 Welcome back to theCUBE's continuing coverage of Snowflake Summit 22, live from Caesar's Forum in Las Vegas. I'm Lisa Martin, my co-host Dave Vellante. We've been here the last day and a half, unpacking a lot of news, a lot of announcements, talking with customers and partners, and we have another great session coming for you next. So we've got a customer and a partner talking tech and data mesh. Please welcome Matesh Shah, the VP in market strategy at Elation. And Ash Naseer, great to have you. Senior Director of Data Engineering at Warner Brothers Discovery. Welcome, guys. Thank you for having me. It's great to be back in person and to be able to really get to see and feel and touch this technology, isn't it? Yeah, it is. Yeah, I mean, two years or so. Yeah, great to feel the energy in the conference center. Yeah. Snowflake was virtual, I think, for two years, and now it's great to kind of see the excitement firsthand, so it's wonderful. The excitement, but also the boom in the number of customers and partners and people attending, they were saying the first, there were the summit in 2019 at about 1,900 attendees and this is around 10,000, so a huge jump in a short time period. Talk a little bit about the Elation-Snowflake partnership and probably some of the acceleration that you guys have been experiencing as a Snowflake partner. Yeah, as a Snowflake partner, I mean, Snowflake is an investor of us in Elation early last year and we've been a partner for longer than that. And good news, we have been awarded the Snowflake Partner of the Year for data governance just earlier this week and that's, in fact, our second year in a row for winning that award, so great news on that front as well. Repeat, congratulations. Repeat, absolutely, and we're going to hope to make it a three-peat as well. And we've also been awarded industry competency badges in five different industries, those being financial services, healthcare, retail technology, and media and telco. Excellent. Okay, I'm going to get into it, data mesh, you guys actually have a data mesh and you've presented at the conference. So take us back to the beginning. Why did you decide that you needed to implement something like data mesh? What was the impetus? Yeah, so when people think of Warner Brothers, you always think of like the movie studio, but we're more than that, right? I mean, you think of HBO, you think of TNT, you think of CNN. We have 30 plus brands in our portfolio and each have their own needs. So the idea of a data mesh really helps us because what we can do is we can federate access across the company so that CNN can work at their own pace. When there's election season, they can ingest their own data and they don't have to bump up against, as an example, HBO if Game of Thrones is going on. So, okay, so the impetus was to serve those lines of business better. Actually, given that you've got these different brands, it was probably easier than most companies because let's say you're a big financial services company and now you have to decide who owns what. What, CNN owns its own data products, HBO, now how do they decide within those different brands how to distribute even further or is it really, how deep have you gone in that decentralization? That's a great question. It's a very close partnership because there are a number of data sets which are used by all the brands. You think about people browsing websites, right? CNN has a website, Warner Brothers has a website. So for us to ingest that data, for each of the brands to ingest that data separately, that means five different ways of doing things and a big environment, right? So that is where our team comes into play. We ingest a lot of the common data sets but like I said, any unique data sets, data sets regarding theatrical as an example, Warner Brothers does it themselves for streaming, HBO Max does it themselves. So we kind of operate in partnership. So you have a centralized data team and also decentralized data teams, right? That's right, yeah. So I love this conversation because that was heresy 10 years ago, five years ago even, because that's inefficient. But I presume you've found that it's actually more productive in terms of the business output. Explain that dynamic. You know, you bring up such a good point. So I consider myself as one of the dinosaurs who started like 20 plus years ago in this industry and back then we were all taught to think of the data warehouse as like a monolithic thing. And the reason for that is the technology wasn't there, the technology didn't catch up. Now 20 years later, the technology is way ahead, right? But like our mindset's still the same because we think of data warehouses and data platforms still as a monolithic thing. But if you really sort of remove that sort of mental barrier, if you will, and if you start thinking about, well, how do I sort of, you know, fabricate everything and make sure that you let folks who are building, who are closest to the customer, who are building their products, let them own that data and have a partnership. The results have been amazing and if we were only sort of doing it as a centralized team, we would not be able to do a tenth of what we do today. So it's that massive scale in our company as well. And I should have clarified, when we talk about data measure, we're talking about the implementing, in-practice Jamakthagani sort of framework, or is this sort of your own sort of terminology? Well, so the interesting part is four years ago we didn't have- It didn't exist. Yeah, it didn't exist, right? And so our principle was very simple, right? When we started out, we said we want to make sure that our brands are able to operate independently with some oversight and guidance from our technology teams, right? That's what we set out to do. We did that with Snowflake by design because Snowflake allows us to, you know, separate those brands into different accounts. So that was done by design and then the magic I think is the Snowflake data sharing, which allows us to sort of bring data in here once and then share it with whoever needs it. So think about HBO Max. On HBO Max, you not only have HBO Max content, but content from CNN, from Cartoon Network, from Warner Brothers, right? All the movies, right? So to see how the Batman movie did in theaters and then on streaming, you don't need, you know, Warner Brothers doesn't need to ingest the same streaming data. HBO Max does it. HBO Max shares it with Warner Brothers. You know, store once, share many times and everyone works at their own pace. So they're building data products. Those data products are discoverable through APIs, I presume, or I guess maybe just, I guess the Snowflake cloud, but very importantly, they're governed. And that's correct where Elation comes in? That's precisely where Elation comes in is where sort of this central, flexible foundation for data governance. You know, you mentioned Datamesh. I think it's what's interesting is that it's really an answer to the bottlenecks created by centralized IT, right? There's this notion of decentralizing that the data engineers and making the data domain owners, the people that know the data the best, have them be in control of publishing the data to the data consumers. There are other popular concepts actually happening right now as we speak around modern data stack, around data fabric that are also in many ways underpinned by this notion of decentralization, right? These are concepts that are underpinned by decentralization. And as the pendulum swings between decentralization and centralization, as we go back and forth in the world of IT and data, there are certain constants that need to be centralized over time. And one of those, I believe, is very much a centralized platform for data governance. And that's certainly, I think, where we come in. We'd love to hear more about how you use Elation. Yeah, so Elation helps us sort of, as you guys say, sort of that map of the treasure map of the data, right? So for consumers to find where their data is, that's where Elation helps us. It helps us with the data cataloging, storing all the metadata, and users can go in, they can sort of find the data that they need, and they can also find how others are using data. So there's a little bit of a crowdsourcing aspect that Elation helps us to do, whereby you can see, okay, my peer in the other group, well, that's how they use this piece of data. So I'm not going to spend hours trying to figure this out. I'm not sure you're going to use the query that they use, so yeah. So you have a master catalog, I presume, and then each of the brands has their own sub-catalogs, is that correct? Well, for the most part, we have that master catalog, and then the brands sort of use it separately themselves. The key here is, all that catalog isn't maintained by a centralized group as well, right? It's, again, maintained by the individual teams, and not only in the individual teams, but the folks that are responsible for the data, right? So I talked about the concept of crowdsourcing. Whoever sort of puts the data in has to make sure that they update the catalog and make sure that the definitions are there and everything sort of in line. So HBO, CNN, they each have their own sort of access to their catalog, but they feed into the master catalog. Is that the right way to think about it? Okay, and they have their own virtual data warehouses, right? They have ownership over that, they can spin them up, spin them down as they see fit, right? And they're governed. They're governed, and what's interesting is it's not just governed, right? Governance is a big word, it's a bit nebulous, but what's really being enabled here is this notion of self-service as well. There's two big sort of rockets that need to happen at the same time in any given organization. There's this notion that you want to put trustworthy data in the hands of data consumers while at the same time mitigating risk and that's precisely what Elation does. So I want to clarify this for the audience. So there's four principles of database, this came after you guys did it and I wonder how it aligns. Domain ownership, data as you were saying to the domain owners who have context, data as product, you guys are building data products and that creates two problems. How do you give people self-service infrastructure and how do you automate governance? So the first two, great, but then it creates these other problems. Is that aligned with your philosophy, where's alignment, what's different? Yeah, data products is exactly where we're going and that's sort of the domain-based design. That's really key as well. In our business, you think about who the customer is as an example, right? Depending on who you ask is going to be, the answer might be different. To the movie business, it's probably going to be the person who watches a movie in a theater. To the streaming business, to HBO Max, is the streamer, right? To others, it's someone watching live CNN on their TV, right? There's yet another group, think about all the franchising we do. So you see Batman action figures and t-shirts and Warner Brothers branded stuff in stores. That's yet another business unit. But at the end of the day, it's not a different person. It's you and me, right? We do all these things. So the domain concept, make sure that you ingest data and you bring data relevant to the context. However, not sort of making it so stringent where it cannot integrate. And then you integrate it at a higher level to create that 360. And it's discoverable. So the point is I don't have to go tap Ash on the shoulder and say, how do I get this data? Is it governed? Do I have access to it? Give me the rules of just, I go grab it, right? And the system computationally automates whether or not I have access to it and it's, as you say, self-service. In this case, exactly right. It enables people to just search for data and know that when they find the data whether it's trustworthy or not through trust flags and the like. So it's doing both of those things at the same time. How is it an enabler of solving some of the big challenges that the media and entertainment industry is going through? We've seen so much change the last couple of years. The rising consumer expectations aren't going to go back down. They're only going to come up. We want you to serve us up content that's relevant, that's personalized, that makes sense. I'd love to understand from your perspective, from an industry challenges perspective, how does this technology help customers like Warner Brothers Discovery meet business customers where they are and reduce the volume on those challenges? It's a great question. And as I mentioned earlier, we had five industry competency badges that were awarded to us by Snowflake and one of those for Media and Telco. And the reason for that is we're helping media companies understand their audiences better and ultimately serve up better experiences for their audiences. But we've got Ash right here that can tell us how that's happening in practice. Yeah, tell us. So I'll share a story. I always like to tell stories, right? Once upon a time before we had Elation in place it was like who you knew was how you got access to the data. So if I knew you and I knew you had access to a certain set kind of data that you know and it was you know your access to the right kind of data was based on the network you had at the company. I had to trust you. Yeah. I might not want to give up my data. That's it. And so that's where Elation sort of helps us democratize it but you know puts the governance and controls, right? There are certain sensitive things as well such as viewership, such as subscriber counts which are very important. So making sure that the right people have access to it that's the other problem that Elation helps us solve. That's precisely part of our integration with Snowflake in particular being able to define and manage policies with Elation saying you know certain people should have access to certain rows doing column level masking and having those policies actually enforced at the Snowflake data layer is precisely part of our value. And that's automated. And all that's automated. Right. So I don't have to think about it and I don't have to go through the tap on the shoulder. What has been the impact dash on data quality as you've pushed it down into the domains? That's a great question. So it has definitely improved but there you know so data quality is a very interesting subject because back to my example of you know when we started doing things we you know the centralized IT team always said well it has to be like this, right? And if it doesn't fit in this then it's bad quality. Well sometime context change, businesses change, right? You have to be able to react to it quickly. So making sure that a lot of that quality is managed at the decentralized level at the place where you have that business context that ensures you have the most up to date quality. We're talking about media industry changing so quickly, right? I mean would we have thought three years ago that people would watch a lot of these major movies on streaming services? I mean you know and but here's the reality, right? But you know so you have to react and you know having it at that level just helps you react faster. So if I play that back, data quality is not a static framework. It's flexible based on the business context and the business owners can make those adjustments because they own the data. That's it? That's exactly it? That's awesome. Wow, that's amazing progress that you guys have made. In quality if I could just add it also just changes depending on where you are in your data pipeline stage, right? Data quality, data observability, this is a very fast evolving space at the moment and if I look to my left right now I bet you I can obviously see a half dozen quality observability vendors right now. And so given that and given the fact that Elation still is sort of a central hub to find trustworthy data, we've actually announced an open data quality initiative allowing for best of breed data quality vendors to integrate with the platform. So whoever they are, whatever tool folks want to use they can use that particular tool of choice. And this all runs in the cloud or is it a hybrid? Everything is in the cloud. We are all in the cloud and you know, again helps us go faster. Let me ask you a question. One of the concepts I could go on forever in this topic, one of the concepts that Jamak Degani put forth is whether it's a Snowflake data warehouse or a Databricks data lake or an Oracle data warehouse they should all be, it should be inclusive. They should just be a node on the mesh. I'm like, wow, that sounds good but I haven't seen it yet, all right? I'm guessing that Snowflake and Elation enable all the self-serve, all this automated governance and that including those other items is got to be a one-off at this point in time. Do you ever see you expanding that scope or is it better off to just kind of leave it into the Snowflake data cloud? It's a good question. You know, I feel like where we're at today especially in terms of sort of technology giving us so many options. I don't think there's a one size fits all, right? Even though we are very heavily invested in Snowflake and we use Snowflake consistently across the organization but you could theoretically, you could have architecture that blends those two, right? Have different types of data platforms like a Teradata or an Oracle and sort of bring it all together. Today we have the technology, you know, that and all sorts of things that can sort of make sure that you query on different databases. So I don't think the technology is the problem. I think it's the organizational mindset. I think that's what gets in the way. Interesting. So I was going to ask you, will hybrid tables help you solve that problem? And maybe not, what you're saying, it's the organization that owns the Oracle database saying, hey, we have our system, it processes, it works, you know, go away. Yeah, well, you know, hybrid tables, I think is a great sort of next step in Snowflake's evolution. I think it's in my opinion, I think it's a game changer. But yeah, I mean, they can still exist. You could do hybrid tables right on Snowflake or you could kind of coexist as well. Yeah. Do you have a thought on this? Yeah, I do. I mean, we're always going to live in a time where you've got data distributed in throughout the organization and around the globe. And that could be even if you're all in on Snowflake, you could have data in Snowflake here, you could have data in Snowflake in EMEA in Europe somewhere, it could be anywhere. By the same token, you might be using. Every organization is using on-premises systems. They have data, they naturally have data everywhere. And so, you know, this one solution to this is really centralizing, as I mentioned, not just governance, but also metadata about all of the data in your organization so that you can enable people to search and find and discover trustworthy data no matter where it is in your organization. Yeah, that's a great point. I mean, if you have the data about the data, then you can treat these independent nodes as just that, right? And maybe there's some advantages of putting it all in the Snowflake cloud, but to your point, organizationally, that's just not feasible. Unfortunately, sorry, Snowflake, all the world's data is not going to go into Snowflake, but they pay a key role in accelerating what I'm hearing, your vision of data mesh. Yeah, absolutely. I think going forward in the future, we have to start thinking about data platforms as just one place where you sort of dump all the data. And that's where the mesh concept comes in. It is going to be a mesh, it's going to be distributed, and organizations have to be okay with that, and they have to embrace the tools. I mean, Facebook developed a tool called Presto many, many years ago that helps them solve exactly the same problem. So I think the technology is there. I think the organizational mindset needs to evolve. Yeah, definitely. Culture is one of the hardest things to change. Exactly. Guys, this was a master class in data mesh, I think. Thank you so much for coming on. We appreciate it, thank you so much. Of course, what Alation is doing with Snowflake and with Warner Brothers Discovery, keep that content coming. I got a lot of stuff I got to catch up on watching. Sounds good. Thank you for having us. Thanks, guys. For Dave Vellante, I'm Lisa Martin. You're watching theCUBE live from Snowflake Summit 22. We'll be back after a short break.