chart? You know, where do you prioritize? And how do you make the move? Because changing databases is not sexy. I mean, let's be honest, everybody changing a database, your end users to your application, see nothing different, You're just swapping out an ingredient in the mixing bowl, and hoping they don't realize that you know if there's any different flavor to what you're cooking up in the kitchen. So we needed to bring forth something that kept the risk minimal, brought a best practice to the table. And also since we have this idea of how we have this Postgres it looks and smells like Oracle, helping them assess if there's any edge cases of things that they did that are kind of awkward, but then work with their developers, because this is, you know, there's a lot of dimensions to this stuff. You've got the data people, you got the infrastructure people, you got your developers, some of the developers, contractors, some are not they still had a guy sitting in California who keeps our mainframe running, you know, it's like, just this idea of change management of people is hard. But also, no matter what, and I have a I was raised by a mother who's a CPA in the oil industry in Oklahoma. That's why I say words like oil, and y'all. And she taught me the young age, nothing happens until the accountant approves it, which is her way to say no to me, but it's true today and everything we do, nobody is doing the same. For more. Nobody's doing more. For more, you got to do more for less money.

Greg Irwin  11:42

I'm gonna Tom, I'm gonna pull your head on that that was better. Go ahead and hear from Peter was cost. All right, yep. I have that preconceived notion. Oracle's expensive Postgres is spring. That's my my view of the world. So that client, how much did they say the net migration?

Tom Rieger  12:00

Great, great question. So they saved, they dropped their overall environmental costs between us, you know, you get you get upgrade of servers and storage memory plus how much less we save that, you know, how much less our licensing costs where they got an ROI back in nine months, and that including labor costs, so it was not it was an 80% reduction in their in their spent from a from a database perspective.

Greg Irwin  12:32

And we dig into this because I see it, I see it in fallacious question. Dennis's question. Yeah. Okay. There's, and I think we all know the components, there's the license costs go, here's the development cost, yeah. And do your planning and development. And then maybe there's cost savings on the other side, but it's never clean. I know, every Oracle contract I know, is all bundled. So it's really messy. In terms of well, just because I'm no longer using it does, you know, did they just, you know, push the cost to my, my Oracle, you know, my Oracle Application? License? Yeah. Yep. Take us through a little bit more here. timeframe, how much development actually went into it? You know, what kind of data it was, let's let's get into some of the some,

Tom Rieger  13:27

let me let me get deeper here. Because I, because I want to get into some particulars here. Because I think a good story needs details, right. And so this was about 300 terabytes of data. And, and they and they needed to move off of it in under 18 months. Because what they were going to do is anything else, they were running on Oracle after that 18 months, they're basically gonna in their contract, and just buy all new licenses, because to your point, Oracle, this lets you just trickle reduce, which you spend. So you either gotta cut ties and start over, or just figure out how to get all of it off. Because you're still a slave to pay. Are you still running an old PeopleSoft general ledger? Are you running some sort of cots app or off the shelf app that only runs on Oracle, but all the rest of it, they want it off? There was also you know, a lot of this is people costs. It's not just software, I mean, the database, you know, in the total scheme of things, that database is probably what 10 15% of the cost. So the people cost their people reeducation, their consultants, everything else, they were able to to retrain who made sense, they're able to reduce their consulting fees, and then also as you're moving into their cloud choices, AWS. So as they're moving things to the AWS, we also then help them make those right decisions in AWS. So that because I feel that there's a lot of opportunities to overspend in this idea of cloud, because if I think about our conversation here today, it's not just how do I get off of something like an Oracle But also, if my destination is a cloud, how do you do that? Right? Because there's a lot of ways to do that wrong. And I hate to tell everybody, nothing is free in these cloud vendors. So you've got to make those right decisions that that is where it comes into. So as I mentioned earlier, we have this thing called advanced Postgres. And what we've done for last 15 years, is we've made it where this advanced Postgres natively inside the optimizer Postgres, because one of the things that makes Postgres very interesting to the world is it's very extensible. Well, that extensibility has allowed us as at EDB, in creating advanced Postgres to run all of those things that you run inside Oracle today, store procedures, triggers, packages, you know, there's a litany of things to lift and shift those into advanced Postgres with no changes at all. Because you're right, your alternatives. And let's pick on the alternatives, the alternatives are, you rewrite everything, nobody wants to do that we spent the 90s, racing off the mainframe, because of y2k, and going on to a whole lot of new things, and spent a lot of money we didn't want to spend. So through the 21st century, we've been kind of, you know, leaning into that hard. So you want to try to minimize change, you want to try to minimize rewrites. And, you know, my my point there is, and I'll use this auto manufacturer out of Plano, Texas, they had systems that were running in there for over 20 years, the code running inside that Oracle, the people who wrote it didn't work there anymore. So you had this, this, this need to make it easy to lift and shift. And then to give them a lower cost platform, get off of Oracle, but still run those things as they ran because they ran perfectly. And I'm talking shop floor control systems, manufacturing, financing, or leasing of a car, spare parts customer success, on and on and on. So they also couldn't go down. Because what also, Dennis, what we've gotten enamored with is this idea of hey, Oracle's really resilient database, and the cloud is resilient, for instance. So how do we get to that resilience as well, but to that point, and when we're done here, I'll work with Greg and I can send some links out to people to kind of see this an actual bit more, so that you can understand when I say advanced Postgres and how we make Postgres look and smell like Oracle, what that really means in detail.

Greg Irwin  17:24

I hope that helps. I think you got to prove it. Right. So I hear it. And as I'm listening to it, I kind of agree with Dennis, it sounds great. If it works. Absolutely. Well, and is it going to work for my code? And is it going to be optimized in tuned for my code, so that when I run it, not only is it running, but it's running, you know, with, with the same IO that I normally expect, or Yeah, same pet tagging that I normally expect. And this

Tom Rieger  17:52

is especially, you know, since Larry came out with exit data a long time ago, and even RAC before, there's a lot of engineered systems out there that have a lot of layers to them to drive a certain level of resiliency and performance. And, and also, Postgres is open source database. It's just the beginning of the conversation. It's here in the middle. There's all these other things around it, that are either open source projects are we bring to the forefront to deal with monitoring, high availability, disaster recovery, backup migration. Now there's all these other things you need to really do what I call tier one, enterprise ready computing. And that's that's the big gap of Postgres, because you kind of you lingered into this little bit, but it's the Hey, Amy, just go download it. So what do I need you guys for awhile to go download Postgres is one thing to go do 5000 concurrent users against 100 terabyte database running your day to day business with zero downtime? That's a whole nother conversation than just downloading Postgres.

Greg Irwin  18:58

Matteo, Mathias, I see your questions in here I'd like to bring in. But before I do that, I want to go to Nick's question, which is pretty fundamental. What are we talking about? You know, are we talking about a data lake, a data warehouse? There are some questions about like, are we talking about time series data? Or are we talking about? I mean, I know vector is the big conversation right now now, and I read a spec sheet on vector and Postgres and sounded pretty good. So maybe first helped define the common architectures, uses and datasets that you're seeing for for enterprise.

Tom Rieger  19:39

And so it just give everybody a little bit more perspective about me before I worked, you know, where I'm working. Now. I used to work at snowflake and I worked with one of the largest clients they have here in Minnesota down the street from me. And so one thing Postgres is not is something competing with a columnar database. So let's start there. But Everything else is your as I think about your list, it can do very well. For instance, that's Don't forget Postgres has extensibility. Extensibility means a lot of things. It has native JSON support. Because JSON is now the de facto communication medium to put something around the data to give it structure to communicate. Things like time series, things like geospatial I do a lot of things with the Forest Service of the United States for the using, you know, us and post GIS, for instance, as the backbone to everything they do in managing everything you can imagine around the Forest Service, and parks and everything else. So these data, these extensible data types, kind of cover the spectrum in terms of time series, JSON, geospatial and beyond. So, but when it comes to things, like, you know, you hear words like data house, and data vary, you know, data lakes, and some sort of water reference and data, right, that's been around since, you know, Hadoop was sexy, and you can do those things. But there's also just certain limitations that sequel has. So I'm not saying that Postgres is the cure all be all. That's why Mongo and others are around and doing quite well right now. But in terms of the swim lane that I'm talking about, especially where it's highly concurrent, sequel driven workloads, that you want to have all the consistency and availability and all that acidity to it, that makes it really easy for you to get off of something you're doing now that's costing you a lot of money like Oracle, and onto something open source. That's Postgres.

Greg Irwin  21:37

Now, one, one clarification, Tom. And I'm sorry, you guys see, I'm, I'll jump in with questions. I like it better when everyone jumps in with questions. So unmute yourselves. And I'll just make sure it stays on top. One thing that I've always snowflakes, great innovation. Now, as I saw it three or four years ago, was separating the compute from the storage. Yep. Like, they enabled you to really distributed scale, can you can you get that same kind of cost performance of enterprise, Postgres,

Tom Rieger  22:13

not for not for data warehouse, if I'm going to go run, you know, really large, ugly queries. So I've been I've been doing just as much as much stuff in the data warehousing world is I haven't the transactional world, because there are two different worlds transactional world, obviously, large concurrency simple stuff, going on inserts, updates, and deletes data warehousing world, big data loads, big, big, big datasets. And what is happening in data warehousing, especially snowflake and data, bricks and others is now you can have just as many people wanting to query the data warehouse as you had working on the transactional systems. But so to your point, that idea of decoupling storage and compute what they've done is, is great, but you know, what that's happening to Postgres as well, this idea of how you can decouple the two and still have that scalable transactional workload out there, but also in the area of the cloud. How do you do things across regions and create, and here's buzzword bingo Thursday from Tom Rieger, some sort of data mesh, where you have this thing that scales that moves across regions in the cloud, because regions in there, I'm assuming there's nobody here from Google, Azure, or AWS. But all of them only give you four nines of availability. Nobody in their right mind would put their tier one workloads in a cloud, we're in a region is only going to live for four nines. I mean, that's downtime that's unreasonable. We have already done things where you can take these workloads and scale them out across regions with active active types of things, and have data consistency. Because what you don't want see data warehouses tend to have a little more forgiving this, when it comes to a little bit of data loss, they just go find that data and reload it. But when you're running a reservation system or a large transactional system, there's zero data loss. And every record is important. We work with some of the largest financial firms and insurance firms in the world. And, you know, every credit card transaction is important. And every claim records is important on and on and on. So zero data loss and zero downtime. And I'll let everybody else in the call confirm or deny what I'm saying. Cloud is either now a strategic imperative. And no matter what we're moving to the cloud, like the world, they are pivoting to the cloud, bar none stop nothing. I could tell you, the particular cell phone provider I work with, they are literally bonusing people based on how many workloads are getting to the cloud. So in some cases, ironically, readiness doesn't freakin matter because they'll make themselves ready. Now organizations who are being a little more pragmatic about it are definitely trying to figure out how do I how do we move to the cloud and still keep the lights on I mean, because move to the cloud, you're having to keep the plane in the air as you're reinventing the plane. Because there's a lot around integrations and security and performance and all the while my mother, the accountant is behind everybody's ear whispering don't spend too much money. So you've got to, you know, maintain all of that. And, you know, it's very case by case, to be honest, because I'm only the data layer. cloud readiness is a much bigger topic than just the data per se. And I think it depends on either in where are your users? Are they all within, you know, a certain geography or the all over the place is that the general public is at the globe. And then I don't want to forget, especially depending on the industry you're in, there's no lack of regulations, laws, governance, risk compliance, some nefarious person from some country somewhere looking to hack in your systems, like you can't buy, like Clorox bleach, they just had to shut down manufacturer walk in, somebody hacked into their systems. So we got to worry about all these things. So cloud readiness, I think requires a fundamental set of, of establishing and resetting certain characteristics of how you want to run your business. And then to put those right gates in place and security protocols in place and budget controls in place. I mean, I'm speaking very abstractly here, because I think we're all coming. I mean, so the BWG people gave me a very just the companies that are coming to this, and we cover the spectrum here in terms of companies. So everybody's got different priorities and what it just described, some people think that, you know, cybersecurity SEC ops, that whole thing is priority number one, some people make a priority number three. So what is the priorities, do stack, rank those priorities, and then just tick them off? I also find certain clouds are better at worst, it's some aspects of these things. For instance, I know some clients who are using Azure versus AWS, because Azure has certain sikap characteristics that are better. And then I would also love to know of everybody on this call, what is your priority cloud, AWS, Azure or Google? Because I think they all come from three very separate dimensions of how they got where they're a cloud player today, and they play off those strengths. Azure is, you know, you think about Office 365. I mean, I think the Fortune 500 490 marohn, office 365. So that opens the door for Azure. AWS, on the other hand, is aided by retailers. They're starting to be hated by healthcare, because they're getting in the in the pharmacy business. But in other industries, they're the leader. Google, because of their advertising and analytical background are really, really strong and retailers or brand-driven companies. So I see them coming from different dimensions. And I see the answers, by the way, cover the spectrum. I mean, it's all over the board. So this idea of, of of what you were asking Mateus around, you know, this idea of cloud and what's important, I think it's a function of which cloud, we're talking about what industry you're in. And also just the motivation of being my minor in college was statistics and business. So I appreciate what the SEC represents. And also, I think it plays in everyone on this call to what is the most important thing the cloud has to do for you first? Is it save money? Is it resiliency? Is it security? And then also then at my layer now that the database layer? How does the database do that for you? For instance, we one of the things we've added to Postgres is this idea of transparent data encryption. So you have this idea of data at rest, encrypted, controlled by the database, because all the cloud vendors do it at the disk level, they just encrypt the whole disk, or is it performance, and you actually brought up a very good point, I've done a lot of benchmarking in the cloud. And there's a lot of ways in which you can, you know, throttle up different types of computes, or throttle up storage with I ops and throughput. And suppose you get to better performance number not, but you know, what, on like the days of a data center, everything that we do in the data now in provisioning something in the cloud. We, all of us on this call affect what is spent on the total environment. It was easy when it was just like, oh, I'll just go get eight Oracle licenses. And I'll, you know, provision them somewhere in my data center. And we'll all just happen. Now we're responsible for are we going over budget or not with our AWS Azure, Google sprint? So but great points, Matthias and let's let's pick up this conversation after. I mean, that's I think that's part of the point here is because I'm hoping that there's some sort of like, either interesting question mark, kind of going over people's heads or like aha moments with light bulb. So definitely want to Try to take this to the next level. I don't care who on this call is a CIO and who is, you know, I'm just an engineer, you're just a contractor. It doesn't matter who we are in the business. We all can we all see the bigger picture, but we also see the details, right? The devils in the details, but how we avoid the devil is up here somewhere. Right? So, and that devil is a technical devil and a financial devil.

Greg Irwin  30:23

Tom, can you knock off the quick questions on Oracle advanced features? RAC Data Guard Golden Gate? Yeah, yeah, so he's these, you know, it's fairly straightforward to emulate with Postgres they are.

Tom Rieger  30:36

And we bring that in relation to the forefront. So this idea of, for instance, active, active Postgres, and highly resilient Postgres, and I would challenge for every 100 organizations who bought RAC or bought Exadata. They're really not using it for all of those things, either. That's always what I found. Things like Golden Gate are on heterogeneous replication, we have a version of that to go help move things from, let's say, Oracle into Postgres, and vice versa. So it really comes down to also why you want to do these things. But the moral the story is, yes. So there's a couple of things into play here. One, there's this idea of active active has certain things that get kind of scary, like, what about conflict resolution? And to your point, what about things like, you know, serial types, it's like, I've got a unique key that's evolving? Well, we we've added into Postgres, this concept called a globally allocated key. So for instance, if you write something in US West, in AWS, and he gets a certain serial number, and let's say, for that, US West, we have 5 million of those serial numbers assigned to US West. And I'm in US East two, and I'm doing some things and inserting a record in the same table, he gets a different unique key. So what we do is we have these this concept of globally, globally allocated data, the serial numbers, so you don't have conflicts at the at the unique key perspective. But below all of this, what if you change someone's record and you lock it, and over here, I'm locking on this other server, and then we both hit commit? What happens? Well, one of the things that we're evolving into is column level conflict resolution. So if you're changing a zip code, or I'm changing a phone number, guess what? They're both required. So there's a lot that goes on under here to manage the what ifs around that conflict resolution. But it's not just that it's also there's the horizontal scaling. And there's also mundane stuff that, personally, I still think is amazing. On September 21, and 2023. How many of you still deal with Oh, my gosh, the software's end of end of life and of service, Oracle 12 went into life earlier this year, SQL Server 2012 would end a life this year. And everybody's like, Oh, my gosh, you went into life. And that's not what we got to make a change. You go to the cloud, or you even think about in your data center with this idea of active active, you could do rolling upgrades. So you upgrade each component and never bring the environment down. You know, and that's true, whether you're on somebody that went into life or in Postgres, that column security vulnerabilities, SV ease, and so you need to apply that patch. Because Matthias is here working with the SEC, and they can't have any security vulnerabilities, you got to upgrade that thing right now or on your iPhone, there's a new release valve with a bunch of security patches. How do we make those things get applied, and not bring the system down? So there's the the idea of scaling, and how do we handle conflict? There's the idea of, like, for instance, let's not forget us East two went down for 220 minutes earlier this summer. How do we make sure you know, systems still run? But also, how do I then do rolling upgrades and also, as a database as a service, part of what we bring to the table is we'll do the major upgrade versions, Postgres 16 just came out, it's got a lot of new cool stuff. We want you to be able to take advantage of that stuff will do those upgrades for you. So you don't worry about that anymore. If for those you have a Tesla, what's cool about a Tesla, that some reason tomorrow morning, you'll have some new features in the car automatically, because he just got downloaded to your car. We want the same thing, when from a database in the cloud. In my time, especially snowflake, I was purely dedicated to a a competent sounds a lot like who you work for. And it isn't just that data sizes are frightening. Or the fact that we're dealing with data types like old XML stuff, or new JSON stuff to add insult to injury. And, but you can also have no records left behind, because when I think about if that patient needs their Revlimid tomorrow in a UPS box, you can't let that fall through the cracks or bad things happen to that person. Right. So every record is important period, along with you're trying to scale. I mean, think about what happened to I don't think anyone from Southwest Airlines, Southwest Airlines, they like almost shut down for a couple of days because you couldn't scale the business because they had a bunch of 20 year old servers running stuff. So now to your point, it's not like there's less data coming at you guys. Shivani there's more. So how do you scale that and also be cognizant of the accountant behind the scenes whispering in your ear, don't spend more money. So you're always having to, you know, squeeze more, you know, what out of the nickel as my grandfather would say, and, and oh, by the way, keep things secure, keep things resilient, keep things you know, all the same stuff. So again, that's the doing more with less. And that's where we come into play, especially in this idea of, if open source is gonna save you so much money, give you greater innovations give you greater flexibility and how you use it. How do we help you realize that nirvana?

Greg Irwin  35:48

Hey, convinced me not to use Mongo? I assume that's Mongo Atlas. And I know the only one good reason is cost. But, Tom, maybe we pick it up there. It's, it's an awful lot of transactional data. Can you tell us a little bit about how Postgres best way to do is be talking about a customer? You know, do you have any similar customers that that kind of scale, removing that kind of data? And why they chose Postgres compared to a Mongo or an Oracle?

Tom Rieger  36:21

You know, it's, it's interesting. So I think comparing Mongo to Postgres to Oracle, is not as straightforward as we'd all like to think, you know, is are you going SQL or no SQL? And also, are you trying to just lift and shift and save money? Are you trying to re architect things? And also, what is the data? And do you have certain, you know, availability, consistency, you know, integrity requirements, the data is this, you know, what, what's your tooling? And what are you developing over the over the top of it? I personally, very little, I don't run into the Postgres versus Mongo debate very much. In our, in our world, it just people have kind of made the Postgres decision.

Greg Irwin  37:08

So that's what that's going to be requirements, gathering that article do in a different different spot. It's not nothing that we're going to talk about here. But I think it would be interesting to say, all right, politics dealing with intimate, and we all understand your tax data and tax transactions carry a certain level of reliability requirements.

Tom Rieger  37:30

Well, and it also security requirements, because trust me, let's let's not forget the three most important secret of things, we all have our age, our weight and our income. Well, needless to say Intuit knows one of those, probably too, because I got my birthday in there, too. So I heard it, I want you to be successful brother.

Greg Irwin  37:49

So tell us about one customer, if it's public, wonderful. If it's not just talk about them, that a company that were did make sense for them to go. Postgres in the cloud,

Tom Rieger  38:02

I'll use one. There's a company called a light in Chicago, and they are in the insurance business. So if you go open and roll, odds are you go do your open enrollment work through maybe some a light front end for your employer. And they are getting off of an old Z 13 DB two based environment plus a lot of Oracle stuff all on-prem, and they're moving into AWS using us. Because Thank you, hard to use them. So. And they you know, all the while they're looking for that flexibility, that innovation, but also the two most fundamental things, I got to stay up to date with my versions of stuff. And I gotta save money. Because those two things right there, it's, it's hard to get all sexy and innovated by technology. If I'm on a 10-year-old version of a database, and I'm on a stoic platform I can't innovate around, because one of the things we get with the clouds, whether it's Google Azure, or AWS is they all have these really cool things, I can now connect into this, right like a some sort of lambda thing feeding my AWS work over here, or Azure Data Factory feeding the data over here, or the list goes on and on and on. Right, that's innovation in the cloud. But they wanted to start with, we need to modernize. We need to move to a more microservices architecture. And we're, we're, we are, it's the end of the lease on our data center ends on this date and 2024, right, it's that fundamental, because I will say by the way, if you're looking to change database and kind of make a move to the cloud, do that first before you start trying to innovate now, have one big change happen at a time. Don't do both. I saw that happen with a client and they were like, wow, we've decided we want to rewrite some code as we're changing the database and, and it turns into this thing where one week turns into 10 you know, kind of thing. So pick your battles. So the first battle is, let's get off of Oracle. Let's get out of that data center. And then let's innovate, because then you have an idea of the art of the possible. Otherwise, it's all just blogs and documentation on the website, and maybe some random video or a call like this, where people have this aha moment. They're like, well, he mentioned Azure Data Factory into Azure, and they can connect to on-prem. No, no, no, no, let's just let's figure out what makes sense to do first, life is full of small successes, it is not one big wallop. So, you know, that's kind of my take on all this. The beauty though, is unlike, let's say, an Oracle to a Mongo, that, to me is a much bigger leap of faith because you deal with two very different architectures. But when it comes to something like an Oracle, or SQL Server or anything, oh, my God, I had a client who has a bunch of Sybase and I was like, Oh, let me please help you to a Postgres. It's still a very, it's, it's, it's, it's the it's a different slang of the same English. You know, it's like, you know, somebody from Scotland speaking to somebody in Australia, they're still speaking English. It's just a slightly different, you know, nuanced English. So that's easy. And also, one of the things we've discovered too, is we're also the largest creators and providers of education for Postgres. In fact, we have a class called, you know, Postgres for Oracle DBAs. Because we see that sometimes the biggest people who are the headwinds in this journey of innovation are the same people who put the Oracle systems in 20 years ago, they're like, no, no, we're staying warm over the sink. It's like a nice warm fire with s'mores. We don't want this to go away. It's like, to your point, Felicia, there are certain things that are needed little more dictatorial, but I think the world's figured out that they're done paying for Larry's Island, I say sarcastically and that the way forward is something else, just like when I was a snowflake, the choice is 20 years ago by an IBM or a net Teza box or a Teradata box and Oracle box for data warehousing. That was a great decision 20 years ago, today, you can be up and running and snowflake and three hot minutes. So this idea of change as we're the brokers have changed, ironically. But yet, we're the ones who are least likely to want to change. It's not that Postgres and AWS compete if anything, you know, an EC two or, or a cloud platform is really just, it's the new data center. So how to use this new data center paradigm to manage data the right way, so that, you know, all of us sleep better at night and not worry about, oh, my gosh, is it going to run tomorrow? Oh, my gosh, you know, the AWS region went down. Oh, my gosh, it performed great yesterday, not today. Because we're moving away from building the car and buying the car and driving the car and insuring the car to just hitting the Uber button. So the world is changing a lot when it comes to compute.

Greg Irwin  42:52

I think most of us are probably looking at our next one. So Tom, I'm going to wrap us up here. Yep. Everyone, thank you all for taking the time. And, again, if there's an opportunity or you want to dig deeper, that's that's a whole that's the whole point. We'll be being contacted. In the meantime. Thank you all and everybody. Have a great day, Tom. Great job. Thank you.

Tom Rieger  43:15

Thank you, everybody.