 Hello, this is your host, Roland Bhatia, and today we are talking to Christian Luz, creator, founder, and CEO of Create.io. What is Create.io? Create.io specializes in distributed SQL databases that help customers in querying and storing machine data. So, let's go and meet Luz. We started in 2013 and released a product in 2014 on GitHub as an open source product. We're based in San Francisco, Berlin, and in Austria. And our product is a real-time SQL database specifically built for machine data use cases. So machine data means all types of automatic data that is produced from sensor logs, from infrastructure, from variables, GPS devices, and all of that. And the characteristics of machine data means you have fire hoses of data, all kinds of data coming from different kind of sources at very high speed from highly concurrent channels. And the data is complex. It's not just, let's say, a temperature reading. It's actually a topology of a sensor that is in a JSON document that you need to store. And then you have to access this in real-time. So a traditional approach would be you use a relational database and a NoSQL database to solve that issue because you need to query stuff in SQL and you need to scale out the when stable on NoSQL. But that means you have to run at least two, maybe three different databases. And we offer a solution where you can store all of that in one database. And you query that in standard SQL. And that's basically the advantage of CREATE and also compared to cloud-only solutions, CREATE DB runs hybrid. So you run it on any cloud you wish, you run it on prem, but also in the same functionality you even run it on a very small host directly on the edge, even on the mobile edge device. How much of your code is open source? So we offer a full version of the product that is open source under patchy2. And why are we here at DockerCon? So firstly, our database is built on a shared nothing architecture. So it is natively cloud enabled and it's natively perfect to run in a Dockerized environment or Kubernetes because you can deploy a full database of Docker in just a single container. So you basically just need to say Docker run great and you have a distributed database running when you launch more containers that can see each other. It's super simple to deploy and manage. And so we think, as everybody is talking about modernizing traditional apps, we think the database will be just following that. And compared to other solutions where you have different kinds of containers and dependencies and you need to attach volumes and stuff to make a containerized database environment with CREATE you just deploy a simple container. And because you asked on the product differentiation on the enterprise side, so we do offer, as the open source version obviously is free and open, we offer on top an enterprise product that has a different set of features specifically built for machine data use cases. So going beyond the traditional SQL database approach. So for example, you have an MQTT broker built into the database. You have security and access control and let's say production relevant features are in the enterprise edition as well as the possibility to get a 24-7 SLA agreement from us. Can you talk a bit about what kind of customers are using your products? Before we maybe talk about the specific markets we address, the general pain customers have before they choose CREATE is twofold. Very often, they simply run into a scaling problem with their traditional database. This is very often a traditional SQL database. For example, MySQL, Postgres, Microsoft SQL Server, but also Oracle. And while they're with a scale up approach, you may solve some of the scalability issues. At some point you can't anymore. It gets either too expensive or too complex. It breaks at some point. And here basically CREATE is a solution as it allows you to automatically chart the data, fully replicated, highly available. And for the developer it just feels like a Postgres server from how he's programming it. However, it's actually a fully scalable cluster in the back end. And the second category of customers that choose CREATE are the ones I touched before. So you have very often in the IoT and manufacturing use cases you have very big event tables like billions and hundreds of billions of records you collect, but you also have relational data that you need to somehow refer to and enrich the data maybe. And traditionally this is done by the relational stuff in MySQL Postgres thing and the other stuff in NoSQL database, maybe Cassandra or Elastic or Mongo. But this means you have two systems that you have to sync and to have to maintain. And then with CREATE you get rid of that. And this is not only easier to run because it's one database. It also significantly reduces the cost because you have a smaller footprint in the cloud and you can use just any developer to run a cluster. You don't need a specialist for that. And the type of customers using CREATE for that where we see the biggest and strongest traction currently is in IoT and Industry 4.0. Let's say all the digitization projects. Second, cybersecurity. So software providers that offer cloud security solutions. We have a couple of customers there. And then you have this, let's call it mobility use cases. So this ranges from a golf app that NBC runs successfully in the U.S. with 4 million users with every golf player is tracked with his GPS data and provided in a real-time service to fleet management when it's more than just the simple tracking of a time series data when it's basically the requirement to run more complex queries. So we're often replacing also time series databases that allow you to capture data fast and also basically plot a nice time series chart of a particular matrix. But in the moment when you actually have to interactively run queries or thousands of people have to run queries and you have to enrich this data or you have to make joins and complex queries this time series databases are not capable to. So it's basically the next level then. And that's also a big use case for us. I see that a lot of these customers are from the IoT space. So how many of these customers are from the consumer IoT space versus the industrial IoT space? So I would say currently most of our customers are either industrial IoT manufacturing production and stuff and providers of consumer IoT solutions. So we have one customer that provides for end users a back end so smartwatches, sleep trackers, fitness trackers can offer app service and they are doing this as an OEM to those vendors and they are using us as well. So we directly don't have obviously consumer service or operating that space but we provide IoT platform providers with a scalable SQL solution. I recall earlier you mentioned that some bottling companies are also using your solutions. Can you talk a bit about that? That's an interesting use case. Yeah, one of our really exciting customers and we're at DockerCon showing this the first time publicly now is a company named Alpla. They are very big like 18,000 people, 180 factories in 45 countries and they produce basically the plastic bottle. They produce in the US the Coca-Cola plastic bottles and their factories are very dislocated because they have to be exactly where the Coca-Cola is filled into the bottle so the empty bottles have to transport, not to be transported. So they are produced at the bottler site and that's why they have a lot of factories and for them it's a very modern company and they exist for like 30 years and they optimized like every process in this company to its perfection so they're using lean production, Six Sigma, everything has been super optimized but they cannot get any more efficient and they knew that the next step they can take to be more efficient is basically digitalization and so they started a project in the US to collect all their production data from the sensors of these production machines in real time using Docker as a mechanism to deploy all of these applications and then this data is from all these different factories is collected, sent to the cloud to create database and create then powers basically interactive dashboards for a 24-7 team that they have, they call it Mission Control and this Mission Control remotely manages all these production lines and looks where there is a problem and so they can basically adjust the production before it goes out of the tolerance and waste is produced because once plastic is wasted it's wasted and that's a really interesting use case for Docker and for Crate so deploying the database in a Docker container deploying the apps in Docker containers having a quick and simple rollout in all these different factories and at the same time having a real time powered production management system What are the next big challenges that you see as you move forward? For us it's always a challenge and every customer that we can publicly talk about helps of course, it's always a challenge for a new technology especially for a database to get enough traction so why we passed already 2 million downloads we have more than a thousand clusters running globally and more and more can publicize really cool production use cases as the one with Alpleid just mentioned or we also have Sky High in the valley a security company with a more than 100 nodes create cluster this helps to get credibility because the big challenge of course is that there are so many database companies on the market it's very hard for developers and DevOps to really see through the jungle and pick the right database and we think for the next generation of challenges the explosion of IoT data and at the same time to have a compatible, sequel compatible option that's something that will help us and the main challenge is really to bring out the word and have users talk about it and see more and more use cases so more and more people trust in this technology and use it so DockerCon was great for us because like on the first reception evening couple of folks came to the booth and said hey we're using you guys since one year in production and because of this whole open source thing we had no clue about them, never heard about them and we even have a customer called Coltrick a very successful company in the US doing service they deployed us in six data centers and we're running in a significant part of their backend and they just told us when they basically released it and just informed us and that's the beauty of open source but of course for us also a challenge because we need to grab these guys make them write about it make them talk about it because obviously if a developer hears ALPLA, Skyhack, Coltrick is using Crate that's pretty significant and trustworthy Since you're heavily into open source world this is a question that I would like to ask you I recall my early journalism days where we used to educate companies, organizations and people about the benefits of open source that why they should use open source nowadays almost everybody is using open source they've created a new set of problems the biggest problem is that most of these people don't really understand how open source works they use it but they don't know how the whole ecosystem works how you have to contribute back how you should engage with the community how you should enable your developers to go out and interact with the developers so my question to you is have you come across any customers who really don't understand open source and what kind of challenges do you see there and how do you help them solve this problem? Yeah, I think you're absolutely right there is a big change that happened and is happening like it used to be the case that you have these big purchasing departments with the CIO and some sub-heads that take like the big decision how we do this this totally shifted from my point of view so the power is now with very good developers and DevOps experts and they would, because of open source it's available, they just go out there they try other technology they play around, they figure out it works and then they kind of bottom up put it in the organization and still the top guy has to decide of course he's the guy who signs the purchase order or decides technically and strategically on it but he would not go against all these guys who would say hey we think this is the best thing we should do and this is the total shift that happened so the distribution and the sales force kind of approach changed you need a sales force of course because in the end you need to sell it and we also need to convert open source users to paying users to be able to sustain the open source product but it's a great, it works in both directions and for some customers especially bigger companies it's a challenge, you're right they somehow try to make you produce custom code and they create dependencies that you would never propose and you try to talk them out of it but we have projects for example where customers where they ask us basically to produce code that glues their solution with our solution and we agree on a shared source agreement where they use this code but also we can reuse this code without like obviously custom secret pieces but very often it's just glue code and this is then how we can either re-open source it or at least use it with other commercial projects and I think that's a big shift that is happening also for bigger companies they value now, they value this and for us for example as a database company I think we would not exist without open source because I spoke once to the SVP of Engineering of Sky High when they informed us they have chosen Crate and they started I think with a 10 node cluster it's now north in 100 nodes cluster and I thanked him and said how did you really end up using us and he said well you did this trick of no sequel and sequel but I had our team doing a hackathon on your code base so they basically reviewed the code and they gained the trust that this is like serious work and this was more important to them that we were like a big company and they started trusting on this and grew with us obviously over the last two years but I think without open source if this would have been a closed source thing it would have been very hard if not impossible but now these customers know it's work, it works it has a quality to code and even if this company goes down the code stays still there and other people likely take it over or it restarts somehow because I think we reached this critical mass of traction now with millions of downloads and more than a thousand clusters running globally That was a great conversation and I hope that we'll see each other again in some future open source event Thank you very much Don't forget to subscribe to this channel on YouTube and on SoundCloud and don't forget to follow us on Twitter Thanks for watching and see you next time