 Wow, it's so I mean, it's Wednesday and and if you don't know today's St. Patrick's Day and I feel like I have a little connection to this whole holiday being the Irish guy in this cut in this call So you have no idea how you double down today. Not only are you gonna get some really interesting information? I think I hope but you're also talking to St. Patrick. So Congratulations, you don't even have to wear green. It's a good day. All right. Maybe not. I'm not selling this am I all right? and I also want to clarify too so This is this is a gonna be a lot of information that I think is is new for folks And I'm really excited that we're gonna go down this route together Ask questions. I won't be able to see the questions while I'm giving this because I'm in full-screen mode for the prezo But go ahead and ask the questions. I'll be pulling a time to talk about it afterwards. Let's dig in. Let's go into it I love the Q&A portion almost more than I like doing this. So Like Marissa said, my name is Patrick McFadden. I do develop a relation stuff here at data stacks I also work quite a bit in our open source side. I do open source strategy Spend most of the time giving away software. That's what I do and I work a lot with open source projects I work with CNCF and I work for the Apache Software Foundation quite a bit. So this is just my world I live in but I'm also An Oracle DBA Yeah, for a long long time and I've been building data-driven apps for and I said 19 I'm not gonna tell you when a long time ago But you know, and I was just I was certified as an Oracle DBA because that's how you made money back in the day and You know that this is part of my journey of working in the application space and When we started when I started doing this it was generally the database was the machine over here and then you had some other machine that was running your application logic over here and you could actually see them and touch them and that may sound crazy today, but It meant that You had to really take care of all your infrastructure. So I remember back in like 2000 during the dot-com Every pitch because I was involved in that dot-com craziness But every company that started the first slide that you showed a VC is how you're gonna build your data center and because it was such a critical piece of the puzzle and Along that line, I wrote a really good book You should use a database This is a fake book. By the way, I don't think it's real, but I Think this is this is the most obvious thing that everyone says it back in day I mean, I can't remember applications that were written that didn't need a database and Especially it was one especially in the dot-com era when we were like had tens and hundreds of users We needed a database, but every application is data-driven and Because of that the database is such a critical piece of that That this is what I'm going to tell you today is stop using databases because they're too critical they've gotten way beyond what they need to be for you to be a successful developer or To be an SRE that develops a good application. I'm sorry that deploys applications into infrastructure Just stop using databases That's the end of my talk. I'll take any questions I'm kidding. I'm gonna prove this to you So this is like, you know a good range of applications, right? Is you know, you have these web and mobile you have microservices out there IOT all of these application types require the same thing a database and If you've been building any kind of applications, this is probably similar This looks similar to the stack that you're building, you know If it's if it's microservices, you're probably just distributing containers all over the place If you're doing web and mobile you have like your API gateway with what's happening on your own You know using like react.js or I mean sorry react or some sort of like Swift something like that on the mobile but it's always interacting eventually with the database and What IOT is pretty much the same, but it's very very bespoke right IOT is pouring Tons of data into you a database time series data into a database and in every case That database is kind of the center of the universe for everything. I don't know if anyone has seen this But someone at one point says data is the new oil What is it like slippery and expensive? Yes, it's that So when we when we talk about data is the new oil We want to make sure that you know It's it's not this really hard thing to pull out and we built that congratulations everybody We have Okay, did my slides advanced? Oh boy Hmm. I think we're are we falling behind here on my slides. You should be seeing an eye of sore on at this point mirth Yeah, I'm only seeing that hello. Here's my background slide Really? Hold on I'm gonna I'm gonna kick it See, this is why we we're doing it live Fine. I'm gonna reshare resharing here goes See, I knew that we were gonna be tempting the gods. Can you see that now? Yes Okay I'm gonna try again. I'm gonna hit present. Here we go. Okay. So here's my background Okay, everyone got that right and then you didn't can you see my book now? Yes, oh, yes, oh See We're all just traumatized by zoom at this point. We're not even can I ask questions? They're like, yeah zoom just did this to me. All right. Thanks everyone for putting up with me Now the pictures of things that I was talking about I'll just really catch us up here real quick So, yeah, here's my fake book about using a database and I thought that was kind of funny and without the visual cue It's not even funny anymore. I'm gonna I'm not gonna cry St. Patrick's Day. Let's go get beer Okay, so I'm telling you to stop using databases. This is the thing. I was talking about like everything that you do has one of these right and Then this is the slide that I was hoping you would see is and Marissa, do you see a really? Really bad Like I have so run slide. I sure do. Okay. Good. See I'm just gonna keep checking on this I'm gonna do check sums every like five slides. Okay. All right, so This is the world we live in right when the database becomes like this. It's tyrannical It's just like this is the center of the universe and we have to like go ask it Please can I have data? I wrote this perfectly in a perfect SQL statement and it's like hmm. I don't know and The issues that we've always run into is that Developers especially have to really know a lot about databases to use them to be effective and more importantly to get the most out of it If you don't if you're using something like my sequel it's very different from implementation from say Oracle and and if you're using No sequel database like Cassandra also very different and it requires it puts a lot of knowledge and This this thing where you have to go pray to the altar of the database to get the best Performance and the best use of that database I've been working on this for so long. I Forget sometimes that you know, I know how to optimize an SQL query for Oracle because I know how to optimize like table scans and row scans and Cassandra same thing. I know how to build partition keys and clustering columns that are proper that will give you submittal second Performance and works well in a huge thousand node cluster That's a lot of to expect from somebody who's starting or Let's say the first day that they're there and then so what we've done is we've created this whole the order of the sacred DBAs That we all have to like ask very politely and I was one of those people This is a group of people that whenever things are not going quite the way you expected them to That you call and say hey oracle DBA my query is slow and I I When I'm in the middle of doing like work as an oracle DBA I'm in this mode where hey I'm gonna listen to your question, but I don't think Sorry, just a minute The when I'm in the mode of being an oracle DBA I'm gonna be listening to questions and I'm gonna try to answer them Well in that mode I I'm in the I'm just gonna be like well I'm gonna help you the best I can but I'm gonna be hoping that someone's gonna know what they're doing because that's a lot of Questions I'm gonna get and you know this if you're a DBA. It's like hey, how come I'm you know my My query is two milliseconds slower than the last one. Well, I don't know Let me go check the index or things like that I mean it's just a lot of deep knowledge and in years when you start learning things like explain plan, right? And then this is I think this is where we we land with using databases is we have these design time trade-offs Design time trade-offs application and you know, we're gonna be building our application and it's this meeting You know where you're on the whiteboard and you're saying okay. Look we have this thing. What are we gonna build and In that meeting you have to make trade-off discussions saying What are my what are the needs of the application? What are the scaling requirements? How do I deploy it and each one of those comes with a trade-off and the trade-offs usually are your best guess and Sometimes they're not the best best guess because whenever you deploy it you learn that oh wait The scaling requirements are completely different or something else so this is a really hard and probably most stressful part of building your application like in a design phase is just making the right choices and Luckily in 2021 we have better choices that give us less of those problems And then here's this It's funny because we're doing this with a Linux foundation and CNCF. Of course we love cloud, but y'all love some cloud and That has changed so much of how we do things now and running databases now I think it from from a standpoint of Cloud and database not a whole lot has changed there has been some changes But you're still probably doing like a bare metal-ish type of deployment You're still under trying to understand a database, but that is changing that is hopefully going to change and we're getting to get to that in a minute, but The fact that you're deploying in cloud and you're probably not Racking up hardware anywhere because I can't even remember the last time I did that It gives us a whole lot of new opportunities, right and then Kubernetes come on now that is it's done Don't ask questions. Of course, it's taken over and for good reason The thing that I have really started to love about where Kubernetes lands in the in the grand scheme of building applications is You know, we've we've slowly abstracted layers over the years where we went from bare metal to virtual machines to containers We went from you know from bare metal that we install ourselves to instance types then eventually to Running containers in the cloud, but now Kubernetes and and all of those you had to kind of run yourself Now Kubernetes is abstracting the next layer, which is abstracting entire data centers You know Kubernetes is a data center abstraction is virtualized data center so I deployed an entire data center worth of an application and It is really amazing how that is the timing is perfect Because it keeps us from getting too locked into one thing like we want to move from on-prem But that means that we need to really understand the entire picture of deploying an application with something like Kubernetes because we're building the entire Virtual data center and it needs to be able to hold up to our application needs So this is where we get to the data services So I'm saying stop using databases, but use data services and this is a fundamental shift in your thinking in your deployment everything And I'm gonna go into why I don't think this will be too much of a shock but this is a this is something that I feel like we need to as a Community of application builders and database professionals. We need to start thinking about things in terms of data services Just like how we went from deploying containers on our own data centers to deploying in Kubernetes So a data service is pretty simple and there have been versions of this in the past This isn't too crazy, but this is we're getting into more a defined space now Data services are an abstraction layer The underlying the top line is the API's the developers need probably some sort of HTTP API like REST, GraphQL grpc the SQL cql document gremlin, you know, those are protocols that do work And can work on your database. However, the data service really wants to abstract the layer but he wants to hide the implementation from the underneath and This is what I think if you look at like the left side This is probably what you want to use mostly anyway because you're not thinking when you're using REST You're not thinking wow am I using You know am I using the right keys here for an optimized query? Well, that keeps our developers from having to overthink about what's going on below the line and the below the line It's the actual database itself And yes, we're still using someone's still using a database, but the data service is abstracting that layer away And you think about it in terms of how we deploy it. That's where it gets interesting So so there's some great examples of these out there this is there's So there's a lot of these, you know out there already and Hang on So whenever we have I'm sorry, there's a guy something happened outside. So Accio is a really interesting project by Facebook. It's a data service that has some really cool features like it abstracts is Geographics and things like that At no point does the Engineers that use it have any idea what database they're using or where it is and that's the point they want to bring in Engineers that are Gonna be ready to use or to build the front end part of Facebook and Instagram on the first day they don't want to have them to go through multiple days of Learning or multiple months of learning all the back-end data systems That's just not important and Accio takes away a lot of the really interesting things that That are great for keeping your application online But developers shouldn't have to worry about like data placement and a replication things like that So in it when we're using a database You have some things you can do here like for instance You're something you have to do like you have to make sure that the drivers installed that you have to Initialize that connection to the database you have to create a prepared statement Then you execute the statement and then you finally parse and use the result This is a blog diagram, but this is pretty much every database that you use There are some you know, there's some details in there, of course But I mean that's mainly what we have to do every time we want to use a database It doesn't matter if you're using Node.js or Java or whatever And I'm gonna check are my are my slides advancing Yes, they are okay, so you have to do that check some So When you use a database Is this is the way it should work, right? Well, what happens when using data services when we cut out a lot of that stuff from here, we're just executing a statement and parsing and using the results and What this eliminates is a lot of stuff a lot of the ceremony and we like to call it that but it's the Am I using the correct driver for the correct database? am I managing my connections properly and If I'm not what does that do to the database? Yeah, you know if you work a lot with like Oracle keeping open database connections forever is really a bad idea and so Pooling connection pooling that sort of things it puts a lot of burden inside of your application code and then eventually for application developers because all that gets You can you can hide it in implementation, but eventually, you know there's gonna have to be dealt with and When you're using data services over HTTP APIs, you're you're just providing an endpoint and those data services provide enough get-and-put operations so that developers on the front end can just keep moving and That's what we really want. That's the dream, right? and This is a great slide to show first of all danger iceberg, but it's also Shows the implementation like this is what we want. We want this top line to look. Oh, it's just this cute little iceberg out here It's very easy to use but that messy big mess details underneath is hidden and So there's a there's some stuff happening it I can't name there's a lot of these projects out there that I can't talk about I will talk about one that I can but because in the reason I can't talk about them because they haven't quite gone open source yet and but Think of all the hyperscalers the big companies that are doing scale operations They're all moving to data services and staying away from databases and mainly because of the messy details Why not have a small core group of people that manage database operations? Like what if I want to migrate from one database to another? Or I want to migrate from one data center to another You know things like that or I need to do some operations on my database that generally require me to have a Negotiation with developers. No, just abstract the whole thing as long as the API is online. Everything's good It gives you the ability to change things underneath All right. There we go oops so Here are the three things that I would consider criteria for data services oops back back You need to have on-demand scaling You need to be elastic meaning you just pay for what you need. So it goes up. It goes down and you need near 100% up time These things give you that maximum flexibility with a minimum amount of trade-offs and when you think about when you're building applications and You're having You're having these discussions built in this application build time These are three things that you would really love to say these are not at the problem. Yeah, are we gonna have enough? Yeah, we'll get whatever we want when we need it. Is it gonna be online no matter where we are in the world Yeah, should be up. No problem. That's what it's built to do And then finally is it gonna blow out our budget? Nope It's not gonna do that because if we don't use it It's gonna scale back down and we're not gonna have to pay for it great Okay, moving on to the other thing spaces or tabs I'll let you figure that out for yourself But we have less of a Conversation around just data and man, I'll tell you this could really make your life better now open source is coming to the rescue on this one because cloud databases Have started to going down this route and we know that you know, there are cloud Surrealist databases out there or data services that basically are there to lock you into their service and It's pretty well known, you know, they they want you in their wall garden so if you're in cloud X and they have a Very bespoke data service that works inside of their cloud. They know if you move your data into it You're probably gonna be there for life, right? Because you know how it goes. Oh, someday We'll move our database to this other database and we'll be fine. It's no lock-in. Yeah That's called technical debt and you're never gonna get out of that so This is what's been really fun to watch with kubernetes is because we're virtualizing data centers and we can do this kind of stuff like, you know Create the entire application including the data layer. We can move this thing around So if I you know if I want to running GCP Amazon and Azure at the same time Totally fine. If I want to just run run at a time totally fine But it the data portion is the key to portability And open source to the rescue Open source helped us a long time ago. I remember when I was back in the 90s remember that I Was working with we had operating systems. We had to pay for a lot. I mean, it was Solaris was not cheap But you know here comes Linux and it changed our scale equation like instead of installing one big server We started installing tons of little ones and they were great and because we were using Linux open source change the economics of it same with all sorts of infrastructure Databases my sequel Cassandra Analytics like spark they've changed the the the whole game when it comes to economics of scale and Same thing's gonna happen with data services now I work like I said with kubernetes quite a bit now and This portability is really important, especially when it comes to building applications developers should be able to use their laptop You could run it, you know, if you still have on-prem software hardware great if you want to run it in the cloud great All of those things Should work just fine and not have to be really super different like you should not have to say well I'm gonna deploy on my laptop. It's gonna be completely different from what I'm gonna deploy in production Using kubernetes now we have the ability to build not only what we need but put it into a CICD pipeline So that the configuration is stored and get our virtual data centers run in any place that we can run kubernetes. So This this equation has changed quite a bit So I work on a couple of open source projects and I'm gonna kind of get into some like more specific examples and this is why I'm really one of the reasons I'm really passionate about it is because as someone who's been working with With open source databases and like Cassandra My sequel for years I'm seeing like this is like the next wave is where we now take the data layer data services into kubernetes and We just use all that good stuff that we've built over the years. I don't think we need a new database Not at all There's lots and great databases out there that do lots of great things, but What I do think is we need a new relationship with our databases in a cloud native architecture so I'm gonna go over some of these Real quickly and just give you a highlight of what they do and why we're you know What these open source projects are about why I think this is the right direction? So Stargate is a not only a cool name for a project But it also does things like Create HTTP API is like REST graph to L Document API's and it even does native format just like I was showing you before in another in another site but It it is built for deploying on kubernetes now This gives you this a lot of options and for developers it gives you lots of options because you can just choose the right API or client or framework and It abstracts away a lot of stuff and it's it's not easy because sometimes you have to think about What's underlying you know There is some thought when you deploy it about how you deploy it with your software or with your underlying database But you know, this is left up to the SREs and if we can provide that as or as SREs Our developers just are fine. They're gonna be just fine Because they're gonna want to use something like no JS and a REST API or graph QL and not think about oh Do I need to download a driver? That's don't need to do that. So It really gives us the ability to separate out our Traffic from our data traffic from the actual database itself and they can scale independently and Then as new API start popping up we can add those and start building out this whole infrastructure the architecture right now for what Stargate does is it has a the first layer is it's basically just it accepts HTTP calls from a variety of places and Then negotiates with the underlying databases right now it uses Cassandra because that was the first implementation there are other databases that are in In the pipeline as they say There's other open-source database projects that will that will participate in Stargate and make those happen And that's the point of this is making it just so that we could have a Meeting place for databases to present themselves as a data service And if you think about it in terms of like this is very modular just like how Kubernetes is very modular We can build out all the different parts in those parts Or how we deploy things into our infrastructure just like as we build storage network and Storage network and compute in Kubernetes. That's how we consume things. We can do the same things with our data layer the other project that I've been pretty heavily involved with is a project called Kate Sandra and This is a little more into the database side for sure, but what we're trying to do is take Cassandra and completely eliminate it as a an operations problem and make it easier for people to deploy it into Kubernetes, but Ultimately with a goal of making it so that it's just a data layer for code Kubernetes You're not really thinking about what the database is Yeah, I know the word Kate Sandra is a little bit of a mashup and it's cute, but You should be able to deploy it and have The three things that I talked about you should be able to have an always-on Elastic scaling high scaling data service without a lot of work It's it should be an SRE topic that's pretty simple and your developers should get what they need out of it so the Things that are built on we use an operator internally for Cassandra We we use Stargate so it implements Stargate to provide a data gateway. So when you deploy it it has It has all the services that you your developers need and then it uses It works with traffic in this case, but it works with others as well for Kubernetes ingress and then it has operations tools that run automatically in the background for all the operations tasks that normally you would have to do manually It does those automatically and this is When we put all this together with metrics and observability You get a pretty usable package that you deploy pretty easily you just deploy it Developers can use it in any way they need to you can deploy it anywhere And this is where open source really look at every one of these things as open source So if you want to deploy it on Amazon you want to deploy in Google you want to deploy it on your laptop totally cool and It's portable and it's you know We have an open source of freeze and freedom and freeze and beer This is free isn't freedom. You don't have to be tied up into a certain walled garden you can You can move it around however you need to and because these projects are all open source You can also do a lot of manipulation and changes It's one of the things with Kate Sandra. That's really exciting is It's a lot less about code Contributions from the community and if you look at the PRs It's a lot. It's it's a lot less about code and a whole lot more about configuration Deploying like how you deploy things so it's it's turning into a like a collection area of SRE knowledge around deploying scale applications and Kubernetes And then we use helm eventually for a lot of this stuff to get deployed Helm makes it easy. It's right now. This it's the the first class in the project, but We will I think we'll see soon it. It'll be departing. It'll split away from helm So the the idea here is whenever you're building applications and if you think about like all the things that interact with your data layer it's creating some distractions inside of Kubernetes that allow for that to happen. So if your web and mobile apps microservices, etc. Need a data service This will provide that in Kubernetes and it'll run on any infrastructure because it's just running in Kubernetes The the keys are really just making sure that you know, you have a stable or Kubernetes cluster and Some of the knowledge that we are building now is around things like making good storage choices how to work with ingress that sort of thing So back to my slide So Stargate is kind of the tip of the iceberg in this case where it's just like here's the here's this little nice light Little piece of ice floating on the water nothing to see here all that messy details handled automatically underneath the case Sandra and This is an interesting. This is what's interesting about this is that we're building up projects within projects. We're starting to wrap projects and projects, but Kubernetes is the destination for I think all of our application infrastructure eventually at least in the next 10 years and We're really trying to make this a place where everyone can gather and bring knowledge and share and create a community around Building data services and no matter what underlying database you have no matter what deployment infrastructure you're using So that is all I have and I promised I would leave plenty of room for questions. So I'm gonna do that real quick. All right Got a few I got few questions. I got one really good one and feel free to ask any questions you'd like in here I Let's see. I'm gonna should I it's my screen still sharing right now It is. Yep. All right Do you see my entire screen or do you see I? Don't even know what I'm showing anymore It's the full presentation screen the full PowerPoint screen. Okay. Got it. Okay. Um Um Do I have a series of sample projects based on this deployment? Um, well Kavita you asked a good question. There are a Lot of these parts like this. I like I'm assuming that you're talking about like the key Sandra Stargate Funny enough Data sacks Astra which is our is data sacks. It's a standard as a service Runs completely on this stack So we are we actually run this We have a We have a team and open source team that spends a lot of time on it But it's also being used to deploy our own service. So we not only believe in it. We Got a better company on it There are Yelp just did a really cool presentation lately It was in a meetup in London, but it's on YouTube, of course Yelp did one on how they use Stargate internally. Um, they're slowly migrating their things over to Kate sander as well Um, they're because sander shop Um, there are a few others that have not been public yet, but they are coming soon um For uh, you know, but if you look on the Kate sander website There's there is some discussion about like how like different deployment scenarios and how you would deploy it in your infrastructure as well Um Who are the main competitors to Stargate? uh It is the cloud databases and this is one of the reasons we believe that open source would be a really good move for this particular thing um And the reason being is that uh When you're like I said the like the dynamo db's the cosmos db's and I forget what to want to google a fire store You know they these are really good databases. Um, there's nothing wrong with them, but they are specific to that cloud and You know, it's just like there's nothing wrong with oracle oracle is a great database nothing wrong at all. Um, I made a great career building good stuff on it, but open source changes the economics and the dynamics of how you do your applications and If you're looking at a cloud database and that that can in that space Then yes, those are a competitor to what Stargate would be and you should look them in the same way. Um There's uh other open source projects that um, probably are come along Um, this is just starting to crack open. Um, I think that occhio Facebook is never open source that they just keep talking about it and I don't know if they ever will open source I talked to that team several times They seem pretty happy just developing and internally and talking about it I am okay so Maybe somebody I would love to see it out there Just because I think it's an open source ecosystem really thrives on having a lot of different choices and different different opinions Um, I don't have a demo. Um, I don't generally I should have a demo um It there are there are a lot of workshops around this if you were you know if you hit google look for a stargate Um, Kate sandred, uh, there's some workshops on actually how to do it And that's that's something that our developer relations team does quite a bit of um So information security, uh, good question. Um, and that it's actually one of the reasons that data services I didn't put it as a main reason, but it is definitely on the list of a good reason to have it and so thinking of in terms of like, uh Info sec when it comes to databases is I think one of the it it keeps um, sysos up at night mainly because You can break into a web server an app server You know if a if someone if a bad actor breaks into a web server or an app server The the thread exposure is a lot lower than if they get into your database and You know, that's you hear about like the worst hacks out there It's usually so and so got ahold of all the data in the database and that's just horrible um we uh We we've looked at like how to do some of these things first of all from an Information security abstraction is always the best plan But what it does inside of like for instance with stargate is it allows the implementation to maintain certain security standards So for instance, um inside a kubernetes deployment, you're going to have a certain security stands already You're going to have shared secrets and that sort of thing but um, one of the things that stargate enforces is things like tls And making sure that everything is firewalled and no open ports So, I mean If you look at like some of the best practices of running a database, this is fitting into that realm pretty well So, uh, let's see next question Does kesandra provide a feature for maintenance task? Yes, actually that is um, What there's two projects to patchy projects in there. Um, there's reaper and medusa, which are great greek names um They uh, they are there to do operations on kesandra in the back end. So you don't have to um reaper does uh, does repair maintenance, which is just part of an eventually consistent database and medusa does backups, um and there you can Is just part of your Deployment inside a kubernetes you can you can play it and specify how often you want those to run um You can deploy with backups on and uh destinations for backup files So for instance, if you want to dump them into a an object storage like s3 or something like that um All of those things should run automatically And it will give you the full observability and you know, it pipes all this stuff out to prometheus So you can in grafana so you can look and see exactly what's happening, but mostly it's just set it up run it and it'll run itself um Question was my sequel is open source. And why do we have to move on to stargate or case sandra? Yes, my sequel is an open source database. Um, my My proposition to you is why? And so it's kassandra kassandra is a database, but um my sequel kassandra I mean, there's all of these are open source postgres But what i'm talking about right now is data services In a cloud native way my sequel is a A database that can run in kubernetes But it does not run in the same way that kubernetes wants to run. For instance, if you need a lot more my sequel You you can't just add more nodes. Um, it doesn't do active active replication. So for instance, it's it's it's hard to case stay resilient There are ways to make it more resilient with like read replicas and things like that But mostly the data services and i'll tell you this I wouldn't be surprised if there was a my sequel implementation of stargate soon. I'm just gonna Here's a little preview. So, um stargate plus my sequel might be an option here for you Um, how long does this stack take to set up from scratch on a new project? Well with a helm install, uh, it's actually pretty easy. Um, I I have it running on my local laptop right now It it takes about five or ten minutes to if you're completely scratched to download all the containers You know get all the infrastructure going If you go to katsandra.io, there's a getting started button Um, and it will walk you through the basics. Um, if you're using something like kind on your desktop or mini cube And I think mostly it's just the time required is how much time it takes for the the containers to download Um, and then when you're setting up an actual cluster like inside katsandra cluster, um, depending on the size because we, uh Per node we put about a one to two minute delay on deploying each one just to let everything inside kubernetes settle down So, you know, that that's how long to go um from a from a developer standpoint, um Not very long. Um, we do a two hour workshop on katsandra and stargate and You can go from zero to fully running application within two hours Which and that's with a lot of explanation and blah blah blah and we even play games in our workshops. So Um, so not very long and that's the point is we want this to be a very fast and easy operation And if it isn't here's the thing I beg of you Please participate in the project by giving your opinion. We have a discord Um, there is prs and issues on github Like if it's like, hey, you know what this doesn't work the way I want it to Speak up. We would love to hear about it. That's what open source is about I mean, we really do value that kind of feedback and be a part of the project Um, is there an option for dynamic data masking? No, not yet, but that's interesting because that that fits inside of the security part and This is actually something that I think we'll find its way here pretty soon because this is something that, um Uh SREs especially Are looking for for instance if I have a data service that's pulling customer information And I need to put a rule on it say based on gdpr I should be able to do that at the operation deployment level Based on parameters and not leave it up to the developers because you know what happens whenever we leave it up to developers, right? Yep, that's right problems Um, how many data centers do I have around the world? Uh, I don't know have any data centers. I read everything Um, yeah, this isn't this isn't a uh astro just uses all the clouds So wherever the clouds have data centers, that's where we are. Um, but it's up to you Um, this is an open source project. You can deploy it anywhere you want so Um, but that's an interesting question Um ml support. Yes. Thank you. Great questions. Um, so ml pipeline support is not there yet However, um, there are some api considerations with this. Um I was just in a conversation this week about with somebody about Using this was like a patch of airflow and that sort of thing Um because of where it sits inside of the infrastructure um It it does lend itself to potentially being a part of an ml pipeline Um, and I think a lot of that is implementation like using apache pulsar flink Um airflow. I mean, there's there's a few options in there where you can start seeing how this might fit in a kubernetes deployment. Um the I'd say just stay tuned on this, but this is I this has been a topic that's coming up Especially in the data mesh world. I don't know if you've been involved in data mesh yet, but um kubernetes data mesh these are all topics that are starting to converge and uh Data service is a key part of that all right Any clue on the release of kassandra 4.0? Oh Pedro Yes, soon. Um, I will tell you uh as someone who's just as inside as everyone else, uh the There are only a couple of tickets left in beta 4 and we're I would suspect we're going to see a release candidate here pretty soon um and those of you who don't know about kassandra, um The the whole thing around kassandra 4.0 It's been a long time since we've released a major release and the reason being is because The biggest companies in the world that use kassandra are spending a lot of time to make sure that it is the most stable database on the planet Period because it runs the world's biggest workloads. So these are like the alibaba's the huawei's the apples Yahoo japan they're all really intimately involved with the stability of this database and so um The the project management committee will not release 4.0 until It is completely green on data all the data loss data Governance everything So and I think it's there. We actually have a lot of folks using kassandra 4.0 beta 4 in production right now If you go to the kassandra dot apache dot org, you'll see some of those um, but it's uh I think it's it's um Uh, it'll be soon um because we're just we're running out of things to break and it's the first time I've seen Any database this stable in my career? so That's my long answer, but yeah, thanks for asking And I think we're out of questions other than happy st. Patrick's day make tristam That's not a question is it? Okay, beautiful. Thank you so much patrick for your time today and thank you everyone so much for joining Just as a quick reminder, uh, this webinar was recorded and it will be posted to the linux foundation youtube later today Okay, thanks so much everyone. Bye