Greg Irwin 0:18

Good morning everybody. My name is Greg Irwin. I'm one of the partners at BWG. We've been hosting series around different tech categories for about nine years. And database data structures has been a key focus for for an awful long time. I'm thrilled to be here co hosting with Jan Karremans and Abdul Sheikh. These guys are from enterprise EDB EDB and Cintra respectively. I'm going to give these guys the the forum here for a second just to make sure we have introductions, so you know who you're all speaking with. But before we do that, just a couple guidelines for how we run this, um, you know, I'm going to moderate, I'm going to basically take advantage of the fact that we have Jan and Abdul here and do a little bit of q&a with them for about 10 15 minutes on the kinds of things that they're seeing from different enterprises. I think if you've been on my sessions, you know, I really want to try and also go around the group and hear the stories from everyone. What are you thinking about? Maybe it's fairly plain vanilla, simple, simple stuff. Or maybe it's a fairly complicated environment where you have challenges that might be interesting for the group. The key thing I'm pushing for always on these is the story. What's the story is the story performance is the story. data growth is the story optimization, maybe maybe different data structures, who knows that sort of governance, it could be anything, we're going to talk about it. As we go, please use the chat window. Just do if you want to direct it to somebody say at Terrell at Tiffany, whatever it is, and ask the question, or someone sharing a story that resonates for you layer on layer on your own comments. And if somebody asks a question, please ask questions. Feel free to to provide your feedback and comments, I promise. Some of the best input here may not come from JanAbdul or myself, but from everyone else in the group. This is a it's an interactive forum. So let's take full advantage on one last thing, and this is a sweetener in the spirit of Thanksgiving. Let's do some give back. So what I'd like to do is, it's always fun to spend the wheel at the end of our call, we'll put everyone's name into one of those random name generators, and whoever whoever gets selected will get $200 to a charity of your choice. So a nice little way to get back at we'll do that at the end of the call. And just just a way to another incentive to keep us all properly engaged here through our session. So without further ado, Jan, great to co host with you do us a favor, please introduce yourself and enterprise EDB.

Jan Karremans 3:13

Yes, thank you very

much. Great. My name is Jan Karremans. Sounds very Dutch. Well, that's true because I am based out of what I always say middle east of the Netherlands. I've been personally active in and around databases for the last 25 to 30 years. And like the last five years kind of saw the light of open source migrated myself from Oracle to Postgres, and then kind of joined this company called EDB. quickly figured out you know, EDB really is to Postgres with redhead is to Linux. So what we tried to do his work with our esteemed partners like Cintra engage with enterprise customers in their journey with Postgres. So from from very small initial steps with perhaps departmental or non mission critical applications, really two brand critical applications, making them run on Postgres and well this is what we've been doing for the last 17 odd years or something to fall very nice success I say. There is a lot of attention there is a lot of power around Postgres at the moment. So that's that's what we're really sharing with our customers and we hope to share with you here this morning.

Greg Irwin 4:39

Wonderful and lastly, Jan, what do you what do you do at EDB

Jan Karremans 4:43

my role is on my business card it says product market here and then really kind of entails a little bit the customer facing side of Product Management. But I have from from Long here done a different bunch of different things within the dB. So I tried to get around speaking and kind of work with partners and customers closely together really to get that successful, Postgres established.

Greg Irwin 5:16

Perfect. Perfect. Thank

you. Jan. All right, Abdul, let's do it. Please give a little intro here on Cintra.

Abdul Sheikh  5:23

Thank you. Good morning, everybody. Abdul Sheikh CTO at Cintra. We are a multi cloud multi database systems integrator, with a long heritage in Oracle. I've been in the Oracle world going back to something like version four of the database. starting my career back in European Space Agency in Amsterdam, actually, ironically, Jan is in Holland. Some good roots back there. But yeah, we are a global organization based out of New York and UK covering most of us as well as UK and European operations. But we find ourselves as the architects helping customers assess roadmap, design, build, migrate, and manage their database infrastructures, similar to the ENSET, through and through an Oracle, architect and DBA, historically, but as a global CTO, now listening to our customers and helping them understand, you know, how they evolved from the challenges of legacy architectures on the running on Oracle on prem and where to take them. Myself, we're steeped in Oracle history, but listening to our customers understanding that they want freedom, database freedom, and what that means is commercial freedom, as well as technical freedom. And now he be offering some viable options. This is exciting times in terms of the transition to help customers understand what workloads they have, what fits, you know, what should remain what fits and what did effectively evolve into EDB, Postgres and how to go about doing that. So our mission statement is architecting your success. That's how we portray ourselves as a customer, advocate, customer, trusted advisor. And as many of you understand, in the Oracle world, you don't always go to Oracle directly. For that advice, you need an independent, trusted voice to help you on that journey turns assess current state very quickly, we have some interesting IPLS that I speak with, with the Lizama assessment ruling as well. So yeah, happy to share our experiences so far. And also listen to the feedback we get straight from the audience.

Greg Irwin 7:25

Wonderful. Well, let's get right at it. Abdul I'm doing I'm gonna put you on the spot first to talk about customer journeys. But let me remind people as you can tell, I want to I want everyone to kind of get into the chat. So So please do that as as Abdul sharing his story, do me a favor, and everybody I'm Viola, I'm gonna I'm gonna put you on mute a little background noise. So everyone careful, a background noise, happy for somebody to jump in with a question. But while while everyone's multitasking, please just keep this on mute. So as we go, what I'd like everyone to do is to share one thing that you'd most like to hear about from the others on the call. Maybe it's performance, maybe it's how you manage how you manage environment, how you migrate an environment, what's the one thing you most want to hear about today? So that this is a productive session for you, Abdul

I'm now gonna ask you to tell us as you're speaking with organizations, what are the problems that you're typically uncovering? In terms of their environment or their journey? What are the typical problems? You're seeing enterprise?

Abdul Sheikh 8:37

Yeah, number one, step one is understand where a customer is today. So an inventory a deep detail industry inventory of what they're seeing on in terms of their Oracle estates, summarize some big customers or you know, smaller nature, just a few servers and a few databases, that it's fairly easy to get our heads around. Our large enterprise customers are running hundreds of servers, servers and hundreds of databases. That's always a challenge, right? So some of you may have a very well implemented OEM Enterprise Manager on the Oracle side gives you a nice inventory. And so the Software Asset Management of your of your database is stable. So understand what you have not in terms of just pure configuration, but in terms of SLA tearing, which are the most important databases, you know, how do they support the business? What's the revenue profile? What's the SLA profile? What's the security stance of each database? Getting a true detailed view of current state and understanding? You know, what are the candidates? Where should I start in terms of the low hanging fruit? Where can I take some risk? You know, where can I demonstrate some of the immediate short term wins? The first step typically is where am I today? And how do I show management that this is a viable journey to take? The business in general is going to be risk averse, in terms of, you know, taking our crown jewels and our most critical high performance databases that work well on Oracle you know, let's let's submit the Oracle Good engine. However, some of the challenges around business practices or the commercial licensing side of things are some of the drivers here. But ultimately, what do I have today? And where do I go and what I've one of the immediate candidates, in our experience in general, all customers want to show some short term wins, get the attention of management in terms of it is a viable journey. Here's some success in the short term, unless get those early successes under our belt, to them gain the build, build, build, build a business case, for both the freeing up the cost base of what you're sitting on today, in terms of the Oracle cost profile, whether it's infrastructure, licensing, software manpower, and freeing that up to justify the journey to Postgres. And, you know, we'll talk today about very simply, what are the easy, medium and more complex journeys? Which databases have limited business logic, you know, sitting in a database, and those will be fairly easy. migrations, which ones are more complex and require higher availability. So, where am I today? Where, you know, where am I going? What are the short, what are the immediate, low hanging fruit candidates? And how do I build a business case,

Greg Irwin 11:09

I'm going to try and dig in with you here to maybe hit on some of those where you've seen those, I love it the the easy wins. Also, we talk a lot about eliminating some of the license cost of Oracle maintenance cost of Oracle, I'd love to talk about real world. In a just because you drop some clusters, can you truly recover the license cost? Or are there gotchas in the licensing model that doesn't make it so easy to to recover the cost? So get in go go a level deeper in terms of how people are actually able to start chatting real, real cost? So let's disable that. Let's Oh, no, I'm sorry. If you're if you're able to respond on that in terms Yeah, really

Abdul Sheikh  12:00

quickly. So those the key net is wins. But the key today is a to two to two fold strategy. Number one, technical architecture understanding, you know, what are the low hanging fruits in Android yet? As I said, the second is commercial architecting. So understanding how your contracts are structured, do I have lots of multiple small CSAs? With Oracle? Do I have a single big Ula, that we have not only technical expertise, but also commercial expertise of Central, we can actually run the same audit that Oracle LMS run, but we do in a soft manner under NDA with customers to dig into what our license usage footprint is, and also what our licensed construct is. That'll give us a sense of, Am I able to start to carve up my Oracle license ownership and start to shelve or cancel some of those existing CSAs in a manner that's aligned to my roadmap for my application migrations to EDB, Postgres as well, really, that's the answer, a, an alignment of technical architecting, roadmapping, as well as commercial roadmapping. And very simply, and we can help with that, sort of, we have the expertise. A lot of X Oracle commercial licensing expert sits within our global commercial Oracle practice. So happy to have a conversation.

Greg Irwin 13:13

Excellent.

Abdul. Thank you. Jan, I'd like to turn to you at the same. The same question. Your your conversations, share with us, share with us a story. Sure, if you can choose a name of a company, you're allowed or not, don't worry about it doesn't matter. What we care about is the story. Yeah.

Jan Karremans  13:35

So this this, the first thing here would be a story about one of the largest privately owned pharmaceutical company, they did kind of fairly bold approach. So when when they contacted us for the first time, their story was, you know, we're gonna run Postgres on containers, and we want to do a POC to see if that's a viable thing. So we got into contact with them, we kind of started investigating, you know, how do you run a database in a container and mind you this is, this is for four and a half years ago, I guess now. And at that time, we were in this far up front with that. So it was it was a precarious effort. But when we got into contact with the customer, actually, the story was very much different. They had already decided that they would going to implement and execute an Oracle two zero program, meaning everything that they were doing with Oracle, not just from a database perspective, but also from a support perspective, from an hardware perspective, and from an application perspective would be completely removed from the equation there and that was a very interesting twist that we had. So it For them, it was really all of these things you hear in the relationship between customers and Oracle on reliability on cost on. Yeah, well, actually translating to business criticalities, kind of beat them in, in the bind to put it like that. And that is what then made them decide to kind of start on this, this truly Oracle to Zero program. And that is what they're executing at this point. And that that also goes back to what I'm doing,

Greg Irwin 15:34

how embedded were they with Oracle was Oracle, the heart and lungs and yeah, and nervous system of the company.

Jan Karremans 15:42

So these are, these are very old company, I think they were created 1870s or something, right. So in when computers came, the first database that they started using was Oracle. And that kind of really, in the beginning, it was good to deal with Oracle, right, I come from an Oracle background myself. And if I look back 25 years, Oracle was a good company to work with, they had a lot of technology that will drive. And there was still a lot of focus on technology. So that's what they also started with. And this started obviously changing over the course of the years, as we've as we probably all know, here on the call. Right? And that kind of prompted them to take this fairly rigorous decision.

Greg Irwin 16:29

Jan, how's it going?

Now, tell us about tell us about some of the details on you know.

Jan Karremans 16:43

Yeah, so that's kind of what I was saying to the point of dual it's, it's a commercial architecture, and it's a technical architecture. So they're now still figuring out how to start dealing with ULAs and how exactly, they're going to translate some of this, some of these changes to actual cost efficiency for them. But that kind of brings us to the point cost wasn't their first driver, it might be your first thing that you look on. But in these kinds of migrations, at some point, cost doesn't become the first driver anymore. There are other benefits you get from working with the free open platform, that that kind of come to the forefront, right. So the commercial architecture is currently being done. And if you start looking at the technical architecture, it is very much you know, database classification, start with the low hanging fruit. Figuring out how you do basic things like backup recovery, how do you integrate with your enterprise backup technology,

and some of the ideas that seem very logical, because that is the way that Oracle preached it, you know, you use our men, and you have the specific setups to do an enterprise backup. But they aren't necessarily always true with Postgres, there is a new and shiny way, if you'd like to look at some of these, some of these functions that you need around database, disaster resilience, high availability, or to extreme high availability, and all of these things, you need to kind of step back and reconsider

Greg Irwin 18:30

a new thought, again, I want to keep going with this because this is the this is the good part. This is this is the meat of the conversation. When you throw away the old paradigm, and you go to, if you want to take backup and recovery, what's the advantage? What's the tangible advantage that you can realize by coming to, you know, backup on on with an open source, Enterprise DB model.

Jan Karremans  18:58

And that kind of goes back to how Postgres was built, right. So based on the original ideas of Stonebreaker, to build a very flexible, agile platform, combined with the fact that it's open source, you know, it's used by the people that need to use it every day, or it's built by the people that need to use it everyday. So it's extremely easy to kind of script and work with. But that also means that were in the past, you would have a solution, and you would start looking okay, so I have a database solution, or I have another kind of solution in my organization. What are some of the elements that I can start fitting around, you have a database, you have an Oracle database, and that's cool that you can make that work with this and this and that solution. If you take a Postgres database, it's so much more flexible and malleable, that you actually say You know, this is this is the IT architecture we have in the organization. How do we let Postgres fit in that architecture, rather than the other way around? Right? It's a little bit of the paradigm shift we had in the 80s, where we said, you know, we're going to buy this enterprise application. And now we're going to restructure the way we work to fit with, let's name it SAP Business suite or whatever. We don't do that today. Today, we say, this is the way we work in an organization. This is the way you know, we've architected and this is what feels good for us. And now we need tools that support that. And Postgres is one of those tools that you can mold in such a way that it will support the way that you want to work.

Greg Irwin 20:44

Yeah, yeah,

I understand that. The concepts, how much money is the pharma company, going to save at the end of this journey?

Jan Karremans 20:55

I'm

that is hard to say, because they haven't given us a lot of insight into their cost structures. But if you simply start counting course.

Greg Irwin 21:09

Yeah, give it give us a round round

number. Is a millions of euros, or dollars and two cents. Do they

ever you said Money isn't the same? Isn't the key as the key driver, you're able to break the paradigm, we have efficiency? Yes. Are there any other major KPI improvements, performance,

Jan Karremans  21:36

were the biggest the second biggest. So the first one was really to get to break the bonds of this vendor that dictates what needs to happen, right? As a pharma company they're on, on in specific areas on the cutting edge of research and development for drugs for both animals and humans. Right. So having a tool and a platform that they can extend a mold to help in that area, and is one of the things that is bringing benefit to them. Another thing is they're able to use this to start saying to vendors, you know, you've been coming to us with these, and these are those requirements. And, you know, we now demand from you that you make this work on Postgres and kind of fit also fit in the existing architecture that we have. So those are kind of the things that in that area are changes as well.

Greg Irwin 22:40

Very good. Well, I want to just give a shout out here to Jenny, Roger Terrell, Geo, and Andrew for adding some questions. And let me ask everybody else, please ask or add your own. And let's get into some of this. Andrew, Andrew, you and I have spoken from time to time over the years. It's nice to speak with you again. Do us a favor, you asked a question about latency for front end applications. Tell us a little bit about the environment that you're you're thinking about? Sure.

Andrew 23:14

Hi, Greg. Thanks, everyone. So we already have like an enterprise data warehouse, the thought process, which we are, which I have is, so there are some legacy warehouses in Oracle. So and the front end application is like, probably I won't give like more details. Like let's say, if you think of like a trade application, like if you log into fatality, like the different types of details we would see Right? Like, how much positions I have, how much holdings I have, like, what are the trades I did in the past few weeks, few months, etc, like all of those details. So right now, those details are fed directly from market, like I would say Oracle Exadata. So my question was like, would that use case obviously, there are like various tools which we can put in like add scale or stos. And maybe like Tableau or something on top of it. But right now, what happens is, there is like an UI, which goes and fetches the data from Oracle Exadata for every request, and the latency is like, it's very good. Like it's in like milliseconds. And that's good. So my question is, with this platform, would that be achievable? Yeah, that's like a very open ended question. There is a lot of parameters which goes into it. But at a high level, would that use case be possible with this database platform?

Greg Irwin 24:33

Do you want to take a shot? You've done? You've done these kinds of architectures and designs, what are the challenges?

Abdul Sheikh  24:40

Let's translate the word latency, right means performance in terms of responsiveness of the database, not necessarily latency in terms of network. So we always just generally, in terms of topology, the app tier and the database tiers are relatively close, close by, as a side note, any customers tried to put the database in the database in the cloud and left the app tier on prem. Those always go issues, uh, specifically regarding accidents. And we have a world of experience on Exadata. And some of the larger databases, especially data warehouses are approaching a 50 6070 to 100 terabytes in size, they're always going to be a challenge exudates, it does a pretty good job of, you know, breaking down that workload and responding in a, in a good manner. So I think this also kind of relates back to a question Terrell has in the chat as well, which is where are we in the adoption? I think that's important to understand. It's still early, right, where we see some early adopters, we've done I think it'd be can speak to that over the last few years. But in terms of the journey, now is where really the inflection point and a real turning point has been this year. Ironically, probably COVID, has really given people a chance to take a look and some time sitting at home to assess where they are. But we're finding that customers have just taken a breather during the last 18 months and said, okay, there is a viable option. Where do I start? Where do I assess? And to be frank, those big databases, those monster databases sitting on Exadata are not on the front of the list, there are no more hard, complex bucket, but they're always gonna be difficult. And I think that the pattern we're seeing there is not a major lift and shift of 100 terabyte database in GDB, you know, but what we're seeing is customers dividing up those respective workloads within that large accident, and saying, should this go to BigQuery? Should this go to snowflake? Can this go to Postgres, and so on. So breaking, conquering and dividing and conquering that workload into manageable pieces? Because at the end of the day, I think generally customers feel that databases have grown organically. data archiving is not always at the forefront of list of, you know, priority projects. And they grow, they get very large, customer base organizations get very dependent on them. And extra data more often than not, is usually the only answer. So you have to take a good hard look at how do we architect these databases going forward? Break them down into manageable pieces, and then move those pieces into respective platforms? That makes sense. And that's certainly what you know, what we see is the the trend there. Gotcha.

Yeah, Jan, what's your experience there any anything to add?

Jan Karremans 27:13

It's very much kind of depends on where and I spoke about this earlier, where they are on their posters journey on the post Christian maturity level, what we do see his his organizations like for instance, MasterCard, who have since years and years moved all of their IT infrastructure to Postgres. So they are also running these kinds of extremely critical solutions with extreme high workloads on Postgres, you know, and if somebody comes to me and say, you know, I have accelerator, it doesn't really mean, much to me personally, because Exadata in the end, it's a machine, right? We actually have customers running Postgres on Exadata, because it's a very good platform for running transactional database loads. Um, so it really and that that goes to your point Abdul as to is to what is actually running on that Exadata machine? What was workload build up? How was it split, et cetera? Where are you in your Postgres journey? Do you have the trust that Postgres can do it, because it's, it's something you need to experience I do. We've all had, like, the opportunity to work with, with Oracle systems for for many years, and we've seen how these kinds of workloads behave. But we haven't done that necessary on Postgres. So we need to build that maturity, we need to build that confidence. So we start with smaller applications, just to get crews and organizations used to working with Postgres, and that can scale up quite quickly, right? We run typically do, I don't know, 10,000 TPS on a Postgres database sizes, and also necessarily the biggest challenge, it's about, you know, how big is your active works that? Can you put it in memory? What kind of queries are you running? How many updates? How many changes? Are you running, etc. So really classifying workloads, before making the right decision on what a Postgres architecture would look like? Is really some of the things that need to be done. Got it?

Greg Irwin 29:26

Got excellent, no, this is this, these questions, create a nice base for conversation. So please keep the questions going. I want to go go to Oh, did I miss? I think it was, it was Jeremy, who shared some of kind of the real world challenges of a mixed database environment. That Jeremy Would you mind just sharing a little bit of color to your story of pay starting to introduce new databases? How do you handle patching? How do you handle you know, compliance and And and backup across these heterogeneous environments, maybe share a little bit of an anecdote or or question. And we'll let JanAbdul and others jump in with some stories in terms of how how to still optimize around to mixed database environment. Yeah, sure.

Jeremy 30:19

So we're both an Oracle and Microsoft SQL Server shop, the biggest workload runs on Oracle. We, we aren't running like some companies run, you know, one application, right or two applications, that's their big thing, like a retail or runs. We run cart. Cargill is a multinational company. So we're running their 78 division. So the company we're running, every type of application that exists in the world we're running. So it's not like one thing where we're tackling, you know, this ERP. And now we want to move that we've got just about every ERP that exists. We've got trading systems, we've got plant, floor systems, everything. So it's very difficult to just pick up an application and move it because we have hundreds of them. In our Oracle environment is extremely mature, very automated, our operations, tasks. Backups, everything is automated. There's so little, so few tickets being generated, because we we intercept issues with Enterprise Manager, we've customized that we react to things. So now we're starting out brand new with Postgres and MySQL. Somebody asked why to? Well, we not all applications that we have, the vendors don't support both Postgres and MySQL. So we, we need both of them. So how do we mature those operationally? What tools are people, other companies using? So far, we're going with community support. But in trying to save right, because of the high Oracle costs, support costs, licensing costs, one of the things I mentioned in the chat was, we're running, we've consolidated on Exadata. But again, it's an exit, we've got Exadata is running 300 databases. And they're not they're 300 Different applications for each database is a different application. So what's the goal? How do we how do we mature this

Greg Irwin 32:50

journey? What's the five year goal? If there is a CIO level goal for your data architecture? Is it? Is it costing rising costs? Yes,

Jeremy 33:00

it's reducing Oracle's reducing Oracle licensing costs. That's, that's really it. Yeah, very simple. And as I mentioned in the chat, is trying to use some of the features that make Oracle the value, right, you're trying to use some Oracle features that the other vendors don't have. Except to put that on Exadata. If we've got 32 cores on it, or 64 cores on there, and only one database needs it. Right, it doesn't make financial sense. Oracle makes us license all the cores.

Jan Karremans  33:41

There, there is

a little bit of a double approach to that. First one, and I think Greg also asked them, what is what is, so the goal is cost reduction, and the first thing, and so we can check that off? Right? And the other thing

Giovanni 33:57

is, is this something

Jan Karremans  34:01

where it's a feature place? Or do we see with some of our customers saying, you know, I want to have Active Data Guard, it doesn't, not necessarily for its functionality, but just you know, somebody told me that for this kind of functionality, we need Active Data Guard, and specifically, if you then start looking at, okay, but what are actually the requirements that need to be fulfilled, then you break it down, kind of, you know, these are the four or five themes around RTO RPO, around a high availability. So a number of nines of availability. If you really, truly start digging into that and get honest answers from the business. I'm not just saying you know, we need seven nines of availability just because it sounds nice, but what does it translate to? What are costs of these? You know, I got four I got five, six and seven nines of availability, and what are actually the costs to achieve each and every one of those, and then start looking, how can I start filling those in would be another. And then again, back to that, like, very well aware of that I sound a little bit like a broken record. Again, this, you know, where are you in your Postgres majority, you're saying, you know, we are using Postgres, we've got community support. That sounds to me like a very, very first step. So I wouldn't in that sense, even start thinking about taking applications from an extra data, moving those to Postgres, I would really first start, you know, internal systems, perhaps a ticketing system, put that on Postgres and get get the experience figure out in what way you would need a vendor like EDB on the one hand, or partner like Cintra, on the other hand, to kind of give you a vendor quality support for Postgres for azules. And some of the other things, but also for guidance in experience building for Postgres, and actually learning these specific and sometimes a little bit more, sometimes a little less subtle differences between Oracle and Postgres, get your head around that and then start deciding. Okay, so now we've got this baseline set, what would be a good next step of applications to look like there are a lot of third party vendors these days that support Postgres because it's everywhere, right? So this is what you start seeing the company actually that I talked to building an airline analysis to the vendors, either you guys support Postgres, or we don't buy your application, or we're going to phase it out. Right. So those are also things you could good start considering, as you start building up that letter of Postbus. Maturity. So the same. Let me

Abdul Sheikh  37:03

add to that. Yeah. And if again, Jeremy, I understand where you're coming from, let's stick step back, understand why we're an extra day, right? One of the biggest drivers for consolidation in the Oracle world is cost, right? Why am I, if I run lots of silos of different servers, and Oracle is going to have a hay day and they come and license the hell out of me. So ultimately, your consolidated and Alexa data because it provides the lowest CPU footprint overall, for a larger database state and the lowest possible license footprint, as opposed to wait, put it across lots of commodity hardware, and even VMware, you get, again, on the licensing side of things. So I kind of agree down in terms of you're saying, but I think there is opportunity when customers put all their eggs in one large data basket to see that there are some critical databases that but there may be also some tier two, tier three apps that just find themselves on that platform. Because it's, you know, that's where we've invested in the licensing. That's where, therefore we let's enroll there. So we do find that our data customers are either, you know, a high degree of OLTP consolidation, and there's all TP databases are candidates within within there for the Postgres migration. And then some of the the larger data warehouse is, of course, more difficult to deal with. So I think there's both sides of the argument. They're valid.

Jan Karremans  38:23

I absolutely agree. And that was also in the beginning, why I asked, you know, if, if this is also polished and ready, why would you look to move to Postgres in the first glimpse, right. So if it's a pure monetary exercise, you really need to also kind of do the math, and make sure that it actually works out for you in the end, because a migration is an effort, that that obviously has a cost factor in it. So it makes sense to do ROI is for for those migrations, and to really get a firm grasp on on, you know, if I put in this effort, does it bring me what I expect? Do I actually reduce costs? Or am I just adding the next database platform due to the group of platforms that I already have?

Abdul Sheikh  39:13

I think there's a there's a question that's related from Roger in the chat, which is, if I'm if what is the cost profile comparison? How does Postgres MySQL compare? But just to my point, if you're if we find a high degree of consolidation in the Oracle, because of the cost drivers, then can I unconsolidated and go with a more, you know, dedicated server to database type approach in the Postgres world? Can you talk about that? Yeah. And in terms of, obviously, there's there's much less of a restriction in terms of needing to consolidate because of the how does the Plus how does the EDB license model compare, allowing us to have a more disparate kind of deployment model.

Jan Karremans  39:53

So in that sense, and in my head, I was already looking a little bit at containerization. I'll come to that in a second.

Greg Irwin 40:02

If you use something

Jan Karremans 40:03

on a machine, and if you use so the approach would be I think that's a better angle, the approach would be to go to a much more separate environment where you would say, you know, an application has a Postgres cluster that lives on a VM. In that sentence, you would only have to license that or get subscriptions. Actually, for that one machine. The interesting part is that there is, you know, if you've got a Postgres or an or an EDB, Postgres is on server subscription for a machine for a specific set, of course, it already includes everything right, there is no, with the exception of one, there is no need to get extra functionality at it. So everything from backup, high availability, and etc, is already included in the subscription. So that's, that's purely looking from a commercial perspective. If I have a 16 core machine, I get 16 subscriptions. And I've covered doesn't matter what you want to do from a monitoring backup disaster recovery perspective. The only exception is extreme high availability. So if you have, like what I said five, five nines of availability requirements, then we have one additional tool that is that is sold separately. That's called post res BDR, which is actually a multi master for geo distributed implementation for Postgres. So that would be the only additional tool that would come into that, but to also just go for the course that actually use it. Right. So does that answer the question?

Greg Irwin 41:52

Yes. Let's keep going around. I think the stories are interesting here. Roger, I'm going to ask you, if you're able to just state your comments in terms of support and put them and let's lay that one bare in terms of available resources around Postgres versus MySQL. And also let's let's try and hit square, like some of the the trade offs. Postgres versus MySQL. Roger, wants you want you to state your question for the group? Chair? Well, I'm concerned

Roger 42:24

about putting in production, a Postgres database that I how can I get support? Does that now mean, I have to have a relationship with a consulting firm forever to maintain that, can I get expertise in that? Can I get DBAs that really understand it down to that level? That that easily? So that's, that's one of my concerns about going into production?

Greg Irwin 42:44

Let's, what's the what's the experience here of know, for those who have put in production systems at scale? What's their ability to find skilled resources to train resources? What's the what's the real situation?

Abdul Sheikh  42:57

I can pick up the neck, Greg, if you like. We've taken the decision that there's this fairly good enough compatibility between Oracle and Postgres and some understanding of concepts that we retrained our entire staff from Oracle, Dorking, Postgres, so rather than, for example, on the SQL server side, it's a specific, you know, tech stack. So we have a specialized team on SQL Server, bought on the Oracle to Postgres, specifically, because of our consultants and our managed services teams are managing both platforms, we've retrained all of our staff from directly from our clusters. There are SMEs and experts out there, you know, there are early adopters and DBAs, who saw this ahead of time and, you know, moved into them move their career into this couple of years ago. And there's, those gems are difficult to find. But certainly, what we've done is identified a few of those, take them on board within our organization a few years ago, and then help them you know, actively train the trainer kind of concept to retrain the rest of our staff.

Jan Karremans  43:57

From a very, very personal perspective, I can answer that as well. In previous lives, I have been an Oracle DBA. And obviously, when you joined EVP that doesn't really make much sense, right to be an Oracle DBA and Postgres shop. So it literally took me a week to take most of my Oracle skills and actually translate them to Postgres skills. The technologies on the one hand are so similar, that you know, all the concepts that you have around relational databases that both systems are based on the work that Goldendale did back in the day, right. So if you got that concept, you can translate that in your head fairly easy. So an Oracle DBA would be, you know, given a little bit of guidance and a little bit of, of enthusiasm, migrate himself from Oracle to Postgres also fairly quickly.

Greg Irwin 44:57

That's a great feedback. Thank you. I just saw

Jan Karremans  45:03

from from Giovanni, the question around Kubernetes. And it was the other topic that I would wanted to make on a previous question. In consolidation, what you see happening these days is, with the way that Postgres is built, it actually runs extremely well in Kubernetes. So EDB is part of the Kubernetes community, we are founding member of the data on Kubernetes. community, it's also that's a sub community in the Kubernetes community really focused, really emphasizing on running Postgres in the Kubernetes environment. So if you start looking at what we bring to market currently, as well as what we call cloud, native Postgres, which are immutable container images, combined with a Kubernetes operator, actually gives you the ability to run Postgres in the Kubernetes environment. And we have a bunch of customers that already do that in production for a variety of workloads. So if you want to D consolidate. So all the effort we did in the past to consolidate on these big machines, is now being broken down to kind of D, consolidate, using Kubernetes to get every application its own database, or if you if you start converting to microservices, one application cannot have 10 to 100 databases in itself, right. Running in a fully microservices architecture.

Greg Irwin 46:46

Let's, let's keep going. Looking at our clock we've got about just over 10 minutes. So let's let's finish strong here. We've got great questions in and let's try and get a couple of them in and then we'll we'll spin the wheel. And and I'm sure everyone has a busy has a busy Monday in front of them. Kimberly, I'd like to invite you in to share some of your story because I hear the same thing going Oracle to Postgres or Oracle. So my question for you, Kimberly is what's the number one macro challenge you see in the migration that you and your team are trying to design?

Oh, Kimberly, are you with us? Nope.

Oh, we see you but we aren't here you evening. Stay active in the chat. We'll keep we'll keep going around. If you're able to jump in, please do. Pop Up. Burke. verkeerd been kind enough to turn on your your video. Do us a favor. Tell us a little bit about one. What what what what's your interest here on the topic? And what are some of the challenges you want to understand?

Burke 48:01

Hello, everyone. My name is work. I'm from, and I'm a DBA database administrator and application switcher in Scotiabank ganadores expert. We have the traditional RDBMS systems, the Scotiabank, which you might guess these are the Oracle and SQL databases on top of that we have Postgres SQL, databases and DB two databases. So the general approach that we are following is that so we have these applications like and I was listening to Jeremy and he was saying that he had all these kinds of applications that they're supporting. And sometimes it's not up to us to use different kinds of databases like Postgres SQL database, although we would have liked to use Postgres SQL, and other databases. And the thing is that it's just the applications that support these a conventional databases like Oracle and SQL. So we are a little bit limited on that side. But we also have our own custom projects that we want to use these databases on, for example, we have aI projects that we want to be leveraging. And for these kinds of projects, we need to be having file based databases or no SQL databases. And I was also looking into those aspects of Postgres SQL and how supportive they are on those kinds of aspects. So that's the thing right now. That's, that's the approach that we are trying to achieve. Let's say that's the goal that we are trying to achieve.

Greg Irwin 49:48

But I'm sorry, I think I'm afraid I missed. What's, what's the operational improvement over the next several years you're trying to get at? Is it just better support for what you have or is there is there A sense of transformation could be as simple as move databases in the cloud.

Abdul Sheikh 50:05

Greg, it sounds like innovation, not to interrupt, but innovation sounds like you're your driver. Right? Well,

Burke 50:11

so one aspect is licensing, of course, we want to be minimizing the Oracle licensing costs as much as possible, in addition to the IBM licensing costs. But the second thing is like, like Abdul was saying is like, we want to leverage the features, the cutting edge features of the of these databases as much as possible. So how can we do that? So we are trying to create our own projects, just like I said, we are trying to bring in machine learning in AI, as well. And so are we going to do that? So we need some very fast databases in different kinds of platforms, like PL SQL could be one of them, but could be my school as well. In order to manage the these kinds of Yeah, thanks.

Greg Irwin 51:07

For that, thank you very much. I appreciate it. Appreciate it. Abdul I'm going to come to you because I think that you hit it, which, as somebody who's trying to drive innovation, I heard IoT I heard I've, we've not talked very much about AI or cognitive services. But what's the ability for Postgres to perform at those levels for potentially pretty, pretty, pretty, pretty high performance hurdles?

Abdul Sheikh 51:37

I think it's a product question more for yen but don't speak at a high level in terms of the constraints that were dealt with in the Oracle world don't exist in the EDB Postgres world. So open source generally lends itself to innovation, you know, strategies, AI, ml, anything that's, you know, cloud native effectively. And that's what we see, generally, Postgres is see a high degree of production, disability. The break point, though, we also see that organizations are split between vendor based applications and custom internal applications, those custom internal initiatives are where the innovation happens. And businesses should be striving towards doing less of the running, maintain, keeping the lights on or have the legacy platform and driving all their resources and manpower as much as possible to drive that cutting edge in our business. But maybe you can speak to the specifics of Postgres in those workloads. Yeah,

Jan Karremans  52:30

thanks a little. This, this is this is exactly what you see happening, right? If you look at the development of Postgres, as a project, it is actually always developing like seven or eight different directions all at the same time. So one of the one of the leading things is some of these new technologies. So a colleague of mine, who is actually a Postgres core team member by the name of Dave page, wrote a blog on how to do artificial intelligence inside Postgres livery, leveraging Python, for instance. So that is stuff that is actually being done already. If you then start speaking about, you know, how fast is the database? How fast is Postgres? How fast is Postgres in document management? So with with Jason, Jason data types, I think we, some of us have seen the announcement that Oracle now has native JSON compatibility with, with Oracle 21 C, or something, which is, like 789 years after Postgres already had native Jason D, document data type in, in the database. So you can imagine the advantage on handling JSON documents inside of databases Postgres compared to some of these newcomers like Oracle, right? So the innovation level in a Postgres database really is dependent on what where the market goes, the product is in the end built by those who are actually the market, right. So

these innovations are happening first and foremost in Postgres, simply because the folks that build the product are the folks that need it. And it's not dependent of a single vendor saying, you know, I've got 10,000 development hours to spend on this, there is an infinite amount of capacity to build features and functionalities inside Postgres, as long as the demand is there, right. So these are some of the things we see around distributed databases around.

Greg Irwin 54:50

Machine learning

Jan Karremans  54:51

AI, and there are a bunch of these directions where Postgres currently is being driven.

Greg Irwin 54:57

Excellent. Yeah. Jan. Thank you, folks. Let's go one more thread. I'd like to invite Giovanni and geo Giovanni, thank you for your your comments in the chat. Do us a favor, what I'd like to hear about is this idea of managing a heterogeneous environment. And it sounds like just from your comments in the chat, that you've had some experience with that. So I'd love to hear maybe some of the successes or challenges you've seen in terms of that, that this group can all think about, in terms of, you know, making this

Giovanni 55:31

journey. So you we use, like, last night, I was on a company that I work with SQL Server, PostgreSQL, MySQL, MongoDB. And Cassandra, the greatest challenge is, and also a great team of five, or six, database professionals. That's a lot of work for a small team. So the first challenge is monitoring we use carefully we use data, which allows us to monitor to centralize monitoring on on, which is a big challenge. Another thing is, is is, and this is why we should when we bring in a new technology, we should be very careful. Because having a if we have when you have to deal with cannot with another, right? People always come and go is really hard to get a database professional. That is experts on and when I say expert, I mean experts of five platforms like MySQL, it's, I would say it's almost impossible. There might be some gems I like those said, but they're hard to find. So

Greg Irwin 56:58

God, let me just ask, have you had success? Give me Give me a one to 10 Score 10 being best. One being worse? How? How successful have you been been in monitoring, being able to get a quality alerting and telemetry and then doing a root problem solving across a heterogeneous environment? Given give us a score?

Giovanni 57:19

I would say. I would say about a make.

Greg Irwin 57:25

You figured it out? You have a small team? You figured it out?

Giovanni 57:29

Yeah, yeah. But it was a challenge. And about, I mean, Id and the thing is that, that when when you are a DBA, you you don't get to say on which technologies the developers are going to use. So you just have to handle what they throw at you. But if given the chance of, of having a say in the architecture, I will simplify more, for example, MySQL and PostgreSQL. I consider them to believe a bit redundant. And I'm favoring PostgreSQL because it's more robust to my opinion. And also SQL Server is a little bit redundant on I will say PostgreSQL. And SQL Server, I'm more like, I'm not the same, but they're very capable. But it was the question that SQL Server is is is awesome, but is expensive. Yeah. So but yeah, it's, it's ugly, it's really hard to to have multiple it, especially if you if you have low team. Excellent.

Greg Irwin 58:45

Jeremy, thank you very much, folks, we're at our our was very quick, our that's a good sign. That's always a good sign. So a couple of takeaways. First, let's take advantage of everybody who joined today, we're gonna send around a list of names. And I really sincerely want to encourage everybody to connect across this group, not just with ADB, you know, not just with Cintra, but across this group, and basically create your own networks, your own contacts, and it's some outstanding people here and I encourage, look, it takes a village. So I truly believe that. Secondarily, a sincere thanks to Jan and Abdul, these guys, we get the luxury of choosing some partners here and and these gentlemen have fantastic leading organizations in the space. So again, treat them as the experts they are if you have problems issues or, or suggestions, let's please take some time to speak with them in their organizations and how perhaps they can help you with some of the initiatives you have. And they'll reach out and if there's opportunities, great if it's one of your colleagues would be better suited than any of those intros would be Very much appreciated. Lastly, let's spend the wheel and MC who's going to have a you know, who get the what do we call it the this year? So, let's see. Can everybody see this wheel?

Yes, you all

can see it. Okay, I hope I have everybody's name and so I'll take a second is anybody's name missing here? I tried to do it, but I was hard to multitask so but now look, I didn't include I didn't include EDB and Sentra, so sorry about the or BWG but is anybody missing from

the We'll speak now or forever? Hold your peace. Okay, let's let's roll it. Things always fun. No, it's

always a good meeting when you can spin a wheel. Giovanni, congratulations. So Giovanni will reach out to you. And and you can tell us which charity you'd like the donation to go to. With that. Alright folks, let's, let's wrap up our session. Thank you all for taking time and joining us

Roger 1:01:09

speak for the group. Greg, when we all want to thank you for putting us on. Hey, Roger, thank

Greg Irwin 1:01:13

Thank you. Thank you. Great, great discussion, everybody.