 We're back at the Javits Center, the CUBE's coverage of MongoDB World 2022. The first live event in three years, pretty amazing. And I'm really excited to have Tony Coleman here as the CTO of Temenos, who changing the finance and banking industry in Boris Bielek is the global head of industry solutions at MongoDB. Welcome back to theCUBE. Welcome first time CUBE alum. So thanks for coming on. Thank you. Thanks for having us. Tony, tell us about Temenos. What are you guys up to disrupting the finance world? So Temenos is everyone's banking platform. So we are a software company. We have over 3000 financial institutions around the world. Marketing tell me that that works out as over 1.2 billion people rely on Temenos software for their banking and financial needs. So 41 of the top 50 banks in the world run our software and we are very proud to be powering all of those entities on their innovation journeys and bringing that digital transformation that we've seen so much of over the past few years and enabling a lot of the world's unbanked through digital banking to become part members of the banking community. So basically you're bringing the software platform to enable that so somebody doesn't have to build it themselves because they'd never get there. Absolutely. I don't know if you consider that disruptive. I guess I do to the industry to a certain extent but when you think of disruption in that business you think of blockchain and crypto and NFT. Is that a completely separate world or do you guys participate in that as well? Well, I would say it's related, right? I mean, I was doing a podcast recently and they had this idea of a buzzword jail where you could choose words to go into jail. And I said NFT. Not because I think NFTs are intrinsically bad but I think just at the moment they are a rife for scam area. I think it's one of those, one of those technologies and investment area that people don't understand it and there's a lot of mistakes that can be made in there. Icebergs ahead. Yeah, I mean it's a fascinating piece, right? It could be truly transformative if we get it right but it's very emerging. So we'll see. So Temenos don't play a huge part in the blockchain industry directly. We work with partners in that space but in terms of digital assets and that sort of thing, yeah, absolutely. So Boris, you have industry solutions in your title. What does that entail? So basically I'm responsible for all the verticals and that includes great partners like Tony and Temenos. We're doing a lot of verticals by now in MongoDB when you listen to Dave in all his various talks. We have so much stuff ranging from banking, retail, retail, house care, insurance, you name it, we have it by now. And that's obviously the clients moving from these edge solution, like touching a little toe in the water about Mongo to going all in, building bigger solutions. You saw on stage the lady from Wells Fargo this morning. These are not second grade. Yeah, we do something small now. We are part of their transformation journey and this is where Tony and I come regularly together how we transform things and how we build a new way of banking is done with microservices and the technology surrounding it. Yeah, but what about performance in this world? What can you tell me about that? Wow, yeah, this is an interesting thing because people are always challenging what is performance and document databases and Tony challenged us actually six weeks before his own show several weeks ago in London. It says, Boris, let's do a benchmark and maybe Tony, you bring your story because if I get too excited, I fall off the chair. Yeah, sure, I mean, performance and efficiency of topics close to my heart have been for years. And so, yeah, every two or three years we had someone else run a high-water, we call it a high-water benchmark and this year we sort of doubled down, literally doubled down on everything we did previously. So this was 200 million accounts, 100 million customers and we were thrashing through 102,875 friends and the number of transactions a second. Which is a phenomenal number and... How can I do that in the blockchain? Wow, good to say, yeah, exactly, right? So this is, I get asked why we do such high numbers and the reason's very straightforward. If somebody wants 10,000 transactions a second and we're seeing banks now that need that sort of throughput, if I can give them a benchmark report, this is 100,000, I don't need to keep doing benchmarks at 10. Yeah, okay, tell me more about the, any time you get into benchmarks, you want to understand the configuration, the workload, tell me more about that. So we have a pretty well-trodden path of a standard transaction mix. We call it our retail transaction mix and so it tries to, you know, the workload is there to mimic, because it's a simulation, right, around what you would do on your daily basis. You're going to make payments, you're going to check your balances, you're going to see what money's moved on your account. So we do all of that and we run it through a proper production-gade environment and we feel this is really important. This isn't something we do in the lab. You couldn't go live on. This is all, all of the horrible, non-functional requirements around, yeah, high availability. Security passes, private links, all the things. And one thing is they're doing this for a long time. So this is not like, let's define something new for the Mongo world. No, this is something Tony's doing for literally 10, 15 years by now, right? If only it was only 15 years. But yes. And this is your benchmark, right? Yeah. Temenos, Temenos is your benchmark. Temenos developed the benchmark, okay. Yeah, and so we ran it through and yeah, some fantastic numbers and not just on the sheer sort of top-level numbers, you know, 100,000 transactions a second. The response time out of Mongo was fantastic, you know, one millisecond, which is just brilliant. So it means you get these really efficient numbers and what that helped us do with, you know, some of the other partners that were involved in the benchmark as well. It meant that our throughput per core, which is a really good measure of efficiency, is up to four times better than we ran it three years ago. So in terms of a sustainability piece, which is so important, that that's really a huge improvement. And that's down to, you know, application changes, architect changes, as well as using appropriate technology in the right places. How important were things like the number of cores, the memory sizes, the block sizes, all that stuff? We are very tiny. So this is a part when I talk to people, we have what we call MAT system in the back and people look at me, you run, how many transactions on that one? So to be fair, three quarters were MongoDB, one quarter was something else because we're still porting some components over installed procedures, full disclosure. But when I think 75,000 transactions on a single MAT system, which is 32 cores, I'll think correctly, something like that. This is a tiny machine in the world of banking. So before this was a mainframe and now it's one little instance on AWS. And this is really the amazing part, cost that on an environmental footprint is so, so important. And this is a heavy, right heavy environment? So the way we architect the solution is it follows something called a CQRS pattern, so command, query, responsibility, segregation. So what we do, we do all the commands in an appropriate database for that piece and that was running at about 25,000 transactions a second and then we're streaming the data out of that directly into Mongo. So actually Mongo was doing more than the 75,000 queries a second, which is the read heavy part. It was also ingesting 25,000 transactions a second at the same time. And, okay, and the workload had a high locality, medium locality, it was, I mean, just give us a picture of what that's like. Sorry, so the locality? Yeah, we don't have that, that's not what it is. So explain that, that's not, that's kind of not the mindset for a document database. Exactly, exactly. In a document database, you don't have this hotspotting. The one single field of a table which is suddenly hotspotting, and no, you have literally, and query comes up and we say what goes together, gets stored together, belongs together, comes out together. So the number of IOPS, for example, is much, much smaller than a document system than a historically relational system. So IOPS is not a good indicator necessarily, right? Not anymore. That's why the size is so much reduced. The number of access patterns is smaller, I mean, terminals is highly optimized for XML internally as well, the internal structures. So that goes very close to obviously changing. So traditional benchmark would have a cash in front, would have a high cash hit rate, would also, 199% cash, right? That's a high locality reference, but that's irrelevant. It's gone. There's no caching in the middle anymore. It goes straight against the database, all these things are out, and that's what makes this thing so exciting. And all this thing in a real environment, I think Tony, we really need to stress it, right? It's not a test bet at home. It's a real life environment out into the wild with the right benchmark drivers and load drivers. How did your customers respond? You did this for your recent event? Yeah, so we did it for our user conference, our community forum, which is, as Boris said, a few weeks ago in London. And the reaction was certainly, it was a great reception, of course, but the main thing that people were fascinated about how much more efficient the whole platform is. Explain that. So when we can run, and it's a great number that the team pulled out, which is so having doubled the throughput on the platform from what we did three years ago, we're actually using 20% less infrastructure to give double the performance. You know, at a macro level, that's a phenomenal achievement. And that means that these changes that we make, everything that we're doing benefits all of our customers. So all of the banks, when they take the latest releases, they get these benefits, everything is that much more efficient to everybody benefits from every investment. And this was run in the cloud, is that correct? That's correct. So you're running Atlas? So this was? Atlas, MAT, on AWS, with an AWS, Kinesis, and Lambda processes, and, and, so it was a really reality-driven environment. It's a pure cloud native, all using managed services on AWS and then Atlas for the Mongo piece. It's awesome, I mean, so now, how convenient for Mongo, the timing for MongoDB world, how are you socializing this with your community? Yeah, we are having this afternoon, a session as well, where we talk a little bit more detail about that. And Tony has a session as well tomorrow. So we see a lot of good feedback as well when we bring it up with clients. Obviously, some clients get very specific because these reduction footprint is so huge. When you think a client has eight, nine environments from early development systems to production to emergency standbys, maybe in a different cloud. All these things, what Dave talks about, the different Atlas features, multi-cloud, environmentally, all this stuff comes to play and this is why I'm so excited to work with him. We should bring up as well the other things which are available to ready already with your front-end solutions, with the Infinity services because that's the other part. The modernization, the microservices, which Tony is so politely not mentioning. So there's a lot of cool technology into that one which fits to how MongoDB works in microservices, API first, all these what they call mach factors, microservice, API, cloud native, headless. I think that was the right order now. So all these things are reflected as well by Temenos with their leadership what they achieve now. I think a lot of companies have to play catch up now to what Tony and his team are delivering on the bank. Well, this gets the modernization, right? We really haven't explicitly talked about that with everything you just said, talks to modernization. So you typically in financial services find a lot of relational database, right? The 20-year-old, hardened, et cetera. High availability, give them credit to that but a lot of times you'll see them just shift that into the cloud. You guys chose not to do that. What was that modernization journey look like? So it's a bit of, I'm a firm believer in pragmatism and using, I think I've touched on earlier, the appropriate technologies. Horses for courses. Horses exactly to go as right out of my mouth. And I was talking to one of the investor analysts earlier and the exact same question comes up, right? So if you've got a relational database or you've got a big legacy system and you're stuck in a mainframe or whatever it is and you want to pull that over, when you, it's not just the case of moving the data model from one paradigm to another. You need to look at it holistically and you need to be ambitious. I think the industry has got, you know, quite nervous about some of these transformation projects but in some ways, although it might be counterintuitive, I think being ambitious and being embold is a better way through it. You know, take a view. Look at it holistically, lay out a plan. It is hard. It is hard to do these sorts of transformations but that's what makes it the challenge and that's what makes it fun. Take those bold steps. Look at it holistically. Look at an end state and then work out a practical way you can deliver value to the business and your customers as you deliver on that roadmap. So did you migrate from a traditional RDBMS to Mongo? So yeah, this is a fun conversation. So of course in the late 90s, the kind of the phrase document model hadn't really been coined yet and for some of our workloads, the time we referred to it as a hierarchical model and at that point in time, really, if you wanted to sell to a bank, you needed to be running on Oracle. So we took this data model and we got it running in Oracle and then other relational databases as well. But it actually, under the colors there, it is stored as XML. So there is a project that we're looking at to say, well, okay, taking that model which is in a relational database and of course you build over time, you do rely on some of the features of relational databases and moving that over to something like Mongo isn't, you know, it's not quite as simple as just changing the data model. So there's a few bits and pieces that we need to work through but there is a proof of concept that we're running which is looking really promising and they're spurred on by the amazing results from the benchmark. That could be something that's really interesting. Yeah, and I think 20 years ago you probably wouldn't even thought about it. It was just too risky but today with the modern tools and the cloud and you're talking about microservices and containers, it becomes potentially more feasible. But the other side of it is, you know, it's only relatively recently that Mongo has had transaction support across multi-document, multi-collection transactions. And in banking, as we all know, you know, it's highly regulated. It is all of your worst possible non-functional requirements, security, transactionality, atomicity, you know, the whole, the whole shebang, your worst possible nightmare is Monday morning for us. Yeah, yeah. And I think one part which is exciting about this, Tony is a very good practical example about this large-scale modernization and cutting out code base. By cutting off that layer and going back to the hierarchical internal structures, we are simplifying a lot of the backend components of our, because obviously code translation which was done before it's not needed anymore. And that is as well for me an exciting example to see how long it takes, what it is. So Tony is facing my life experiment, so to speak. Well, of course, you're right on because there used to be those migrations where how many lines of code, how long do I have to freeze it? And that, a lot of times led people to say, well, forget it, because our business is going to shut down for a year. But now we do that. We do that. So I'm working, obviously, besides the work with stimulus, there's a lot of financial clients. And by now it's, my joke is normally, shift and lift, pain and no gain. Because the result of the work is when they move everything to the cloud and it was bad before, it will not be better in the cloud only because it's in somebody else's data center. So these modernization and innovation factors, absolutely critical. And as Tony said, the people get it by now. This shift and lift is over. It is how can I innovate? How can I accelerate my innovation? And that leads very quickly to the document model discussion. Yeah, I think the world practitioners will tell you, if you really want to affect the operational model, have a meaningful impact on your business, you have to really modernize. You can't just lift and shift. Is that fair? Yeah, absolutely. I mean, that's, you know, it's a difference between, you know, hundreds of millions or billions in some cases versus, you know, some nice little hits here or there. And we see as well, a lot of clients by now asking for solutions like the terminal solution and like others where there is not any more discussion about how to move to MongoDB. The question is how fast, how can accelerate? We see the services request, the first one. It's amazing. After the event what we had in London, I've 100 clients calling us. So it's not our sales people calling upon the clients. The clients calling in, I saw it. How do I get started? And that is for me from the vendor perspective, so to speak, an amazing moment. Love to sell for Mongo. There you go. Guys, we've got to go. Thanks so much for sharing your story. Thank you. Love to have you back and see how that goes. That, that migration. Happy to come back. That's a big story of, if and when and happens. Yeah, great. All right, keep it right there, everybody. We'll be right back. This is Dave Vellante for theCUBE. You're watching our live coverage of MongoDB World 2022 from New York City.