 The Cube presents Dell Technologies World, brought to you by Dell. Hi everyone, welcome back to Dell Technologies World 2022. You're watching the Cube's coverage of this three day coverage. Walter Wall, my name is Dave Vellante. John Furrier's here, Lisa Martin, David Nicholson. Talk of the town here is data. And one of the big announcements at the show is Snowflake and Dell partnering up, building ecosystems, Snowflake reaching into on-prem, allowing customers to actually access the Snowflake data cloud without moving the data, or if they want to move the data, they can. This is really one of the hotter announcements at the show. Martin Glenn is here as the senior director of storage product management at Dell Technologies. And Clark Patterson is the head of product marketing for Snowflake. Guys, welcome. Thanks for having us. So a lot of buzz around this. And you know, Clark, you and I have talked about the need to really extend your data vision. And this really is the next step, first step ever you've taken on-prem. Explain the motivation for this from your customer's perspective. Yeah, I mean, if you step back and think about Snowflake's vision and our mission of mobilizing the world's data, it's all around trying to break down silos for however customers define what a silo is, right? So we've had a lot of success breaking down silos from a workload perspective where we've expanded the platform to be data warehousing and data engineering and machine learning and data science and all the compute-intensive ways that people work with us. We've also had a lot of success in our sharing capabilities and how we're breaking down silos of organizations, right? So I can share data more seamlessly within my team. I can do it across the totally disparate organizations and break down silos that way. So this partnership is really like the next leg of the stool, so to speak, where we're breaking down the silos of the data and where the data lives ultimately, right? So up until this point, cloud, all focused there. Now we have this opportunity with Dell to expand that into on-premises world and people can bring all those data sets together. And the data target for this, Martin, is Dell ECS, right? Your object store and you've got S3 compatibility. Explain that. Yeah, we've actually got sort of two flavors. We'll start with ECS, which is our turnkey object storage solution. Object storage offers sort of the ultimate in flexibility, potential performance, ease of use, right? Which is why it fits so well with Snowflake's mission for sort of unlocking the data within the data center. So we'll offer it to begin with with ECS and then we also recently announced our software-defined object scale solution. So we'll add even more flexibility there. Okay, and Clark, the way it works is I can now access non-native Snowflake data using what, materialized views, external tables? How does that work? Some combination of all the above. So we've had in Snowflake a capability called external tables, which we refer to. It goes hand-in-hand with this notion of external stages. Basically, through the combination of those two capabilities, it's a metadata layer on data wherever it resides. So customers have actually used this in Snowflake for data lake data outside of Snowflake in the cloud up until this point. So it's effectively an extension of that functionality into the Dell on-premises worlds so that we can tap into those things. So we use the external stages to expose all the metadata about what's in the Dell environment, and then we build external tables in Snowflake so that data looks like it is in Snowflake and then the experience for the analyst or whomever it is is exactly as though that data lives in the Snowflake world. Okay, so for a while you've allowed non-native Snowflake data, but it had to be in the cloud, it was the first time it's on-prem, that's the innovation here. Okay, and if I want to bring it into the cloud, can I? Yeah, the connection here will help in a migration sense as well, right? So that's the good thing is it's really giving the user the choice. So we are integrating together as partners to make the connection as seamless as possible. And then the end user will say like, look, I've got data that needs to live on-premises for whatever reasons, data sovereignty, whatever they decide, and they can keep it there and still do the analytics on it in place. But if there's a need and a desire to use this as an opportunity to migrate some of that data to the cloud, that connection between our two platforms will make that easier. Well, Michael always says, hey, customer choice, we're flexible, so you're cool with that? That's been the mission since we kind of came together, right? Is if our customers needed to stay in their data center, if that makes more sense from a cost perspective or a data gravity perspective, then they can do that. But we also want to help them unlock the value of that data. So if they need to copy it up to the public cloud to take advantage of it, we're going to integrate directly with Snowflake to make that really easy to do. So there are engineering integrations here, obviously. That's required. Can you describe what that looks like? When is this, give us the details on when it's available. Sure, so it's going to be sort of second half this year, as you'll see, we're demoing it this week, but the availability of the second half this year. And fundamentally, it's the way Clark described it, that Snowflake will reach into our S3 interface using the standard S3 interface. We're qualifying between the way they expect that S3 interface to present the data and the way our platform works, just to ensure that there's smooth interaction between the two. So that's sort of the first simplest use case. And then the second example we gave where the customer can copy some of that data up to the public cloud, we're basically copying between two S3 buckets and making sure that Snowflake's snow pipe is aware that data's being made available and can easily ingest it. And then that just goes into a virtual warehouse and the customer doesn't know or care. Exactly. The compute happens in Snowflake the way it does in any other manner. And then I know you got to crawl, walk, run second half of this year, but I would imagine, okay, you're going to start with AWS, correct? And then eventually go to other clouds. I mean, that's going to take other technical integrations. I mean, obviously, so should we assume there's a roadmap here or is there someone in done? I would assume that. I mean, based on our multi-cloud approach, that's kind of our approach at least. Kind of makes sense, right? I mean, that would seem to be a natural progression. My other thought was, okay, I've got operational systems. They might be running in transaction systems running on a PowerMax. Is there a way to get the data into an object store and make that available? Now that opens up even more workloads. I know, you're not committing to doing that, but it just conceptually it seems like something a customer might want to do. Yeah, I 100% agree. I mean, I think when we brought our teams together we started with a blank slate. It was, what's the best solution we can build? We landed on this sort of first step, but we got lots of feedback from a lot of our big joint customers about this system over there, this potential integration over here, and whether it's PowerMax type systems or other file workloads with native Snowflake data types. I think this is just the beginning, right? We have lots of potential here. And I don't think you've announced pricing. It's premature for that, but have you thought about, and how are you thinking about the pricing model? I mean, you're a consumption based pricing. Is that kind of how this is going to work? Or is there sort of a new pricing model or haven't you figured that out yet? I don't know if you've got any of the details on that, but from a Snowflake perspective, I would assume it's consistent with how our customers engage with us today. And we'll offer both possibilities, right? So you can either continue with the standard, sort of capex motion, maybe that's the most optimal view from a cost perspective, or you can take advantage through IAPEX option, right? So you can do consumption on them also. Okay, so it could be a dual model, right? Depending on what the customer wants. If they're a Snowflake customer, obviously, it's going to be consumption based, however you guys price. What's happening, Clark, in the market? Explain why Snowflake has so much momentum and traction in the marketplace. So it's like I spend a lot of time doing analysis of why we win and lose, core part of my role. And there's a couple of, there's really three things that come up consistently as to why people are really excited about the Snowflake platform. One is the most simplest thing of all, it feels like it's just ease of use and it just works, right? And I think the way that this platform was built for the cloud from the ground up all the way back 10 years ago really allows us to deliver that seamless experience of just like instant compute when you want it, it goes away, only pay for what you use, very few knobs to turn and things like that. And so people absolutely love that factor. The other is multi-cloud. So there's definitely a lot of organizations out there that have a multi-cloud strategy and what that means to them can be highly variable. But regardless, they want to be able to interact across clouds in some capacity. And of course we are a single platform, like literally one single interface consistent across all the three cloud providers that we work upon. And it gives them that flexibility to mix and match cloud infrastructure and any Snowflake however they see fit. The last piece of it is sharing. And I think it's that ability as I kind of alluded to around like breaking down organizational silos and allow people to be able to actually connect with each other in ways that you couldn't do before. Like if you think about how you and I would have shared data before, I'd be like, hey Dave, I'm going to unload this table into a spreadsheet and I'm going to send it over an email and there's the whole host of issues that get introduced in that and world. Now it's like instantly available. I have a lot of control over it's governed, it's all these other things. And I can create kind of walled gardens so to speak of how far out I want that to go. It could be in a controlled environment of organizations that I want to collaborate with or I can put it on our marketplace and expose it to the whole world because I think there's a value in that and if I choose I can monetize it, right? So those, the ease of use aspect of it, absolutely it's just a fantastic platform, the multicloud aspect of it and our unique differentiation around sharing and our marketplace and monetization. Yeah and the sharing fund, I mean it's now discoverable. Like if you send me an email, what did you call that, when did you send that email? And then the same time I can forward that to somebody else who's not governed. So that just creates a nightmare for the compliance. Yeah, you think about how you revoke access in that situation, you just don't, right? Now I can just turn it off and you don't get access to that data anymore. Okay and then the other thing I wanted to ask you Clark is snowflake started really as an analytics platform, simplifying data warehousing. You're moving into that world of data science, the whole data lake movement, bringing those two worlds together. You know, I was talking to Benoit about this, maybe there's a semantic layer that helps us kind of talk between those two worlds, but you don't care, right? It's if it's in an object store, it can play in both of those worlds, right? That's right, yeah. It's up to you to figure it out and the customer, but from a storage standpoint, there it is, serve it up. And that's the thrust of this announcement, right? Is bringing together two great companies, the Dell platform, the snowflake platform and allowing organizations to bring that together and they decide, as we all know, customers decide how they're going to build their architecture and so this is just another way that we're helping them leverage the capabilities of our two great platforms. Is this push or pull or a little bit of both? I mean, where'd this come from? Were customers saying, hey, it would be kind of cool if we could have this? Or is it more, hey, what do you guys think? You know, where are you at with that? It was definitely both, right? I mean, so we certainly started with, high level idea that the technologies are complementary, right, I mean, as Clark just described. And at the same time, we had customers coming to us saying, hey, wait a minute, I'm doing this over here and this over here, how can I make this easier? So that was, like I said, we started with a blank sheet and lots of long customer conversations and this is what resulted, so. So what are the sequence of events to kind of roll this out? You said it's second half. You know, when do you start getting customers involved? Have you already, you know, to poke at this and what's that look like? Yeah, sure, I can't weigh in there, so absolutely. We've had a few of our big customers that have been involved sort of in the design already who are, understand how they want to use it. So I think our expectation is that now that the sort of demonstrations have been in place, we have some pre-functionality, we're going to see some initial testing and usage, some beta type situations with our customers and then second half will ramp from there. It's got to be a huge overlap between Dell customers and Snowflake customers. I mean, it's 100 billion, you can't not bump into Dell somewhere. Exactly, yeah. So where do you guys want to see this relationship go? Kind of how should we measure success? Maybe you could each give your perspectives of that. I mean, for us, I think it's really showing the value of the Snowflake platform in this new world where there's a whole new ecosystem of data that's accessible to us, right? So seeing those organizations that are saying like, look, I'm doing new things with on-premises data that I didn't think that I could do before or I'm driving efficiency in how I do analytics and data engineering and data science in ways that I couldn't do before because they were locked out of using a Snowflake-like technology, right? So I think for me, that's going to be that real excitement. I'm really curious to see how the collaboration and the sharing component comes into this, you know, where you can think of having an on-premises data strategy and a need, right? But you can really connect to cloud native customers and partners and suppliers that live in the Snowflake ecosystem, and that wasn't possible before. And so that is very conceivable and very possible through this relationship. So seeing how those edges get created in our world and how people start to collaborate across data both in the cloud and on-premise is going to be really exciting. I remember I asked Frank, it was kind of early in the pandemic, I asked him, come on and tell me about how you're managing things. And he was awesome. And I asked him at the time, you know, you ever going to do, bring this platform on-premise? He's like unequivocal. No way, that's never going to happen. We're not going to do a halfway housework cloud only. And I kept thinking, just God, but there's got to be a way to expand that, Tim, there's so much data out there. And so, boom, now we see the answer. Martin, from your standpoint, what does success look like? I think it starts with our partnership, right? So I've been doing this a long time, probably the first time I've worked so closely with a partner like Snowflake, joint customer conversations, joint solutioning, making sure what we're building is going to be really, truly as useful as possible to them. And I think we're going to let them guide us as we go forward here, right? You mentioned, you know, systems of record or other potential platforms. We're going to let them tell us where exactly the most value will come from this, the integration between the two companies. Yeah, follow the data. I remember the old days, you would, a hardware company like Dell would go to an ISP like Snowflake and say, hey, we ran some benchmarks. Your software runs really fast on our hardware. Can we work together? You go, yeah, of course. No problem. But wow, what a different dynamic it is today. Yeah, yeah, absolutely. All right, guys, hey, thanks so much for coming to theCUBE. It's great to see you. We'll see you at the Snowflake Summit. Snowflake Summit in a month and a half. And we're looking forward to that. All right, thank you again. Thank you, Dave. All right, keep it right there, everybody. This is Dave Vellante, wall-to-wall coverage of Dell Tech World 2022. We'll be right back.