 Hi, this is your simple Bhartiya and welcome to another episode of tfo. Let's talk and today we have with us once again Then Meghupal co-founder and CEO of Hasura Thomas great to have you on the show again. Thanks a lot. Thanks for having me And today we are going to talk about a lot of things of course as a rock on territory some announcements and future of data APIs, but before we go there I'm trying to be a bit greedy here. We also ran a whole series on data for June so I would like to just quickly talk a bit about that because I would love to use, you know, your inputs there is that If you look at from the traditional IT and the whole cloud native Because of Kubernetes centric word, how we have seen the evolution of data. That's a great question I think we're at a tipping point in the industry in the way that we starting to think about data, right? So What happened over the last year was that we realized that we need to get really productive with kind of developing applications of the cloud and One of the big challenges with the developing applications of the cloud was that we needed to have You always grow out of the use case of being able to use just one data source and one database, right? And you always need to have more and more databases That you want to build So that you can kind of scale out right for all kinds of things Maybe your workloads change the amount of data that you have changes and stuff like that. I think What we realize is an industry now is that weird this The route that we went down where we have so many microservices and each microservice is kind of working with each data store is Becoming cumbersome is getting velocity people are getting very frustrated with the total amount of complexity that is increased in the system And so what people want to do essentially is be able to use more data sources now What has happened also over the last few years is that the explosion in the In the capabilities that all the database vendors are offering in the variety of data sources that we have has also exploded Right, so in the last five years There have been about a thousand database and database adjacent startups that have been launched just in the last five years Which is you know, which is insane? So this what what the world is kind of not transitioning to is the idea of saying can we make our database vendors do more work? As long as they have open standards, and I think that's the Evolution that we are seeing now we kind of want to go more from a service first world Into more of a data first world and I think we're at the tipping point where that transition We're starting to see I think especially with the stuff that's happening with AI that value of data and being able To use that data ASAP right take data out from one place use it to do something else integrated with product It's all becoming even more important, right? So I think like we're going to see that transition from being Service first into data first and of course all of the innovation that has happened in the industry is going to help that so far Right all the cloud native work that we've done all the Kubernetes Innovation that has happened. It's all going to help as you're talking about this evolution as we have seen in the early days of the whole cloud movement DevOps movement was there. We have seen DevSecOps movement there. NetOps is there AIOps is there Things are we are talking about shift let movement with a more of a security But what we're talking basically about cultural and people when it comes to data How much do you see we also need the culture should because when we look at data once again in the traditional world It was a silo of folks who knew about data. We still talk about data scientists You know data engineers is not and it is something folks get intimidated and discussion about data doesn't become very interesting topics Of course, generally AI is making things interesting How much you are seeing the importance of cultural shift is needed within organizations. So when we look at the data first word culturally they are Not looking at as a silo or hey, we just got a solution and we're done. We're seeing a massive cultural kind of shift That is being driven by I think the overall macro situation as well, right? So let's let's let's kind of just to break that under some specifics right the first thing is When we talk about data first when we talk about the data first world we often and we think about data scientists and data engineers We're often thinking about OLAP and VI type data But there's also a lot of transactional data that needs to be put to use right There's a lot of like there's it's a spectrum of data that you have that is real time that is streaming That is analytical that is being used for bi now. It's all becoming unified We are no longer living in a world where there's this pure separation between transactional data and analytical data We want to merge those right people are our users are getting wiser They're expecting a product to be integrated automatically with Analytical information to be enriched to be enriched with AI but also to be enriched with just analytical information So a this whole It's it's kind of becoming a continuous spectrum. So this is happening and being driven from users naturally now Now the the the shift the cultural shift that needs to happen Is because ultimately what is happening is that and I think it's mostly driven by the macro All the CEOs in the world right all companies in the world. They all have a mandate saying we have to move faster we have to be more efficient and Um, we all need to get more done with what we have today, right? And that is the biggest need of the market, right? We need to innovate We need to make progress. We need to add end user value We need to create business value and we need to do it With lesser resources. We need to do it more efficiently. Um That was kind of the macro situation that we've been in the last for the last year or so That is still very much the case and with AI it's even more pressing Like we have to get to adding value as fast as possible. Whoever is not adding value is going to suffer Now in this world when you think about being able to go fast We are burdened with too much complexity In our technology stack in our product development stack There is too much process that we are trying to fight against In companies of all sizes, right? Like you said, there's DevOps. There's AI ops. There's There's product ops. There's data ops. It's it's a lot It's a lot. There's too many people. There's too many processes. There's too many ops becoming really hard to think through How to ultimately add value and leverage all this information that we have, right? And so where I see this going in in kind of helping address that complexity Is the culture shift that is required of people, right? I think is going to go in like two directions The first culture shift is that everybody is going to have to become more product and end user focused product product thinking product engineering is going to be more valuable than any other form of kind of engineering Or any other kind of mindset in the organization, right? The the product thinkers in the organization and people who are as close to the end user as possible They are the most valuable people in the organization Empowering them is the way to kind of Deal with all of this chaos, right? Just to say look, I don't as a CEO. I don't understand the complexity But here are the people they are the closest to my end users. They're the closest to my customers. They know what to do Work backwards from them and simplify everything Right that that is kind of one one anchor point that helps us think through how to reduce chaos The second piece here to think through it from the data point of view is that Let us start to leverage Best of breed polyglot data as much as possible We know that we can build really good products if we use the best data sources that are available Right, if I can if I can make my data migrate faster move faster If I can de-risk the adoption of new data Sources and new data vendors then I can build a product faster, right instead of saying Can I scale my legacy system to handle this new kind of workload? No Alongside my legacy system. Can I now use the best of breed data system also for this new workload? Yes You want to move towards that world? Because you know that a big part of product development can be enabled if you choose the best of breed beta Right a best of breed data source best of breed database vendor, you know on whatever side it is Now when you think about that I think the skill that becomes more important from a culture shift point of view and a technical skill point of view People who are from an engineering standpoint familiar With first principles of how data sources and data systems work Right how streaming data works how analytical data works how transaction data works how databases work what the trade-offs are Those people are going to kind of become more important again in the organization because they can unlock value of data They can tell you the best practices of okay. You've got a good amount of data and the relational database We need to move it into the analytical database to do stuff We need to integrate a vector database with our relational database or with our search database so that we can do Genai stuff with it. This is the concurrency we need. This is the latency that we need so you need people who are more familiar with data systems and people you can afford to be You can afford to have lesser expertise in building microservices and in having microservice expertise So I think the microservice expertise is going to shift towards becoming more data expertise and less about building and scaling microservices Because again, like you just said in the beginning You have kubernetes now you have containers now you're serverless now So that expertise around scaling microservices less important because the infrastructure is handling it for us But that data expertise and domain understanding that's critical So those are I think the two big cultureships like product and data Those are the most important things that need to happen to unlock the Velocity and efficiency that we need am I making sense? No, it does make sense and thanks for you know once again explaining till Before we jump to the other topic one more thing that I want to ask before So to conclude this thread is that when we look at once again, if you look at the kubernetes work CSA cloud native word or docker container work in the initial it was all about innovation that was happening around it About adoption and moving things into production And then we started talking about security docker security became a big topic kubernetes security also became a big topic in two or three Years ago if you look at the data, uh, because data I sell once again It's much more sensitive than a lot of other things Data production is important. Of course integrity is important Um, the restore and backup is important high availability is important. So talk about that aspect also that you know, uh, Where are these discussions when we talk about data? And of course after that we'll talk about hasura and hasura con absolutely No, I mean and it ties up really well into uh into kind of the recent work that we've been doing as well but but this is exactly I think um from a industry standpoint a product standpoint uh process standpoint The key thing that we need to solve right so like we just discussed We want to move as quickly as possible And we want to move as efficiently as possible right to do that We know that product thinking and product people need to be enabled because they know what to do and how to add value And to do that we also discussed that um, you need to unlock the value of data that you have or the new data that you want to create Now in this we talked about the drivers, right? What you're kind of talking about is what can we not compromise on? When we want to move fast when we want to be efficient, right? We will we have to we have to change a lot of culture and process in the way that we work When we drive that change What are we not going to compromise on right? So two things that any business cannot compromise on right? reliability And security right these are the two lynch pins You cannot compromise on these two because then you break trust with your users with your customers And you are dead as a business right so reliability and security are the two things that cannot change now And that is kind of what is happening when we talk about shift left or whatever we Whether it's container security api security data security, right? It's all essentially around saying If I want to enable the product people How can I enable them in a way that does not compromise the reliability of like my offering? in a data access World that is often about concurrency and latency right concurrency latency uptime These are the three things when you think about reliability reliability of your offering concurrency Latency and uptime right and then you work backwards from that right for uptime for example Like you said high availability is required or disaster recovery is required right for latency Edge is required again depending on what kind of what your user base is what your user base looks like right for highly concurrent workloads Again depends on the nature of what where that high concurrent usage is coming from but whatever the the end of it You care about reliability and and and key enabler is security If you can handle security and authorization. Well, it's a massive boost to the entire business Now from a data point of view data security and compliance is critical And this is again why you want to change the culture of your organization to be more around data It's so much easier when you think about this nightmare that we have with api security Right, it's so much easier to say I have an account I have an account model in my enterprise right. I have something called an account This account has security policies account can be accessed by the owner account can be accessed by the family member Account can be accessed only these fields can be accessed by the regional manager Right so many stakeholders in your organization can access this this this this domain model called account based on certain context So you want to have domain model based governance What you if you are able to enforce this authorization and security you don't care about api security anymore You don't have to enforce security on every single endpoint every single product feature Because if you bring the shift left all the way to data Right at a domain model level if you're able to guarantee security and authorization at that data level It makes life so much easier right now for a lot of people can go and build as many features as they want They can update very quickly. They can take stuff to market They can validate their hypotheses and they can do that because they are not burdened Saying that in every sprint I have to get a security review Right normally this is what happens right every single sprint you're building a feature you need to do a security review If you if you remove that security concern and you bring that security concern on the data side We have we we are able to kind of unlock a massive amount of productivity right For the product people and we're able to de-risk our Our data story right and inside the organization So that's the way that I think about the importance of kind of some of these other These other linchpins are not reliable in security that we think about. Thank you so much Now I want to shift gear, but not totally changing them is that you folks recently concluded your fourth annual Sura conference talk a bit about the conference also talk a bit about The the the theme that we discussed today security and other aspect. How much that was there at the conference this was a fourth annual conference we do this for for our users and for our developer kind of community and And one of the things that we announced or we kind of previewed Will be releasing the product in august is what we call the data delivery network And so the hasura ddn is essentially like a cdn like a content delivery network, but Bringing that value bringing that infrastructure layer to real-time data transactional data streaming data analytical data and now vector data, right? so With the ddn what we want to do is we want to provide a Infrastructure layer that provides reliability and security guarantees on data So that the this is becomes the edge and the layer That that the product people use to interact with the data people, right? Your data teams your federated data teams can connect their data sources to our ddn to our data delivery network And product people who are building products, you know in whatever context internal products external products can Consume that data from the ddn, right? And the ddn is what is guaranteeing, you know high concurrency low latency Massive amount of uptime. It's not a single point of failure. Just like a cdn. We're distributed with multi cloud multi region We're going to launch with hundred a hundred locations all over the globe so That's the infrastructure layer that we announced And you know, like we were chatting over the last few minutes top of mind for us is How can we enable more data? How can we guarantee reliability? How can we guarantee security and all of this we are doing So that ultimately product people and product builders people you're building api as you're building apps. We are enabled, right? so Entire team of the conference was essentially around, you know our ddn thinking about what are the various features that we're building to solve for security and authorization to solve for Federation being able for multiple data teams kind of coordinating together You know stuff like that on the AI front. I think one of the One of the key things that we're doing that kind of becomes Very natural when you think about from a ddn a data delivery network point of view, right? An enterprise is connecting all of their data into the ddn And product builders are able to access that one of the api is that we provide to access this data We provide a graph dual api rest api What we're also providing is now a knowledge api and that's what we're working on Right now with our community the knowledge api essentially allows people who are accessing data to be able to Ask questions on that internal data that they have Which is a combination of structured analysis and unstructured analysis, right? So you can go to your kind of for example e-commerce inventory data and ask a very specific question and say, you know What pillows under 20 dollars are good for people with small necks? And what the ddn does is that it automatically converts that into a bunch of structured queries That will go and filter products that are under 20 dollars and then go vectorize that and Integrate with an llm or large language model api, you know, say gpt or answer pick or bar and say and and result with saying These this is the meaning of you know, small necks. This is the meaning of a best pillow This is what this is the summarization of information that we're seeing in our database in our unstructured data And so that I think is going to be extremely exciting the ability to kind of connect our internal private data with kind of global human knowledge that is embedded in these large llms and And kind of use that to provide a knowledge api on data, right? And that's I think the final frontier of data that we're all going to be so excited about In the in the industry that conversion of data to knowledge And so that's going to be you know, our our kind of pioneering move on that side Then Matt, thank you so much for joining me today And of course talk about the evolution of data of course the event and your focus and of course You're not announcing the words that you folks are doing So I will look forward to talk to you again in august, you know, when you know You folks are ready with the announcement, but I really appreciate your time today. Thank you. Thanks