>>Good morning. Welcome to the cubes. Continuous coverage of AWS reinvent 2021. I'm Lisa Martin with Dave Nicholson morning, Dave. Good morning. Here we are two sets, two live sets, two remote sets, a hundred guests on the program this week. This is day, second full day for us. We were here. Exactly. The number of announcements has my head spinning typical for AWS. So many. We won't even recap all the pre, but really from a thematic perspective. One of the things that I think we've both noticed is that the theme here is if you're not a data-driven company, if you don't have a data company at your core, you're going to be left behind. >>Yeah, yeah. That was that. That came through in a Swami's keynote this morning. That was the key setup for the announcements of all of the tools that are being rolled out to be able to achieve those goals. >>And we talked to, you know, we've had a number of AWS folks on the program this week, customers partners, and from another theme is this customer obsession, this customer centricity, really working backwards from the customer first to help them understand how do we, how do we build a data core? How do we build our redefine reinvent, Hubba reinvent our business around a data strategy because there is a competitor in the rear view mirror in every industry waiting to take place. If somebody doesn't have. And they really were very clear with, if you don't, if you're not a data company, now you're going to be left behind. >>Yeah, absolutely. The whole, it, especially the keynote this morning, it felt like a, it felt like they were very, very careful to set up the pins and then knock them all down. Talking about exactly what customers want to do, what customers have said in surveys they need to do. Uh, and then, oh, by the way, we have all of these tools. So exactly what you mentioned, not talking about technology first, but talking about the requirements the customers have, where customers say they want to be, where customers need to be, and then proving out that in fact, AWS has the entire tool kit. Uh, I, I mean it is mind boggling. Uh, I, uh, I wouldn't want to be another cloud this morning, waking up and seeing that it looks like I fell months behind somehow. Why not? While I was sleeping. >>One of the things that you and I were talking about as we walked over here this morning is that it's no wonder AWS. And they say they don't look in their rear view mirror. They don't need to. One of the things the magically that Adam talked about in his keynote yesterday was I kept saying the phrase, but you wanted more, you being, the customer did this, but you wanted more. So we did this, but you wanted more. So we did this, the fact that they start backwards, that customer flywheel. The thing that another theme that got me is all of the innovation that's coming out of AWS that comes from the customers, enables AWS to do so much more enables customers across every industry to do so much more and to competitively differentiate themselves if they truly lean in to the power of AWS and the partnership that can deliver. >>Yeah, no, no completely agree. And you know, from, from sort of an analysis perspective, um, I think it's interesting, uh, part of the theme this morning had to do with artificial intelligence and, um, and I, I just personally like to think of it more as augmented intelligence, uh, cause essentially what they're doing is they're building tools that are powerful, that, uh, that amplify human intelligence, it's not that AI replaces human intelligence, maybe it does for certain things, but machine learning, most of that is really machine training, humans, training the machines. Yeah. We'll get to a point. Yes. There are science experiments where, um, where the computer has learned on its own. Uh, but essentially all of the pieces that Swami was talking about putting together today, sources of data, various databases, everything together that equals an amazing amount of opportunity for human beings moving forward. So this isn't about machine technology, replacing the value of what people do, data scientists. I mean, I'm, I'm, I'm texting my kids data science used to be plastics right during their graduate. Now it's all about data science be a data scientist. Well, >>Another thing from a data perspective is the idea of data democratization. We've talked. We talk about that all the time. It's interesting. We were, I was talking about data mesh yesterday as, as a really facilitator and unlocking technical debt facilitating the democratization. So that because data is so scattered these days, I mean, it's one of the things that we saw when the pandemic stuck was people scattered. We're still scattered data sources are only proliferating data volume is only proliferating. Customers need to be able to harness the value of it in real time. I think another thing we learned in the pandemic is of there is no real-time is no longer a nice to have. It is absolutely essential. We've also seen the acceleration and talked a lot in the last couple of days about the acceleration of cloud adoption. And just thinking about the sheer volume of announcements. The last couple of days that AWS has achieved during a global pandemic is phenomenal. Obviously we saw every business have no choice, but to go digital and those that aren't here anymore, weren't able to do that fast enough. And with the scale at which they need to know, meet customer demand. >>Yeah. When you talk democratization often, it's a question of not only can you do it technically, but what is the cost associated with doing it? So when you lower the barrier to entry and you lower the friction associated with all sorts of transactions, not just financial transactions, but a transaction being teasing important information out of a pile of otherwise meaningless data, when you lower the cost of that, like what AWS is doing with all of these tools. I mean, really that's the, that's the end result of this? Uh, I look at these things through the lens of an economist and uh, and it's about decreasing friction and increasing efficiency. And when you do that, you can't imagine what the effect of that is on the entire world, just by lowering the cost of teasing out important information from otherwise useless data. Yeah. >>And another thing that we talked about the vision of this company, this is a 10th reinvent, 15 years of AWS, but the 10th reinvent and wanted to get your take on the last, you know, the first 15 years we'll say of AWS division, obviously this is the first year with a new head, new CEO, Adam. So Lipski, and, uh, we're, we're used to the, the JASSI era. This is now we're looking at like, what's the next 10 years of cloud innovation gonna look like at AWS? >>Yeah. I get a chill when you ask that question. I like, I was there working with Amazon before there was an AWS. So in the early two thousands, uh, working at a very large data storage company and we're working with then Amazon to help them figure out how they might someday put together something that they could not only use for themselves, but also rent out to others, sort of this novel idea. Um, and so that feels like yesterday to me, um, the idea that AWS was ECE to S3 science experiment, not ready for enterprise. All of that feels like yesterday. Those claims most people at the time realized that they were going to be disruptive. I don't think a lot of people realized how relatively quickly it was going to happen. And I don't think anyone could have predicted that in 2021, they would be as dominant as they are with such a small share of the overall it market. >>So onstage, they'll say five to 15% of it is in the cloud, right? Not five to 15% is in AWS, right. Five to 15% is in the cloud. So you think that AWS is a large organization firing on all cylinders. Now they're just at the beginning of addressing the total addressable market. So I, you know, um, uh, Dave Vellante, uh, had a great article on the subject of the natural inclination for people to worry when certain things get to a certain size. Um, I think that that's going to be a part of the conversation moving forward. Uh, uh, you know, just what does it mean when so much critical infrastructure is being handled by a single entity? Uh, from a user's perspective though, Hey, it's all good. The more they integrate, the more they bring things together, the easier they make my life as an it practitioner. So I, I don't, I have no idea five years from now what we're going to be looking at. >>No, and they, you know, one of the things that, that I think I've done maybe three or four reinvents and the, one of the themes always, or, or really kind of taglines is early innings. We're S we're still early days. You talked about, you know, when the 15% of it spending, being in the cloud, they have their, and they're so massive, but there's so much the Tam is, is enormous, absolutely enormous. And we're seeing, you know, the need for as data volumes, we're only going to continue to proliferate. We have to have the artificial intelligence and the machine learning to help the humans process the data, to be able to unlock all that value. Otherwise organizations, risk being left behind when they simply cannot get value out of the data either at all, or fast enough. >>Well, you know, you know, let me make one prediction actually. Okay. Okay. Um, I think that the definition of cloud, which is different in a lot of people's minds already is increasingly going to be broad broadened out to include what AWS at a certain period of time said, didn't matter. And that was what's going on on premises. It was clear that in the early days, AWS adopted an attitude that their, their stratospheric growth could be maintained on the back of net. New stuff. Only let the stuff on, on premises die. The stickiness associated with that on premises stuff has caused them to change direction over time in a very, very smart way. And so my prediction is that in five years, we might not be using the term cloud at all. We might be back to just calling it all it, because when you think about it, cloud represents this idea that the physical location isn't important, it's all virtualized. >>So if that's the case, why does it matter if it's in a data center that I own, or a colo or data center that AWS owns, it doesn't matter. So when you look at things like snowball and outpost and all of these other things where they're taking essentially AWS, and they're sticking it wherever it needs to be fit for function, you have latency, sovereignty, govern governance issues, fine. We'll run it as a black box in your data center. You don't touch it, you consume it, you interact with it just like it's in an AWS data center. So, so that's my prediction is that it's not that, you know, w we're going to start thinking about this comment that X percentage is in the cloud as being sort of like a, well, what do you even mean by that? That's a good point. Is it in the cloud? If it's in an outpost in my data center, right. Is that included in the 15% or does it not? Is that on-prem or isn't it? I think, I think within five years, we're all going to agree that it just doesn't matter. It's all it. And I think AWS will be still dominating the it space. >>Interesting prediction. I liked that you were bold in that, you know it, but it also shows what you talked about, shows how focused AWS is on the customer, really leaning into everything in the public cloud, you know, no more on-prem years ago. And, and then what, what is outpost announced two years ago? So it's, I think it's been a couple years to, >>Yeah, I confused. I can. I confuse announcements with deployments and things like that, but it's been years it's been years now. And I guarantee you that inside the halls of AWS, there were people who said, no, no, no, we are not going to be involved with boxes being shipped to raise data center floors. But like you said, customer it's about the customer, what the customer needs. And the reason why there is, you know, if the number is 85%, whatever the number is, um, there, there's a reason why that still remains in on premises data centers. And it's not primarily because it, people are stupid slash fearful. There are good reasons for it. You know, it has to do with the ROI of going through modernization and migration. And so what does AWS do? They figure out how to make the math work so that it makes sense for people to do things. >>And we've been talking for the last couple of days with a lot of the really, really important we'll continue to the really, really important service providers that are partners because AWS and the rest of the cloud providers can come up with as much technology as they want to, but they need partners to bridge the divide, right? To tease out the human value, the business value from the technology they provide. So again, opportunity people who look at AWS and say, wow, that's scary. They're so big. No, no, no, no, no. They're creating an ecosystem of opportunity around them. So that's just, >>Yeah, no, their ecosystem is huge. We talked about this cube being sponsored by AMD. What we talked about, the ecosystem of partners at the beginning of the segment, it's, it's, it's massive. We've talked with a lot of different partners, huge partners, smaller partners, overall, at the end of the day, what we see, what they create, all of the technologies that they're creating in response to customer needs. You know, we have to address it there. That the more data we have, the more things grow, the more complex and complexity has to be dealt with. And there's a lot of their ecosystem partners that are really have business models designed around helping customers. What does seamless actually mean? How do you actually become efficient? How do we do this in a frictionless way? Those are all great marketing terms, but they need this ecosystem of partners to help the customers actually make those reality. Right? And otherwise they're just words on a, on a piece of paper >>And it's not easy. Implementation is not easy. Sales is easy, implementation, not easy, right. And, uh, you know, the, the, the risk associated with failure can be very, very high, especially from a financial perspective. You know, the endless project that you're sinking money into that doesn't deliver any ROI just doesn't work. No, you can't, you can't do it. So, yeah, I, I I'm, I'm blown away by just today. Today's keynote going through knocking down those pins one just one after the other. Oh, CQL got it now, you know, for, for, for certain environments where before, yeah. We've got Oracle database support for this function. Now we have SQL also, it's just, you know, it's like this thousand by thousand grid check box, right? Check, check, check, check, check. And I can imagine all of those service providers just salivating, because they're thinking, oh my gosh, practice new practice that we need to spin up to be able to do this. Yep. Which points to a huge challenge. And that is actual practitioners where the rubber meets the road. All of these things require smart people to be able to implement. >>And the good news is a lot of those smart people are on the program this week here, you're going to be able to get, to hear so much more about the innovations that AWS and its ecosystem of partners are doing. We have a LA a full day again today, Dave, and again, tomorrow looking forward to hearing about the conversations that you had, I'm looking forward to interviewing guests and we look forward to having you continue to watch the cube, the global leader in live tech coverage.

>> Welcome to theCUBE, I am Dave Vellante. And today we're going to explore the ebb and flow of data as it travels into the cloud and the data lake. The concept of data lakes was alluring when it was first coined last decade by CTO James Dixon. Rather than be limited to highly structured and curated data that lives in a relational database in the form of an expensive and rigid data warehouse or a data mart. A data lake is formed by flowing data from a variety of sources into a scalable repository, like, say an S3 bucket that anyone can access, dive into, they can extract water, A.K.A data, from that lake and analyze data that's much more fine-grained and less expensive to store at scale. The problem became that organizations started to dump everything into their data lakes with no schema on our right, no metadata, no context, just shoving it into the data lake and figure out what's valuable at some point down the road. Kind of reminds you of your attic, right? Except this is an attic in the cloud. So it's too big to clean out over a weekend. Well look, it's 2021 and we should be solving this problem by now. A lot of folks are working on this, but often the solutions add other complexities for technology pros. So to understand this better, we're going to enlist the help of ChaosSearch CEO Ed Walsh, and Thomas Hazel, the CTO and Founder of ChaosSearch. We're also going to speak with Kevin Miller who's the Vice President and General Manager of S3 at Amazon web services. And of course they manage the largest and deepest data lakes on the planet. And we'll hear from a customer to get their perspective on this problem and how to go about solving it, but let's get started. Ed, Thomas, great to see you. Thanks for coming on theCUBE. >> Likewise. >> Face to face, it's really good to be here. >> It is nice face to face. >> It's great. >> So, Ed, let me start with you. We've been talking about data lakes in the cloud forever. Why is it still so difficult to extract value from those data lakes? >> Good question. I mean, data analytics at scale has always been a challenge, right? So, we're making some incremental changes. As you mentioned that we need to see some step function changes. But in fact, it's the reason ChaosSearch was really founded. But if you look at it, the same challenge around data warehouse or a data lake. Really it's not just to flowing the data in, it's how to get insights out. So it kind of falls into a couple of areas, but the business side will always complain and it's kind of uniform across everything in data lakes, everything in data warehousing. They'll say, "Hey, listen, I typically have to deal with a centralized team to do that data prep because it's data scientists and DBAs". Most of the time, they're a centralized group. Sometimes they're are business units, but most of the time, because they're scarce resources together. And then it takes a lot of time. It's arduous, it's complicated, it's a rigid process of the deal of the team, hard to add new data, but also it's hard to, it's very hard to share data and there's no way to governance without locking it down. And of course they would be more self-serve. So there's, you hear from the business side constantly now underneath is like, there's some real technology issues that we haven't really changed the way we're doing data prep since the two thousands, right? So if you look at it, it's, it falls two big areas. It's one, how to do data prep. How do you take, a request comes in from a business unit. I want to do X, Y, Z with this data. I want to use this type of tool sets to do the following. Someone has to be smart, how to put that data in the right schema, you mentioned. You have to put it in the right format, that the tool sets can analyze that data before you do anything. And then second thing, I'll come back to that 'cause that's the biggest challenge. But the second challenge is how these different data lakes and data warehouses are now persisting data and the complexity of managing that data and also the cost of computing it. And I'll go through that. But basically the biggest thing is actually getting it from raw data so the rigidness and complexity that the business sides are using it is literally someone has to do this ETL process, extract, transform, load. They're actually taking data, a request comes in, I need so much data in this type of way to put together. They're literally physically duplicating data and putting it together on a schema. They're stitching together almost a data puddle for all these different requests. And what happens is anytime they have to do that, someone has to do it. And it's, very skilled resources are scanned in the enterprise, right? So it's a DBS and data scientists. And then when they want new data, you give them a set of data set. They're always saying, what can I add to this data? Now that I've seen the reports. I want to add this data more fresh. And the same process has to happen. This takes about 60% to 80% of the data scientists in DPA's to do this work. It's kind of well-documented. And this is what actually stops the process. That's what is rigid. They have to be rigid because there's a process around that. That's the biggest challenge of doing this. And it takes an enterprise, weeks or months. I always say three weeks or three months. And no one challenges beyond that. It also takes the same skill set of people that you want to drive digital transformation, data warehousing initiatives, motorization, being data driven or all these data scientists and DBS they don't have enough of. So this is not only hurting you getting insights out of your day like in the warehouses. It's also, this resource constraint is hurting you actually getting. >> So that smallest atomic unit is that team, that's super specialized team, right? >> Right. >> Yeah. Okay. So you guys talk about activating the data lake. >> Yep. >> For analytics. What's unique about that? What problems are you all solving? You know, when you guys crew created this magic sauce. >> No, and basically, there's a lot of things. I highlighted the biggest one is how to do the data prep, but also you're persisting and using the data. But in the end, it's like, there's a lot of challenges at how to get analytics at scale. And this is really where Thomas and I founded the team to go after this, but I'll try to say it simply. What we're doing, I'll try to compare and contrast what we do compared to what you do with maybe an elastic cluster or a BI cluster. And if you look at it, what we do is we simply put your data in S3, don't move it, don't transform it. In fact, we're against data movement. What we do is we literally point and set that data and we index that data and make it available in a data representation that you can give virtual views to end-users. And those virtual views are available immediately over petabytes of data. And it actually gets presented to the end-user as an open API. So if you're elastic search user, you can use all your elastic search tools on this view. If you're a SQL user, Tableau, Looker, all the different tools, same thing with machine learning next year. So what we do is we take it, make it very simple. Simply put it there. It's already there already. Point us at it. We do the hard of indexing and making available. And then you publish in the open API as your users can use exactly what they do today. So that's, dramatically I'll give you a before and after. So let's say you're doing elastic search. You're doing logging analytics at scale, they're lending their data in S3. And then they're ETL physically duplicating and moving data. And typically deleting a lot of data to get in a format that elastic search can use. They're persisting it up in a data layer called leucine. It's physically sitting in memories, CPU, SSDs, and it's not one of them, it's a bunch of those. They in the cloud, you have to set them up because they're persisting ECC. They stand up same by 24, not a very cost-effective way to the cloud computing. What we do in comparison to that is literally pointing it at the same S3. In fact, you can run a complete parallel, the data necessary it's being ETL out. When just one more use case read only, or allow you to get that data and make this virtual views. So we run a complete parallel, but what happens is we just give a virtual view to the end users. We don't need this persistence layer, this extra cost layer, this extra time, cost and complexity of doing that. So what happens is when you look at what happens in elastic, they have a constraint, a trade-off of how much you can keep and how much you can afford to keep. And also it becomes unstable at time because you have to build out a schema. It's on a server, the more the schema scales out, guess what? you have to add more servers, very expensive. They're up seven by 24. And also they become brutalized. You lose one node, the whole thing has to be put together. We have none of that cost and complexity. We literally go from to keep whatever you want, whatever you want to keep an S3 is single persistence, very cost effective. And what we are able to do is, costs, we save 50 to 80%. Why? We don't go with the old paradigm of sit it up on servers, spin them up for persistence and keep them up 7 by 24. We're literally asking their cluster, what do you want to cut? We bring up the right compute resources. And then we release those sources after the query done. So we can do some queries that they can't imagine at scale, but we're able to do the exact same query at 50 to 80% savings. And they don't have to do any tutorial of moving that data or managing that layer of persistence, which is not only expensive, it becomes brittle. And then it becomes, I'll be quick. Once you go to BI, it's the same challenge, but the BI systems, the requests are constant coming at from a business unit down to the centralized data team. Give me this flavor of data. I want to use this piece of, you know, this analytic tool in that desk set. So they have to do all this pipeline. They're constantly saying, okay, I'll give you this data, this data, I'm duplicating that data, I'm moving it and stitching it together. And then the minute you want more data, they do the same process all over. We completely eliminate that. >> And those requests are queue up. Thomas, it had me, you don't have to move the data. That's kind of the exciting piece here, isn't it? >> Absolutely no. I think, you know, the data lake philosophy has always been solid, right? The problem is we had that Hadoop hang over, right? Where let's say we were using that platform, little too many variety of ways. And so, I always believed in data lake philosophy when James came and coined that I'm like, that's it. However, HTFS, that wasn't really a service. Cloud object storage is a service that the elasticity, the security, the durability, all that benefits are really why we founded on-cloud storage as a first move. >> So it was talking Thomas about, you know, being able to shut off essentially the compute so you don't have to keep paying for it, but there's other vendors out there and stuff like that. Something similar as separating, compute from storage that they're famous for that. And you have Databricks out there doing their lake house thing. Do you compete with those? How do you participate and how do you differentiate? >> Well, you know you've heard this term data lakes, warehouse, now lake house. And so what everybody wants is simple in, easy in, however, the problem with data lakes was complexity of out. Driving value. And I said, what if, what if you have the easy in and the value out? So if you look at, say snowflake as a warehousing solution, you have to all that prep and data movement to get into that system. And that it's rigid static. Now, Databricks, now that lake house has exact same thing. Now, should they have a data lake philosophy, but their data ingestion is not data lake philosophy. So I said, what if we had that simple in with a unique architecture and indexed technology, make it virtually accessible, publishable dynamically at petabyte scale. And so our service connects to the customer's cloud storage. Data stream the data in, set up what we call a live indexing stream, and then go to our data refinery and publish views that can be consumed the elastic API, use cabana Grafana, or say SQL tables look or say Tableau. And so we're getting the benefits of both sides, use scheme on read-write performance with scheme write-read performance. And if you can do that, that's the true promise of a data lake, you know, again, nothing against Hadoop, but scheme on read with all that complexity of software was a little data swamping. >> Well, you've got to start it, okay. So we got to give them a good prompt, but everybody I talked to has got this big bunch of spark clusters, now saying, all right, this doesn't scale, we're stuck. And so, you know, I'm a big fan of Jamag Dagani and our concept of the data lake and it's early days. But if you fast forward to the end of the decade, you know, what do you see as being the sort of critical components of this notion of, people call it data mesh, but to get the analytics stack, you're a visionary Thomas, how do you see this thing playing out over the next decade? >> I love her thought leadership, to be honest, our core principles were her core principles now, 5, 6, 7 years ago. And so this idea of, decentralize that data as a product, self-serve and, and federated computer governance, I mean, all that was our core principle. The trick is how do you enable that mesh philosophy? I can say we're a mesh ready, meaning that, we can participate in a way that very few products can participate. If there's gates data into your system, the CTL, the schema management, my argument with the data meshes like producers and consumers have the same rights. I want the consumer, people that choose how they want to consume that data. As well as the producer, publishing it. I can say our data refinery is that answer. You know, shoot, I'd love to open up a standard, right? Where we can really talk about the producers and consumers and the rights each others have. But I think she's right on the philosophy. I think as products mature in this cloud, in this data lake capabilities, the trick is those gates. If you have to structure up front, if you set those pipelines, the chance of you getting your data into a mesh is the weeks and months that Ed was mentioning. >> Well, I think you're right. I think the problem with data mesh today is the lack of standards you've got. You know, when you draw the conceptual diagrams, you've got a lot of lollipops, which are APIs, but they're all unique primitives. So there aren't standards, by which to your point, the consumer can take the data the way he or she wants it and build their own data products without having to tap people on the shoulder to say, how can I use this?, where does the data live? And being able to add their own data. >> You're exactly right. So I'm an organization, I'm generating data, when the courageously stream it into a lake. And then the service, a ChaosSearch service, is the data is discoverable and configurable by the consumer. Let's say you want to go to the corner store. I want to make a certain meal tonight. I want to pick and choose what I want, how I want it. Imagine if the data mesh truly can have that producer of information, you know, all the things you can buy a grocery store and what you want to make for dinner. And if you'd static, if you call up your producer to do the change, was it really a data mesh enabled service? I would argue not. >> Ed, bring us home. >> Well, maybe one more thing with this. >> Please, yeah. 'Cause some of this is we're talking 2031, but largely these principles are what we have in production today, right? So even the self service where you can actually have a business context on top of a data lake, we do that today, we talked about, we get rid of the physical ETL, which is 80% of the work, but the last 20% it's done by this refinery where you can do virtual views, the right or back and do all the transformation need and make it available. But also that's available to, you can actually give that as a role-based access service to your end-users, actually analysts. And you don't want to be a data scientist or DBA. In the hands of a data scientist the DBA is powerful, but the fact of matter, you don't have to affect all of our employees, regardless of seniority, if they're in finance or in sales, they actually go through and learn how to do this. So you don't have to be it. So part of that, and they can come up with their own view, which that's one of the things about data lakes. The business unit wants to do themselves, but more importantly, because they have that context of what they're trying to do instead of queuing up the very specific request that takes weeks, they're able to do it themselves. >> And if I have to put it on different data stores and ETL that I can do things in real time or near real time. And that's game changing and something we haven't been able to do ever. >> And then maybe just to wrap it up, listen, you know 8 years ago, Thomas and his group of founders, came up with the concept. How do you actually get after analytics at scale and solve the real problems? And it's not one thing, it's not just getting S3. It's all these different things. And what we have in market today is the ability to literally just simply stream it to S3, by the way, simply do, what we do is automate the process of getting the data in a representation that you can now share an augment. And then we publish open API. So can actually use a tool as you want, first use case log analytics, hey, it's easy to just stream your logs in. And we give you elastic search type of services. Same thing that with CQL, you'll see mainstream machine learning next year. So listen, I think we have the data lake, you know, 3.0 now, and we're just stretching our legs right now to have fun. >> Well, and you have to say it log analytics. But if I really do believe in this concept of building data products and data services, because I want to sell them, I want to monetize them and being able to do that quickly and easily, so I can consume them as the future. So guys, thanks so much for coming on the program. Really appreciate it.

>>Welcome to the cube. I'm Dave Volante. Today, we're going to explore the ebb and flow of data as it travels into the cloud. In the data lake, the concept of data lakes was a Loring when it was first coined last decade by CTO James Dickson, rather than be limited to highly structured and curated data that lives in a relational database in the form of an expensive and rigid data warehouse or a data Mart, a data lake is formed by flowing data from a variety of sources into a scalable repository, like say an S3 bucket that anyone can access, dive into. They can extract water. It can a data from that lake and analyze data. That's much more fine-grained and less expensive to store at scale. The problem became that organizations started to dump everything into their data lakes with no schema on it, right? No metadata, no context to shove it into the data lake and figure out what's valuable. >>At some point down the road kind of reminds you of your attic, right? Except this is an attic in the cloud. So it's too big to clean out over a weekend. We'll look it's 2021 and we should be solving this problem by now, a lot of folks are working on this, but often the solutions at other complexities for technology pros. So to understand this better, we're going to enlist the help of chaos search CEO and Walsh and Thomas Hazel, the CTO and founder of chaos search. We're also going to speak with Kevin Miller. Who's the vice president and general manager of S3 at Amazon web services. And of course they manage the largest and deepest data lakes on the planet. And we'll hear from a customer to get their perspective on this problem and how to go about solving it, but let's get started. Ed Thomas. Great to see you. Thanks for coming on the cube. Likewise face. It's really good to be in this nice face. Great. So let me start with you. We've been talking about data lakes in the cloud forever. Why is it still so difficult to extract value from those data? >>Good question. I mean, a data analytics at scale is always been a challenge, right? So, and it's, uh, we're making some incremental changes. As you mentioned that we need to seem some step function changes, but, uh, in fact, it's the reason, uh, search was really founded. But if you look at it the same challenge around data warehouse or a data lake, really, it's not just a flowing the data in is how to get insights out. So it kind of falls into a couple of areas, but the business side will always complain and it's kind of uniform across everything in data lakes, everything that we're offering, they'll say, Hey, listen, I typically have to deal with a centralized team to do that data prep because it's data scientist and DBS. Most of the time they're a centralized group, sometimes are business units, but most of the time, because they're scarce resources together. >>And then it takes a lot of time. It's arduous, it's complicated. It's a rigid process of the deal of the team, hard to add new data. But also it's hard to, you know, it's very hard to share data and there's no way to governance without locking it down. And of course they would be more self-service. So there's you hear from the business side constantly now underneath is like, there's some real technology issues that we haven't really changed the way we're doing data prep since the two thousands. Right? So if you look at it, it's, it falls, uh, two big areas. It's one. How do data prep, how do you take a request comes in from a business unit. I want to do X, Y, Z with this data. I want to use this type of tool sets to do the following. Someone has to be smart, how to put that data in the right schema. >>You mentioned you have to put it in the right format, that the tool sets can analyze that data before you do anything. And then secondly, I'll come back to that because that's a biggest challenge. But the second challenge is how these different data lakes and data we're also going to persisting data and the complexity of managing that data and also the cost of computing. And I'll go through that. But basically the biggest thing is actually getting it from raw data so that the rigidness and complexity that the business sides are using it is literally someone has to do this ETL process extract, transform load. They're actually taking data request comes in. I need so much data in this type of way to put together their Lilly, physically duplicating data and putting it together and schema they're stitching together almost a data puddle for all these different requests. >>And what happens is anytime they have to do that, someone has to do it. And it's very skilled. Resources are scant in the enterprise, right? So it's a DBS and data scientists. And then when they want new data, you give them a set of data set. They're always saying, what can I add this data? Now that I've seen the reports, I want to add this data more fresh. And the same process has to happen. This takes about 60 to 80% of the data scientists in DPA's to do this work. It's kind of well-documented. Uh, and this is what actually stops the process. That's what is rigid. They have to be rigid because there's a process around that. Uh, that's the biggest challenge to doing this. And it takes in the enterprise, uh, weeks or months. I always say three weeks to three months. And no one challenges beyond that. It also takes the same skill set of people that you want to drive. Digital transformation, data, warehousing initiatives, uh, monitorization being, data driven, or all these data scientists and DBS. They don't have enough of, so this is not only hurting you getting insights out of your dead like that, or else it's also this resource constraints hurting you actually getting smaller. >>The Tomic unit is that team that's super specialized team. Right. Right. Yeah. Okay. So you guys talk about activating the data lake. Yep, sure. Analytics, what what's unique about that? What problems are you all solving? You know, when you guys crew created this, this, this magic sauce. >>No, and it basically, there's a lot of things I highlighted the biggest one is how to do the data prep, but also you're persisting and using the data. But in the end, it's like, there's a lot of challenges that how to get analytics at scale. And this is really where Thomas founded the team to go after this. But, um, I'll try to say it simply, what are we doing? I'll try to compare and stress what we do compared to what you do with maybe an elastic cluster or a BI cluster. Um, and if you look at it, what we do is we simply put your data in S3, don't move it, don't transform it. In fact, we're not we're against data movement. What we do is we literally pointed at that data and we index that data and make it available in a data representation that you can give virtual views to end users. >>And those virtual views are available immediately over petabytes of data. And it re it actually gets presented to the end user as an open API. So if you're elastic search user, you can use all your lesser search tools on this view. If you're a SQL user, Tableau, Looker, all the different tools, same thing with machine learning next year. So what we do is we take it, make it very simple. Simply put it there. It's already there already. Point is at it. We do the hard of indexing and making available. And then you publish in the open API as your users can use exactly what they do today. So that's dramatically. I'll give you a before and after. So let's say you're doing elastic search. You're doing logging analytics at scale, they're lending their data in S3. And then they're,, they're physically duplicating a moving data and typically deleting a lot of data to get in a format that elastic search can use. >>They're persisting it up in a data layer called leucine. It's physically sitting in memories, CPU, uh, uh, SSDs. And it's not one of them. It's a bunch of those. They in the cloud, you have to set them up because they're persisting ECC. They stand up semi by 24, not a very cost-effective way to the cloud, uh, cloud computing. What we do in comparison to that is literally pointing it at the same S3. In fact, you can run a complete parallel, the data necessary. It's being ETL. That we're just one more use case read only, or allow you to get that data and make this virtual views. So we run a complete parallel, but what happens is we just give a virtual view to the end users. We don't need this persistence layer, this extra cost layer, this extra, um, uh, time cost and complexity of doing that. >>So what happens is when you look at what happens in elastic, they have a constraint, a trade-off of how much you can keep and how much you can afford to keep. And also it becomes unstable at time because you have to build out a schema. It's on a server, the more the schema scales out, guess what you have to add more servers, very expensive. They're up seven by 24. And also they become brittle. As you lose one node. The whole thing has to be put together. We have none of that cost and complexity. We literally go from to keep whatever you want, whatever you want to keep an S3, a single persistence, very cost effective. And what we do is, um, costs. We save 50 to 80% why we don't go with the old paradigm of sit it up on servers, spin them up for persistence and keep them up. >>Somebody 24, we're literally asking her cluster, what do you want to cut? We bring up the right compute resources. And then we release those sources after the query done. So we can do some queries that they can't imagine at scale, but we're able to do the exact same query at 50 to 80% savings. And they don't have to do any of the toil of moving that data or managing that layer of persistence, which is not only expensive. It becomes brittle. And then it becomes an I'll be quick. Once you go to BI, it's the same challenge, but the BI systems, the requests are constant coming at from a business unit down to the centralized data team. Give me this flavor of debt. I want to use this piece of, you know, this analytic tool in that desk set. So they have to do all this pipeline. They're constantly saying, okay, I'll give you this data, this data I'm duplicating that data. I'm moving in stitching together. And then the minute you want more data, they do the same process all over. We completely eliminate that. >>The questions queue up, Thomas, it had me, you don't have to move the data. That's, that's kind of the >>Writing piece here. Isn't it? I absolutely, no. I think, you know, the daylight philosophy has always been solid, right? The problem is we had that who do hang over, right? Where let's say we were using that platform, little, too many variety of ways. And so I always believed in daily philosophy when James came and coined that I'm like, that's it. However, HTFS that wasn't really a service cloud. Oddish storage is a service that the, the last society, the security and the durability, all that benefits are really why we founded, uh, Oncotype storage as a first move. >>So it was talking Thomas about, you know, being able to shut off essentially the compute and you have to keep paying for it, but there's other vendors out there and stuff like that. Something similar as separating, compute from storage that they're famous for that. And, and, and yet Databricks out there doing their lake house thing. Do you compete with those? How do you participate and how do you differentiate? >>I know you've heard this term data lakes, warehouse now, lake house. And so what everybody wants is simple in easy N however, the problem with data lakes was complexity of out driving value. And I said, what if, what if you have the easy end and the value out? So if you look at, uh, say snowflake as a, as a warehousing solution, you have to all that prep and data movement to get into that system. And that it's rigid static. Now, Databricks, now that lake house has exact same thing. Now, should they have a data lake philosophy, but their data ingestion is not daily philosophy. So I said, what if we had that simple in with a unique architecture, indexed technology, make it virtually accessible publishable dynamically at petabyte scale. And so our service connects to the customer's cloud storage data, stream the data in set up what we call a live indexing stream, and then go to our data refinery and publish views that can be consumed the lasted API, use cabana Grafana, or say SQL tables look or say Tableau. And so we're getting the benefits of both sides, you know, schema on read, write performance with scheme on, right. Reperformance. And if you can do that, that's the true promise of a data lake, you know, again, nothing against Hadoop, but a schema on read with all that complexity of, uh, software was, uh, what was a little data, swamp >>Got to start it. Okay. So we got to give a good prompt, but everybody I talked to has got this big bunch of spark clusters now saying, all right, this, this doesn't scale we're stuck. And so, you know, I'm a big fan of and our concept of the data lake and it's it's early days. But if you fast forward to the end of the decade, you know, what do you see as being the sort of critical components of this notion of, you know, people call it data mesh, but you've got the analytics stack. Uh, you, you, you're a visionary Thomas, how do you see this thing playing out over the next? >>I love for thought leadership, to be honest, our core principles were her core principles now, you know, 5, 6, 7 years ago. And so this idea of, you know, de centralize that data as a product, you know, self-serve and, and federated, computer, uh, governance, I mean, all that, it was our core principle. The trick is how do you enable that mesh philosophy? We, I could say we're a mesh ready, meaning that, you know, we can participate in a way that very few products can participate. If there's gates data into your system, the CTLA, the schema management, my argument with the data meshes like producers and consumers have the same rights. I want the consumer people that choose how they want to consume that data, as well as the producer publishing it. I can say our data refinery is that answer. You know, shoot, I love to open up a standard, right, where we can really talk about the producers and consumers and the rights each others have. But I think she's right on the philosophy. I think as products mature in this cloud, in this data lake capabilities, the trick is those gates. If you have the structure up front, it gets at those pipelines. You know, the chance of you getting your data into a mesh is the weeks and months that it was mentioning. >>Well, I think you're right. I think the problem with, with data mesh today is the lack of standards you've got. You know, when you draw the conceptual diagrams, you've got a lot of lollipops, which are API APIs, but they're all, you know, unique primitives. So there aren't standards by which to your point, the consumer can take the data the way he or she wants it and build their own data products without having to tap people on the shoulder to say, how can I use this? Where's the data live and, and, and, and, and being able to add their own >>You're exactly right. So I'm an organization I'm generally data will be courageous, a stream it to a lake. And then the service, uh, Ks search service is the data's con uh, discoverable and configurable by the consumer. Let's say you want to go to the corner store? You know, I want to make a certain meal tonight. I want to pick and choose what I want, how I want it. Imagine if the data mesh truly can have that producer of information, you, all the things you can buy a grocery store and what you want to make for dinner. And if you'd static, if you call up your producer to do the change, was it really a data mesh enabled service? I would argue not that >>Bring us home >>Well. Uh, and, um, maybe one more thing with this, cause some of this is we talking 20, 31, but largely these principles are what we have in production today, right? So even the self service where you can actually have business context on top of a debt, like we do that today, we talked about, we get rid of the physical ETL, which is 80% of the work, but the last 20% it's done by this refinery where you can do virtual views, the right our back and do all the transformation need and make it available. But also that's available to, you can actually give that as a role-based access service to your end users actually analysts, and you don't want to be a data scientist or DBA in the hands of a data science. The DBA is powerful, but the fact of matter, you don't have to affect all of our employees, regardless of seniority. If they're in finance or in sales, they actually go through and learn how to do this. So you don't have to be it. So part of that, and they can come up with their own view, which that's one of the things about debt lakes, the business unit wants to do themselves, but more importantly, because they have that context of what they're trying to do instead of queuing up the very specific request that takes weeks, they're able to do it themselves and to find out that >>Different data stores and ETL that I can do things in real time or near real time. And that's that's game changing and something we haven't been able to do, um, ever. Hmm. >>And then maybe just to wrap it up, listen, um, you know, eight years ago is a group of founders came up with the concept. How do you actually get after analytics at scale and solve the real problems? And it's not one thing it's not just getting S3, it's all these different things. And what we have in market today is the ability to literally just simply stream it to S3 by the way, simply do what we do is automate the process of getting the data in a representation that you can now share an augment. And then we publish open API. So can actually use a tool as you want first use case log analytics, Hey, it's easy to just stream your logs in and we give you elastic search puppet services, same thing that with CQL, you'll see mainstream machine learning next year. So listen, I think we have the data lake, you know, 3.0 now, and we're just stretching our legs run off >>Well, and you have to say it log analytics. But if I really do believe in this concept of building data products and data services, because I want to sell them, I want to monetize them and being able to do that quickly and easily, so that can consume them as the future. So guys, thanks so much for coming on the program. Really appreciate it. All right. In a moment, Kevin Miller of Amazon web services joins me. You're watching the cube, your leader in high tech coverage.

>>Really excited about this year's data, citizens with so many of you together. Uh, I'm going to talk today about accessible data, because what good is the data. If you can get it into your hands and shop for it, but you can't understand it. Uh, and I'm here today with, uh, bald, really thrilled to be here with Paul. Paul is an award-winning author on all topics data. I think 20 books with 21st on the way over 300 articles, he's been a frequent speaker. He's an expert in future trends. Uh, he's a VP at cognitive systems, uh, over at IBM teachers' data also, um, at the business school and as a champion of diversity initiatives. Paul, thank you for being here, really the conformance, uh, to the session with you. >>Oh, thanks for having me. It's a privilege. >>So let's get started with, uh, our origins and data poll. Um, and I'll start with a little story of my own. So, uh, I trained as an engineer way back when, uh, and, um, in one of the courses we got as an engineer, it was about databases. So we got the stick thick book of CQL and me being in it for the programming. I was like, well, who needs this stuff? And, uh, I wanted to do my part in terms of making data accessible. So essentially I, I was the only book that I sold on. Uh, obviously I learned some hard lessons, uh, later on, as I did a master's in AI after that, and then joined the database research lab at the university that Libra spun off from. Uh, but Hey, we all learned along the way. And, uh, Paula, I'm really curious. Um, when did you awaken first to data? If you will? >>You know, it's really interesting Stan, because I come from the opposite side, an undergrad in economics, uh, with some, uh, information systems research at the higher level. And so I think I was always attuned to what data could do, but I didn't understand how to get at it and the kinds of nuances around it. So then I started this job, a database company, like 27 years ago, and it started there, but I would say the awakening has never stopped because the data game is always changing. Like I look at these epochs that I've been through data. I was a real relational databases thinking third normal form, and then no SQL databases. And then I watch no SQL be about no don't use SQL, then wait a minute. Not only sequel. And today it's really for the data citizens about wait, no, I need SQL. So, um, I think I'm always waking up in data, so I'll call it a continuum if you will. But that was it. It was trying to figure out the technology behind driving analytics in which I took in school. >>Excellent. And I fully agree with you there. Uh, every couple of years they seem to reinvent new stuff and they want to be able to know SQL models. Let me see. I saw those come and go. Uh, obviously, and I think that's, that's a challenge for most people because in a way, data is a very abstract concepts, um, until you get down in the weeds and then it starts to become really, really messy, uh, until you, you know, from that end button extract a certain insights. Um, and as the next thing I want to talk about with you is that challenging organizations, we're hearing a lot about data, being valuable data, being the new oil data, being the new soil, the new gold, uh, data as an asset is being used as a slogan all over. Uh, people are investing a lot in data over multiple decades. Now there's a lot of new data technologies, always, but still, it seems that organizations fundamentally struggle with getting people access to data. What do you think are some of the key challenges that are underlying the struggles that mud, that organizations seem to face when it comes to data? >>Yeah. Listen, Stan, I'll tell you a lot of people I think are stuck on what I call their data, acumen curves, and you know, data is like a gym membership. If you don't use it, you're not going to get any value on it. And that's what I mean by accurate. And so I like to think that you use the analogy of some mud. There's like three layers that are holding a lot of organizations back at first is just the amount of data. Now, I'm not going to give you some stat about how many times I can go to the moon and back with the data regenerate, but I will give you one. I found interesting stat. The average human being in their lifetime will generate a petabyte of data. How much data is that? If that was my apple music playlist, it would be about 2000 years of nonstop music. >>So that's some kind of playlist. And I think what's happening for the first layer of mud is when I first started writing about data warehousing and analytics, I would be like, go find a needle in the haystack. But now it's really finding a needle in a stack of needles. So much data. So little time that's level one of mine. I think the second thing is people are looking for some kind of magic solution, like Cinderella's glass slipper, and you put it on her. She turns into a princess that's for Disney movies, right? And there's nothing magical about it. It is about skill and acumen and up-skilling. And I think if you're familiar with the duper, you recall the Hadoop craze, that's exactly what happened, right? Like people brought all their data together and everyone was going to be able to access it and give insights. >>And it teams said it was pretty successful, but every line of business I ever talked to said it was a complete failure. And the third layer is governance. That's actually where you're going to find some magic. And the problem in governance is every client I talked to is all about least effort to comply. They don't want to violate GDPR or California consumer protection act or whatever governance overlooks, where they do business and governance. When you don't lead me separate to comply and try not to get fine, but as an accelerant to your analytics, and that gets you out of that third layer of mud. So you start to invoke what I call the wisdom of the crowd. Now imagine taking all these different people with intelligence about the business and giving them access and acumen to hypothesize on thousands of ideas that turn into hundreds, we test and maybe dozens that go to production. So those are three layers that I think every organization is facing. >>Well. Um, I definitely follow on all the days, especially the one where people see governance as a, oh, I have to comply to this, which always hurts me a little bit, honestly, because all good governance is about making things easier while also making sure that they're less riskier. Um, but I do want to touch on that Hadoop thing a little bit, uh, because for me in my a decade or more over at Libra, we saw it come as well as go, let's say around 2015 to 2020 issue. So, and it's still around. Obviously once you put your data in something, it's very hard to make it go away, but I've always felt that had do, you know, it seemed like, oh, now we have a bunch of clusters and a bunch of network engineers. So what, >>Yeah. You know, Stan, I fell for, I wrote the book to do for dummies and it had such great promise. I think the problem is there wasn't enough education on how to extract value out of it. And that's why I say it thinks it's great. They liked clusters and engineers that you just said, but it didn't drive lineup >>Business. Got it. So do you think that the whole paradigm with the clouds that we're now on is going to fundamentally change that or is just an architectural change? >>Yeah. You know, it's, it's a great comment. What you're seeing today now is the movement for the data lake. Maybe a way from repositories, like Hadoop into cloud object stores, right? And then you look at CQL or other interfaces over that not allows me to really scale compute and storage separately, but that's all the technical stuff at the end of the day, whether you're on premise hybrid cloud, into cloud software, as a service, if you don't have the acumen for your entire organization to know how to work with data, get value from data, this whole data citizen thing. Um, you're not going to get the kind of value that goes into your investment, right? And I think that's the key thing that business leaders need to understand is it's not about analytics for kind of science project sakes. It's about analytics to drive. >>Absolutely. We fully agree with that. And I want to touch on that point. You mentioned about the wisdom of the crowds, the concept that I love about, right, and your organization is a big grout full of what we call data citizens. Now, if I remember correctly from the book of the wisdom of the crowds, there's, there's two points that really, you have to take Canada. What is, uh, for the wisdom of the grounds to work, you have to have all the individuals enabled, uh, for them to have access to the right information and to be able to share that information safely kept from the bias from others. Otherwise you're just biasing the outcome. And second, you need to be able to somehow aggregate that wisdom up to a certain decision. Uh, so as Felix mentioned earlier, we all are United by data and it's a data citizen topic. >>I want to touch on with you a little bit, because at Collibra we look at it as anyone who uses data to do their job, right. And 2020 has sort of accelerated digitization. Uh, but apart from that, I've always believed that, uh, you don't have to have data in your title, like a data analyst or a data scientist to be a data citizen. If I take a look at the example inside of Libra, we have product managers and they're trying to figure out which features are most important and how are they used and what patterns of behavior is there. You have a gal managers, and they're always trying to know the most they can about their specific accounts, uh, to be able to serve as them best. So for me, the data citizen is really in its broadest sense. Uh, anyone who uses data to do their job, does that, does that resonate with you? >>Yeah, absolutely. It reminds me of myself. And to be honest in my eyes where I got started from, and I agree, you don't need the word data in your title. What you need to have is curiosity, and that is in your culture and in your being. And, and I think as we look at organizations to transform and take full advantage of their, their data investments, they're going to need great governance. I guarantee you that, but then you're going to have to invest in this data citizen concept. And the first thing I'll tell you is, you know, that kind of acumen, if you will, as a team sport, it's not a departmental sport. So you need to think about what are the upskilling programs of where we can reach across to the technical and the non-technical, you know, lots and lots of businesses rely on Microsoft Excel. >>You have data citizens right there, but then there's other folks who are just flat out curious about stuff. And so now you have to open this up and invest in those people. Like, why are you paying people to think about your business without giving the data? It would be like hiring Tom Brady as a quarterback and telling him not to throw a pass. Right. And I see it all the time. So we kind of limit what we define as data citizen. And that's why I love what you said. You don't need the word data in your title and more so if you don't build the acumen, you don't know how to bring the data together, maybe how to wrangle it, but where did it come from? And where can you fixings? One company I worked with had 17 definitions for a sales individual, 17 definitions, and the talent team and HR couldn't drive to a single definition because they didn't have the data accurate. So when you start thinking of the data citizen, concept it about enabling everybody to shop for data much. Like I would look for a USB cable on Amazon, but also to attach to a business glossary for definition. So we have a common version of what a word means, the lineage of the data who owns it, who did it come from? What did it do? So bring that all together. And, uh, I will tell you companies that invest in the data, citizen concept, outperform companies that don't >>For all of that, I definitely fully agree that there's enough research out there that shows that the ones who are data-driven are capturing the most markets, but also capturing the most growth. So they're capturing the market even faster. And I love what you said, Paul, about, um, uh, the brains, right? You've already paid for the brains you've already invested in. So you may as well leverage them. Um, you may as well recognize and, and enable the data citizens, uh, to get access to the assets that they need to really do their job properly. That's what I want to touch on just a little bit, if, if you're capable, because for me, okay. Getting access to data is one thing, right? And I think you already touched on a few items there, but I'm shopping for data. Now I have it. I have a cul results set in my hands. Let's say, but I'm unable to read and write data. Right? I don't know how to analyze it. I don't know maybe about bias. Uh, maybe I, I, I don't know how to best visualize it. And maybe if I do, maybe I don't know how to craft a compelling persuasion narrative around it to change my bosses decisions. So from your viewpoint, do you think that it's wise for companies to continuously invest in data literacy to continuously upgrade that data citizens? If you will. >>Yeah, absolutely. Forest. I'm going to tell you right now, data literacy years are like dog years stage. So fast, new data types, new sources of data, new ways to get data like API APIs and microservices. But let me take it away from the technical concept for a bit. I want to talk to you about the movie. A star is born. I'm sure most of you have seen it or heard it Bradley Cooper, lady Gaga. So everyone knows the movie. What most people probably don't know is when lady Gaga teamed up with Bradley Cooper to do this movie, she demanded that he sing everything like nothing could be auto-tuned everything line. This is one of the leading actors of Hollywood. They filmed this remake in 42 days and Bradley Cooper spent 18 months on singing lessons. 18 months on a guitar lessons had a voice coach and it's so much and so forth. >>And so I think here's the point. If one of the best actors in the world has to invest three and a half years for 42 days to hit a movie out of the park. Why do we think we don't need a continuous investment in data literacy? Even once you've done your initial training, if you will, over the data, citizen, things are going to change. I don't, you don't. If I, you Stan, if you go to the gym and workout every day for three months, you'll never have to work out for the rest of your life. You would tell me I was ridiculous. So your data literacy is no different. And I will tell you, I have managed thousands of individuals, some of the most technical people around distinguished engineers, fellows, and data literacy comes from curiosity and a culture of never ending learning. That is the number one thing to success. >>And that curiosity, I hire people who are curious, I'll give you one more story. It's about Mozart. And this 21 year old comes to Mozart and he says, Mozart, can you teach me how to compose a symphony? And Mozart looks at this person that says, no, no, you're too young, too young. You compose your fourth symphony when you were 12 and Mozart looks at him and says, yeah, but I didn't go around asking people how to compose a symphony. Right? And so the notion of that story is curiosity. And those people who show up in always want to learn, they're your home run individuals. And they will bring data literacy across the organization. >>I love it. And I'm not going to try and be Mozart, but you know, three and a half years, I think you said two times, 18 months, uh, maybe there's hope for me yet in a singing, you'll be a good singer. Um, Duchy on the, on the, some of the sports references you've made, uh, Paul McGuire, we first connected, uh, I'm not gonna like disclose where you're from, but, uh, I saw he did come up and I know it all sorts of sports that drive to measure everything they can right on the field of the field. So let's imagine that you've done the best analysis, right? You're the most advanced data scientists schooled in the classics, as well as the modernist methods, the best tools you've made a beautiful analysis, beautiful dashboards. And now your coach just wants to put their favorite player on the game, despite what you're building to them. How do you deal with that kind of coaches? >>Yeah. Listen, this is a great question. I think for your data analytics strategy, but also for anyone listening and watching, who wants to just figure out how to drive a career forward? I would give the same advice. So the story you're talking about, indeed hockey, you can figure out where I'm from, but it's around the Ottawa senators, general manager. And he made a quote in an interview and he said, sometimes I want to punch my analytics, people in the head. Now I'm going to tell you, that's not a good culture for analytics. And he goes on to say, they tell me not to play this one player. This one player is very tough. You know, throws four or five hits a game. And he goes, I'd love my analytics people to get hit by bore a wacky and tell me how it feels. That's the player. >>Sure. I'm sure he hits hard, but here's the deal. When he's on the ice, the opposing team gets more shots on goal than the senators do on the opposing team. They score more goals, they lose. And so I think whenever you're trying to convince a movement forward, be it management, be it a project you're trying to fund. I always try to teach something that someone didn't previously know before and make them think, well, I never thought of it that way before. And I think the great opportunity right now, if you're trying to get moving in a data analytics strategy is around this post COVID era. You know, we've seen post COVID now really accelerate, or at least post COVID in certain parts of the world, but accelerate the appetite for digital transformation by about half a decade. Okay. And getting the data within your systems, as you digitize will give you all kinds of types of projects to make people think differently than the way they thought before. >>About data. I call this data exhaust. I'll give you a great example, Uber. I think we're all familiar with Uber. If we all remember back in the days when Uber would offer you search pricing. Okay? So basically you put Uber on your phone, they know everything about you, right? Who are your friends, where you going, uh, even how much batteries on your phone? Well, in a data science paper, I read a long time ago. They recognize that there was a 70% chance that you would accept a surge price. If you had less than 10% of your battery. So 10% of battery on your phone is an example of data exhaust all the lawns that you generate on your digital front end properties. Those are logs. You can take those together and maybe show executive management with data. We can understand why people abandoned their cart at the shipping phase, or what is the amount of shipping, which they abandoned it. When is the signal when our systems are about to go to go down. So, uh, I think that's a tremendous way. And if you look back to the sports, I mean the Atlanta Falcons NFL team, and they monitor their athletes, sleep performance, the Toronto Raptors basketball, they're running AI analytics on people's personalities and everything they tweet and every interview to see if the personality fits. So in sports, I think athletes are the most important commodity, if you will, or asset a yet all these teams are investing in analytics. So I think that's pretty telling, >>Okay, Paul, it looks like we're almost out of time. So in 30 seconds or less, what would you recommend to the data citizens out there? >>Okay. I'm going to give you a four tips in 30 seconds. Number one, remember learning never ends be curious forever. You'll drive your career. Number two, remember companies that invest in analytics and data, citizens outperform those that don't McKinsey says it's about 1.4 times across many KPIs. Number three, stop just collecting the dots and start connecting them with that. You need a strong governance strategy and that's going to help you for the future because the biggest thing in the future is not going to be about analytics, accuracy. It's going to be about analytics, explainability. So accuracy is no longer going to be enough. You're going to have to explain your decisions and finally stay positive and forever test negative. >>Love it. Thank you very much fall. Um, and for all the data seasons is out there. Um, when it comes down to access to data, it's more than just getting your hands on the data. It's also knowing what you can do with it, how you can do that and what you definitely shouldn't be doing with it. Uh, thank you everyone out there and enjoy your learning and interaction with the community. Stay healthy. Bye-bye.

(gentle music) >> From the Cube Studios in Palo Alto and Boston, connecting with thought leaders all around the world. This is theCube Conversation. >> Hi everybody, this is Dave Vellante, we're back at Cube on Cloud, and with me is Muddu Sudhakar. He's a long time alum of theCube, a technologist and executive, a serial entrepreneur and an investor. Welcome my friend, good to see you. >> Good to see you, Dave. Pleasure to be with you. Happy elections, I guess. >> Yeah, yeah. So I wanted to start, this work from home, pivot's been amazing, and you've seen the enterprise collaboration explode. I wrote a piece a couple months ago, looking at valuations of various companies, right around the snowflake IPO, I want to ask you about that, but I was looking at the valuations of various companies, at Spotify, and Shopify, and of course Zoom was there. And I was looking at just simple revenue multiples, and I said, geez, Zoom actually looks, might look undervalued, which is crazy, right? And of course the stock went up after that, and you see teams, Microsoft Teams, and Microsoft doing a great job across the board, we've written about that, you're seeing Webex is exploding, I mean, what do you make of this whole enterprise collaboration play? >> No, I think the look there is a trend here, right? So I think this probably trend started before COVID, but COVID is going to probably accelerate this whole digital transformation, right? People are going to work remotely a lot more, not everybody's going to come back to the offices even after COVID, so I think this whole collaboration through Slack, and Zoom, and Microsoft Teams and Webex, it's going to be the new game now, right? Both the video, audio and chat solutions, that's really going to help people like eyeballs. You're not going to spend time on all four of them, right? It's like everyday from a consumer side, you're going to spend time on your Gmail, Facebook, maybe Twitter, maybe Instagram, so like in the consumer side, on your personal life, you have something on the enterprise. The eyeballs are going to be in these platforms. >> Yeah. Well. >> But we're not going to take everything. >> Well, So you are right, there's a permanence to this, and I got a lot of ground to cover with you. And I always like our conversations mood because you tell it like it is, I'm going to stay on that work from home pivot. You know a lot about security, but you've seen three big trends, like mega trends in security, Endpoint, Identity Access Management, and Cloud Security, you're seeing this in the stock prices of companies like CrowdStrike, Zscaler, Okta- >> Right >> Sailpoint- >> Right, I mean, they exploded, as a result of the pandemic, and I think I'm inferring from your comment that you see that as permanent, but that's a real challenge from a security standpoint. What's the impact of Cloud there? >> No, it isn't impact but look, first is all the services required to be Cloud, right? See, the whole ideas for it to collaborate and do these things. So you cannot be running an application, like you can't be running conference and SharePoint oN-Prem, and try to on a Zoom and MS teams. So that's why, if you look at Microsoft is very clever, they went with Office 365, SharePoint 365, now they have MS Teams, so I think that Cloud is going to drive all these workloads that you have been talking about a lot, right? You and John have been saying this for years now. The eruption of Cloud and SAS services are the vehicle to drive this next-generation collaboration. >> You know what's so cool? So Cloud obviously is the topic, I wonder how you look at the last 10 years of Cloud, and maybe we could project forward, I mean the big three Cloud vendors, they're running it like $20 billion a quarter, and they're growing collectively, 35, 40% clips, so we're really approaching a hundred billion dollars for these three. And you hear stats like only 20% of the workloads are in the public Cloud, so it feels like we're just getting started. How do you look at the impact of Cloud on the market, as you say, the last 10 years, and what do you expect going forward? >> No, I think it's very fascinating, right? So I remember when theCube, you guys are talking about 10 years back, now it's been what? More than 10 years, 15 years, since AWS came out with their first S3 service back in 2006. >> Right. >> Right? so I think look, Cloud is going to accelerate even more further. The areas is going to accelerate is for different reasons. I think now you're seeing the initial days, it's all about startups, initial workloads, Dev test and QA test, now you're talking about real production workloads are moving towards Cloud, right? Initially it was backup, we really didn't care for backup they really put there. Now you're going to have Cloud health primary services, your primary storage will be there, it's not going to be an EMC, It's not going to be a NetApp storage, right? So workloads are going to shift from the business applications, and these business applications, will be running on the Cloud, and I'll make another prediction, make customer service and support. Customer service and support, again, we should be running on the Cloud. You're not want to run the thing on a Dell server, or an IBM server, or an HP server, with your own hosted environment. That model is not because there's no economies of scale. So to your point, what will drive Cloud for the next 10 years, will be economies of scale. Where can you take the cost? How can I save money? If you don't move to the Cloud, you won't save money. So all those workloads are going to go to the Cloud are people who really want to save, like global gradual custom, right? If you stay on the ASP model, a hosted, you're not going to save your costs, your costs will constantly go up from a SaaS perspective. >> So that doesn't bode well for all the On-prem guys, and you hear a lot of the vendors that don't own a Cloud that talk about repatriation, but the numbers don't support that. So what do those guys do? I mean, they're talking multi-Cloud, of course they're talking hybrid, that's IBM's big play, how do you see it? >> I think, look, see there, to me, multi-Cloud makes sense, right? You don't want one vendor that you never want to get, so having Amazon, Microsoft, Google, it gives them a multi-Cloud. Even hybrid Cloud does make sense, right? There'll be some workloads. It's like, we are still running On-prem environment, we still have mainframe, so it's never going to be a hundred percent, but I would say the majority, your question is, can we get to 60, 70, 80% workers in the next 10 years? I think you will. I think by 2025, more than 78% of the Cloud Migration by the next five years, 70% of workload for enterprise will be on the Cloud. The remaining 25, maybe Hybrid, maybe On-prem, but I get panics, really doesn't matter. You have saved and part of your business is running on the Cloud. That's your cost saving, that's where you'll see the economies of scale, and that's where all the growth will happen. >> So square the circle for me, because again, you hear the stat on the IDC stat, IBM Ginni Rometty puts it out there a lot that only 20% of the workloads are in the public Cloud, everything else is On-prem, but it's not a zero sum game, right? I mean the Cloud native stuff is growing like crazy, the On-prem stuff is flat to down, so what's going to happen? When you talk about 70% of the workloads will be in the Cloud, do you see those mission critical apps and moving into the car, I mean the insurance companies going to put their claims apps in the Cloud, or the financial services companies going to put their mission critical workloads in the Cloud, or they just going to develop new stuff that's Cloud native that is sort of interacts with the On-prem. How do you see that playing out? >> Yeah, no, I think absolutely, I think a very good question. So two things will happen. I think if you take an enterprise, right? Most businesses what they'll do is the workloads that they should not be running On-prem, they'll move it up. So obviously things like take, as I said, I use the word SharePoint, right? SharePoint and conference, all the knowledge stuff is still running on people's data centers. There's no reason. I understand, I've seen statistics that 70, 80% of the On-prem for SharePoint will move to SharePoint on the Cloud. So Microsoft is going to make tons of money on that, right? Same thing, databases, right? Whether it's CQL server, whether there is Oracle database, things that you are running as a database, as a Cloud, we move to the Cloud. Whether that is posted in Oracle Cloud, or you're running Oracle or Mongo DB, or Dynamo DB on AWS or SQL server Microsoft, that's going to happen. Then what you're talking about is really the App concept, the applications themselves, the App server. Is the App server is going to run On-prem, how much it's going to laureate outside? There may be a hybrid Cloud, like for example, Kafka. I may use a Purse running on a Kafka as a service, or I may be using Elasticsearch for my indexing on AWS or Google Cloud, but I may be running my App locally. So there'll be some hybrid place, but what I would say is for every application, 75% of your Comprende will be on the Cloud. So think of it like the Dev. So even for the On-prem app, you're not going to be a 100 percent On-prem. The competent, the billing materials will move to the Cloud, your Purse, your storage, because if you put it On-prem, you need to add all this, you need to have all the whole things to buy it and hire the people, so that's what is going to happen. So from a competent perspective, 70% of your bill of materials will move to the Cloud, even for an On-prem application. >> So, Of course, the susification of the industry in the last decade and in my three favorite companies last decade, you've worked for two of them, Tableau, ServiceNow, and Splunk. I want to ask you about those, but I'm interested in the potential disruption there. I mean, you've got these SAS companies, Salesforce of course is another one, but they can't get started in 1999. What do you see happening with those? I mean, we're basically building these sort of large SAS, platforms, now. Do you think that the Cloud native world that developers can come at this from an angle where they can disrupt those companies, or are they too entrenched? I mean, look at service now, I mean, I don't know, $80 billion market capital where they are, they bigger than Workday, I mean, just amazing how much they've grown and you feel like, okay, nothing can stop them, but there's always disruption in this industry, what are your thoughts on that. >> Not very good with, I think there'll be disrupted. So to me actually to your point, ServiceNow is now close to a 100 billion now, 95 billion market coverage, crazy. So from evaluation perspective, so I think the reason they'll be disrupted is that the SAS vendors that you talked about, ServiceNow, and all this plan, most of these services, they're truly not a multi-tenant or what do you call the Cloud Native. And that is the Accenture. So because of that, they will not be able to pass the savings back to the enterprises. So the cost economics, the economics that the Cloud provides because of the multi tenancy ability will not. The second reason there'll be disrupted is AI. So far, we talked about Cloud, but AI is the core. So it's not really Cloud Native, Dave, I look at the AI in a two-piece. AI is going to change, see all the SAS vendors were created 20 years back, if you remember, was an operator typing it, I don't respond administered we'll type a Splunk query. I don't need a human to type a query anymore, system will actually find it, that's what the whole security game has changed, right? So what's going to happen is if you believe in that, that AI, your score will disrupt all the SAS vendors, so one angle SAS is going to have is a Cloud. That's where you make the Cloud will take up because a SAS application will be Cloudified. Being SAS is not Cloud, right? Second thing is SAS will be also, I call it, will be AI-fied. So AI and machine learning will be trying to drive at the core so that I don't need that many licenses. I don't need that many humans. I don't need that many administrators to manage, I call them the tuners. Once you get a driverless car, you don't need a thousand tuners to tune your Tesla, or Google Waymo car. So the same philosophy will happen is your Dev Apps, your administrators, your service management, people that you need for service now, and these products, Zendesk with AI, will tremendously will disrupt. >> So you're saying, okay, so yeah, I was going to ask you, won't the SAS vendors, won't they be able to just put, inject AI into their platforms, and I guess I'm inferring saying, yeah, but a lot of the problems that they're solving, are going to go away because of AI, is that right? And automation and RPA and things of that nature, is that right? >> Yes and no. So I'll tell you what, sorry, you have asked a very good question, let's answer, let me rephrase that question. What you're saying is, "Why can't the existing SAS vendors do the AI?" >> Yes, right. >> Right, >> And there's a reason they can't do it is their pricing model is by number of seats. So I'm not going to come to Dave, and say, come on, come pay me less money. It's the same reason why a board and general lover build an electric car. They're selling 10 million gasoline cars. There's no incentive for me, I'm not going to do any AI, I'm going to put, I'm not going to come to you and say, hey, buy me a hundred less license next year from it. So that is one reason why AI, even though these guys do any AI, it's going to be just so I call it, they're going to, what do you call it, a whitewash, kind of like you put some paint brush on it, trying to show you some AI you did from a marketing dynamics. But at the core, if you really implement the AI with you take the driver out, how are you going to change the pricing model? And being a public company, you got to take a hit on the pricing model and the price, and it's going to have a stocking part. So that, to your earlier question, will somebody disrupt them? The person who is going to disrupt them, will disrupt them on the pricing model. >> Right. So I want to ask you about that, because we saw a Snowflake, and it's IPO, we were able to pour through its S-1, and they have a different pricing model. It's a true Cloud consumption model, Whereas of course, most SAS companies, they're going to lock you in for at least one year term, maybe more, and then, you buy the license, you got to pay X. If you, don't use it, you still got to pay for it. Snowflake's different, actually they have a different problem, that people are using it too much and the sea is driving the CFO crazy because the bill is going up and up and up, but to me, that's the right model, It's just like the Amazon model, if you can justify it, so how do you see the pricing, that consumption model is actually, you're seeing some of the On-prem guys at HPE, Dell, they're doing as a service. They're kind of taking a page out of the last decade SAS model, so I think pricing is a real tricky one, isn't it? >> No, you nailed it, you nailed it. So I think the way in which the Snowflake there, how the disruptors are data warehouse, that disrupted the open source vendors too. Snowflake distributed, imagine the playbook, you disrupted something as the $ 0, right? It's an open source with Cloudera, Hortonworks, Mapper, that whole big data that you want me to, or that market is this, that disrupting data warehouses like Netezza, Teradata, and the charging more money, they're making more money and disrupting at $0, because the pricing models by consumption that you talked about. CMT is going to happen in the service now, Zen Desk, well, 'cause their pricing one is by number of seats. People are going to say, "How are my users are going to ask?" right? If you're an employee help desk, you're back to your original health collaborative. I may be on Slack, I could be on zoom, I'll maybe on MS Teams, I'm going to ask by using usage model on Slack, tools by employees to service now is the pricing model that people want to pay for. The more my employees use it, the more value I get. But I don't want to pay by number of seats, so the vendor, who's going to figure that out, and that's where I look, if you know me, I'm right over as I started, that's what I've tried to push that model look, I love that because that's the core of how you want to change the new game. >> I agree. I say, kill me with that problem, I mean, some people are trying to make it a criticism, but you hit on the point. If you pay more, it's only because you're getting more value out of it. So I wanted to flip the switch here a little bit and take a customer angle. Something that you've been on all sides. And I want to talk a little bit about strategies, you've been a strategist, I guess, once a strategist, always a strategist. How should organizations be thinking about their approach to Cloud, it's cost different for different industries, but, back when the cube started, financial services Cloud was a four-letter word. But of course the age of company is going to matter, but what's the framework for figuring out your Cloud strategy to get to your 70% and really take advantage of the economics? Should I be Mono Cloud, Multi-Cloud, Multi-vendor, what would you advise? >> Yeah, no, I mean, I mean, I actually call it the tech stack. Actually you and John taught me that what was the tech stack, like the lamp stack, I think there is a new Cloud stack needs to come, and that I think the bottomline there should be... First of all, anything with storage should be in the Cloud. I mean, if you want to start, whether you are, financial, doesn't matter, there's no way. I come from cybersecurity side, I've seen it. Your attackers will be more with insiders than being on the Cloud, so storage has to be in the Cloud then come compute, Kubernetes. If you really want to use containers and Kubernetes, it has to be in the public Cloud, leverage that have the computer on their databases. That's where it can be like if your data is so strong, maybe run it On-prem, maybe have it on a hosted model for when it comes to database, but there you have a choice between hybrid Cloud and public Cloud choice. Then on top when it comes to App, the app itself, you can run locally or anywhere, the App and database. Now the areas that you really want to go after to migrate is look at anything that's an enterprise workload that you don't need people to manage it. You want your own team to move up in the career. You don't want thousand people looking at... you don't want to have a, for example, IT administrators to call central people to the people to manage your compute storage. That workload should be more, right? You already saw Sierra moved out to Salesforce. We saw collaboration already moved out. Zoom is not running locally. You already saw SharePoint with knowledge management mode up, right? With a box, drawbacks, you name anything. The next global mode is a SAS workloads, right? I think Workday service running there, but work data will go into the Cloud. I bet at some point Zendesk, ServiceNow, then either they put it on the public Cloud, or they have to create a product and public Cloud. To your point, these public Cloud vendors are at $2 trillion market cap. They're they're bigger than the... I call them nation States. >> Yeah, >> So I'm servicing though. I mean, there's a 2 trillion market gap between Amazon and Azure, I'm not going to compete with them. So I want to take this workload to run it there. So all these vendors, if you see that's where Shandra from Adobe is pushing this right, Adobe, Workday, Anaplan, all the SAS vendors we'll move them into the public Cloud within these vendors. So those workloads need to move out, right? So that all those things will start, then you'll start migrating, but I call your procurement. That's where the RPA comes in. The other thing that we didn't talk about, back to your first question, what is the next 10 years of Cloud will be RPA? That third piece to Cloud is RPA because if you have your systems On-prem, I can't automate them. I have to do a VPN into your house there and then try to automate your systems, or your procurement, et cetera. So all these RPA vendors are still running On-prem, most of them, whether it's UI path automation anywhere. So the Cloud should be where the brain should be. That's what I call them like the octopus analogy, the brain is in the Cloud, the tentacles are everywhere, they should manage it. But if my tentacles have to do a VPN with your house to manage it, I'm always will have failures. So if you look at the why RPA did not have the growth, like the Snowflake, like the Cloud, because they are running it On-prem, most of them still. 80% of the RP revenue is On-prem, running On-prem, that needs to be called clarified. So AI, RPA and the SAS, are the three reasons Cloud will take off. >> Awesome. Thank you for that. Now I want to flip the switch again. You're an investor or a multi-tool player here, but so if you're, let's say you're an ecosystem player, and you're kind of looking at the landscape as you're in an investor, of course you've invested in the Cloud, because the Cloud is where it's at, but you got to be careful as an ecosystem player to pick a spot that both provides growth, but allows you to have a moat as, I mean, that's why I'm really curious to see how Snowflake's going to compete because they're competing with AWS, Microsoft, and Google, unlike, Frank, when he was at service now, he was competing with BMC and with on-prem and he crushed it, but the competitors are much more capable here, but it seems like they've got, maybe they've got a moat with MultiCloud, and that whole data sharing thing, we'll see. But, what about that? Where are the opportunities? Where's that white space? And I know there's a lot of white space, but what's the framework to look at, from an investor standpoint, or even a CEO standpoint, where you want to put place your bets. >> No, very good question, so look, I did something. We talk as an investor in the board with many companies, right? So one thing that says as an investor, if you come back and say, I want to create a next generation Docker or a computer, there's no way nobody's going to invest. So that we can motor off, even if you want to do object storage or a block storage, I mean, I've been an investor board member of so many storage companies, there's no way as an industry, I'll write a check for a compute or storage, right? If you want to create a next generation network, like either NetSuite, or restart Juniper, Cisco, there is no way. But if you come back and say, I want to create a next generation Viper for remote working environments, where AI is at the core, I'm interested in that, right? So if you look at how the packets are dropped, there's no intelligence in either not switching today. The packets come, I do it. The intelligence is not built into the network with AI level. So if somebody comes with an AI, what good is all this NVD, our GPS, et cetera, if you cannot do wire speed, packet inspection, looking at the content and then route the traffic. If I see if it's a video package, but in UN Boston, there's high interview day of they should be loading our package faster, because you are a premium ISP. That intelligence has not gone there. So you will see, and that will be a bad people will happen in the network, switching, et cetera, right? So that is still an angle. But if you work and it comes to platform services, remember when I was at Pivotal and VMware, all models was my boss, that would, yes, as a platform, service is a game already won by the Cloud guys. >> Right. (indistinct) >> Silicon Valley Investors, I don't think you want to invest in past services, right? I mean, you might come with some lecture edition database to do some updates, there could be some game, let's say we want to do a time series database, or some metrics database, there's always some small angle, but the opportunity to go create a national database there it's very few. So I'm kind of eliminating all the black spaces, right? >> Yeah. >> We have the white spaces that comes in is the SAS level. Now to your point, if I'm Amazon, I'm going to compete with Snowflake, I have Redshift. So this is where at some point, these Cloud platforms, I call them aircraft carriers. They're not going to stay on the aircraft carriers, they're going to own the land as well. So they're going to move up to the SAS space. The question is you want to create a SAS service like CRM. They are not going to create a CRM like service, they may not create a sales force and service now, but if you're going to add a data warehouse, I can very well see Azure, Google, and AWS, going to create something to compute a Snowflake. Why would I not? It's so close to my database and data warehouse, I already have Redshift. So that's going to be nightlights, same reason, If you look at Netflix, you have a Netflix and you have Amazon prime. Netflix runs on Amazon, but you have Amazon prime. So you have the same model, you have Snowflake, and you'll have Redshift. The both will help each other, there'll be a... What do you call it? Coexistence will happen. But if you really want to invest, you want to invest in SAS companies. You do not want to be investing in a compliment players. You don't want to a feature. >> Yeah, that's great, I appreciate that perspective. And I wonder, so obviously Microsoft play in SAS, Google's got G suite. And I wonder if people often ask the Andy Jassy, you're going to move up the stack, you got to be an application, a SAS vendor, and you never say never with Atavist, But I wonder, and we were talking to Jerry Chen about this, years ago on theCube, and his angle was that Amazon will play, but they'll play through developers. They'll enable developers, and they'll participate, they'll take their, lick off the cone. So it's going to be interesting to see how directly Amazon plays, but at some point you got Tam expansion, you got to play in that space. >> Yeah, I'll give you an example of knowing, I got acquired by a couple of times by EMC. So I learned a lot from Joe Tucci and Paul Merage over the years. see Paul and Joe, what they did is to look at how 20 years, and they are very close to Boston in your area, Joe, what games did is they used to sell storage, but you know what he did, he went and bought the Apps to drive them. He bought like Legato, he bought Documentum, he bought Captiva, if you remember how he acquired all these companies as a services, he bought VMware to drive that. So I think the good angle that Microsoft has is, I'm a SAS player, I have dynamics, I have CRM, I have SharePoint, I have Collaboration, I have Office 365, MS Teams for users, and then I have the platform as Azure. So I think if I'm Amazon, (indistinct). I got to own the apps so that I can drive this workforce on my platform. >> Interesting. >> Just going to developers, like I know Jerry Chan, he was my peer a BMF. I don't think just literally to developers and that model works in open source, but the open source game is pretty much gone, and not too many companies made money. >> Well, >> Most companies pretty much gone. >> Yeah, he's right. Red hats not bad idea. But it's very interesting what you're saying there. And so, hey, its why Oracle wants to have Tiktok, running on their platform, right? I mean, it's going to. (laughing) It's going to drive that further integration. I wanted to ask you something, you were talking about, you wouldn't invest in storage or compute, but I wonder, and you mentioned some commentary about GPU's. Of course the videos has been going crazy, but they're now saying, okay, how do we expand our Team, they make the acquisition of arm, et cetera. What about this DPU thing, if you follow that, that data processing unit where they're like hyper dis-aggregation and then they reaggregate, and as an offload and really to drive data centric workloads. Have you looked at that at all? >> I did, I think, and that's a good angle. So I think, look, it's like, it goes through it. I don't know if you remember in your career, we have seen it. I used to get Silicon graphics. I saw the first graphic GPU, right? That time GPU was more graphic processor unit, >> Right, yeah, work stations. >> So then become NPUs at work processing units, right? There was a TCP/IP office offloading, if you remember right, there was like vector processing unit. So I think every once in a while the industry, recreated this separate unit, as a co-processor to the main CPU, because main CPU's inefficient, and it makes sense. And then Google created TPU's and then we have the new world of the media GPU's, now we have DPS all these are good, but what's happening is, all these are driving for machine learning, AI for the training period there. Training period Sometimes it's so long with the workloads, if you can cut down, it makes sense. >> Yeah. >> Because, but the question is, these aren't so specialized in nature. I can't use it for everything. >> Yup. >> I want Ideally, algorithms to be paralyzed, I want the training to be paralyzed, I want so having deep use and GPS are important, I think where I want to see them as more, the algorithm, there should be more investment from the NVIDIA's and these guys, taking the algorithm to be highly paralyzed them. (indistinct) And I think that still has not happened in industry yet. >> All right, so we're pretty much out of time, but what are you doing these days? Where are you spending your time, are you still in Stealth, give us a little glimpse. >> Yeah, no, I'm out of the Stealth, I'm actually the CEO of Aisera now, Aisera, obviously I invested with them, but I'm the CEO of Aisero. It's funded by Menlo ventures, Norwest, True, along with Khosla ventures and Ram Shriram is a big investor. Robin's on the board of Google, so these guys, look, we are going out to the collaboration game. How do you automate customer service and support for employees and then users, right? In this whole game, we talked about the Zoom, Slack and MS Teams, that's what I'm spending time, I want to create next generation service now. >> Fantastic. Muddu, I always love having you on you, pull punches, you tell it like it is, that you're a great visionary technologist. Thanks so much for coming on theCube, and participating in our program. >> Dave, it's always a pleasure speaking to you sir. Thank you. >> Okay. Keep it right there, there's more coming from Cuba and Cloud right after this break. (slow music)

(gentle music) >> From the Cube Studios in Palo Alto and Boston, connecting with thought leaders all around the world. This is theCube Conversation. >> Hi everybody, this is Dave Vellante, we're back at Cube on Cloud, and with me is Muddu Sudhakar. He's a long time alum of theCube, a technologist and executive, a serial entrepreneur and an investor. Welcome my friend, good to see you. >> Good to see you, Dave. Pleasure to be with you. Happy elections, I guess. >> Yeah, yeah. So I wanted to start, this work from home, pivot's been amazing, and you've seen the enterprise collaboration explode. I wrote a piece a couple months ago, looking at valuations of various companies, right around the snowflake IPO, I want to ask you about that, but I was looking at the valuations of various companies, at Spotify, and Shopify, and of course Zoom was there. And I was looking at just simple revenue multiples, and I said, geez, Zoom actually looks, might look undervalued, which is crazy, right? And of course the stock went up after that, and you see teams, Microsoft Teams, and Microsoft doing a great job across the board, we've written about that, you're seeing Webex is exploding, I mean, what do you make of this whole enterprise collaboration play? >> No, I think the look there is a trend here, right? So I think this probably trend started before COVID, but COVID is going to probably accelerate this whole digital transformation, right? People are going to work remotely a lot more, not everybody's going to come back to the offices even after COVID, so I think this whole collaboration through Slack, and Zoom, and Microsoft Teams and Webex, it's going to be the new game now, right? Both the video, audio and chat solutions, that's really going to help people like eyeballs. You're not going to spend time on all four of them, right? It's like everyday from a consumer side, you're going to spend time on your Gmail, Facebook, maybe Twitter, maybe Instagram, so like in the consumer side, on your personal life, you have something on the enterprise. The eyeballs are going to be in these platforms. >> Yeah. Well. >> But we're not going to take everything. >> Well, So you are right, there's a permanence to this, and I got a lot of ground to cover with you. And I always like our conversations mood because you tell it like it is, I'm going to stay on that work from home pivot. You know a lot about security, but you've seen three big trends, like mega trends in security, Endpoint, Identity Access Management, and Cloud Security, you're seeing this in the stock prices of companies like CrowdStrike, Zscaler, Okta- >> Right >> Sailpoint- >> Right, I mean, they exploded, as a result of the pandemic, and I think I'm inferring from your comment that you see that as permanent, but that's a real challenge from a security standpoint. What's the impact of Cloud there? >> No, it isn't impact but look, first is all the services required to be Cloud, right? See, the whole ideas for it to collaborate and do these things. So you cannot be running an application, like you can't be running conference and SharePoint oN-Prem, and try to on a Zoom and MS teams. So that's why, if you look at Microsoft is very clever, they went with Office 365, SharePoint 365, now they have MS Teams, so I think that Cloud is going to drive all these workloads that you have been talking about a lot, right? You and John have been saying this for years now. The eruption of Cloud and SAS services are the vehicle to drive this next-generation collaboration. >> You know what's so cool? So Cloud obviously is the topic, I wonder how you look at the last 10 years of Cloud, and maybe we could project forward, I mean the big three Cloud vendors, they're running it like $20 billion a quarter, and they're growing collectively, 35, 40% clips, so we're really approaching a hundred billion dollars for these three. And you hear stats like only 20% of the workloads are in the public Cloud, so it feels like we're just getting started. How do you look at the impact of Cloud on the market, as you say, the last 10 years, and what do you expect going forward? >> No, I think it's very fascinating, right? So I remember when theCube, you guys are talking about 10 years back, now it's been what? More than 10 years, 15 years, since AWS came out with their first S3 service back in 2006. >> Right. >> Right? so I think look, Cloud is going to accelerate even more further. The areas is going to accelerate is for different reasons. I think now you're seeing the initial days, it's all about startups, initial workloads, Dev test and QA test, now you're talking about real production workloads are moving towards Cloud, right? Initially it was backup, we really didn't care for backup they really put there. Now you're going to have Cloud health primary services, your primary storage will be there, it's not going to be an EMC, It's not going to be a ETAP storage, right? So workloads are going to shift from the business applications, and this business App again, will be running on the Cloud, and I'll make another prediction, make customer service and support. Customer service and support, again, we should be running on the Cloud. You're not want to run the thing on a Dell server, or an IBM server, or an HP server, with your own hosted environment. That model is not because there's no economies of scale. So to your point, what will drive Cloud for the next 10 years, will be economies of scale. Where can you take the cost? How can I save money? If you don't move to the Cloud, you won't save money. So all those workloads are going to go to the Cloud are people who really want to save, like global gradual custom, right? If you stay on the ASP model, a hosted, you're not going to save your costs, your costs will constantly go up from a SAS perspective. >> So that doesn't bode well for all the On-prem guys, and you hear a lot of the vendors that don't own a Cloud that talk about repatriation, but the numbers don't support that. So what do those guys do? I mean, they're talking multi-Cloud, of course they're talking hybrid, that's IBM's big play, how do you see it? >> I think, look, see there, to me, multi-Cloud makes sense, right? You don't want one vendor that you never want to get, so having Amazon, Microsoft, Google, it gives them a multi-Cloud. Even hybrid Cloud does make sense, right? There'll be some workloads. It's like, we are still running On-prem environment, we still have mainframe, so it's never going to be a hundred percent, but I would say the majority, your question is, can we get to 60, 70, 80% workers in the next 10 years? I think you will. I think by 2025, more than 78% of the Cloud Migration by the next five years, 70% of workload for enterprise will be on the Cloud. The remaining 25, maybe Hybrid, maybe On-prem, but I get panics, really doesn't matter. You have saved and part of your business is running on the Cloud. That's your cost saving, that's where you'll see the economies of scale, and that's where all the growth will happen. >> So square the circle for me, because again, you hear the stat on the IDC stat, IBM Ginni Rometty puts it out there a lot that only 20% of the workloads are in the public Cloud, everything else is On-prem, but it's not a zero sum game, right? I mean the Cloud native stuff is growing like crazy, the On-prem stuff is flat to down, so what's going to happen? When you talk about 70% of the workloads will be in the Cloud, do you see those mission critical apps and moving into the car, I mean the insurance companies going to put their claims apps in the Cloud, or the financial services companies going to put their mission critical workloads in the Cloud, or they just going to develop new stuff that's Cloud native that is sort of interacts with the On-prem. How do you see that playing out? >> Yeah, no, I think absolutely, I think a very good question. So two things will happen. I think if you take an enterprise, right? Most businesses what they'll do is the workloads that they should not be running On-prem, they'll move it up. So obviously things like take, as I said, I use the word SharePoint, right? SharePoint and conference, all the knowledge stuff is still running on people's data centers. There's no reason. I understand, I've seen statistics that 70, 80% of the On-prem for SharePoint will move to SharePoint on the Cloud. So Microsoft is going to make tons of money on that, right? Same thing, databases, right? Whether it's CQL server, whether there is Oracle database, things that you are running as a database, as a Cloud, we move to the Cloud. Whether that is posted in Oracle Cloud, or you're running Oracle or Mongo DB, or Dynamo DB on AWS or SQL server Microsoft, that's going to happen. Then what you're talking about is really the App concept, the applications themselves, the App server. Is the App server is going to run On-prem, how much it's going to laureate outside? There may be a hybrid Cloud, like for example, Kafka. I may use a Purse running on a Kafka as a service, or I may be using Elasticsearch for my indexing on AWS or Google Cloud, but I may be running my App locally. So there'll be some hybrid place, but what I would say is for every application, 75% of your Comprende will be on the Cloud. So think of it like the Dev. So even for the On-prem app, you're not going to be a 100 percent On-prem. The competent, the billing materials will move to the Cloud, your Purse, your storage, because if you put it On-prem, you need to add all this, you need to have all the whole things to buy it and hire the people, so that's what is going to happen. So from a competent perspective, 70% of your bill of materials will move to the Cloud, even for an On-prem application. >> So, Of course, the susification of the industry in the last decade and in my three favorite companies last decade, you've worked for two of them, Tableau, ServiceNow, and Splunk. I want to ask you about those, but I'm interested in the potential disruption there. I mean, you've got these SAS companies, Salesforce of course is another one, but they can't get started in 1999. What do you see happening with those? I mean, we're basically building these sort of large SAS, platforms, now. Do you think that the Cloud native world that developers can come at this from an angle where they can disrupt those companies, or are they too entrenched? I mean, look at service now, I mean, I don't know, $80 billion market capital where they are, they bigger than Workday, I mean, just amazing how much they've grown and you feel like, okay, nothing can stop them, but there's always disruption in this industry, what are your thoughts on that. >> Not very good with, I think there'll be disrupted. So to me actually to your point, ServiceNow is now close to a 100 billion now, 95 billion market coverage, crazy. So from evaluation perspective, so I think the reason they'll be disrupted is that the SAS vendors that you talked about, ServiceNow, and all this plan, most of these services, they're truly not a multi-tenant or what do you call the Cloud Native. And that is the Accenture. So because of that, they will not be able to pass the savings back to the enterprises. So the cost economics, the economics that the Cloud provides because of the multi tenancy ability will not. The second reason there'll be disrupted is AI. So far, we talked about Cloud, but AI is the core. So it's not really Cloud Native, Dave, I look at the AI in a two-piece. AI is going to change, see all the SAS vendors were created 20 years back, if you remember, was an operator typing it, I don't respond administered we'll type a Splunk query. I don't need a human to type a query anymore, system will actually find it, that's what the whole security game has changed, right? So what's going to happen is if you believe in that, that AI, your score will disrupt all the SAS vendors, so one angle SAS is going to have is a Cloud. That's where you make the Cloud will take up because a SAS application will be Cloudified. Being SAS is not Cloud, right? Second thing is SAS will be also, I call it, will be AI-fied. So AI and machine learning will be trying to drive at the core so that I don't need that many licenses. I don't need that many humans. I don't need that many administrators to manage, I call them the tuners. Once you get a driverless car, you don't need a thousand tuners to tune your Tesla, or Google Waymo car. So the same philosophy will happen is your Dev Apps, your administrators, your service management, people that you need for service now, and these products, Zendesk with AI, will tremendously will disrupt. >> So you're saying, okay, so yeah, I was going to ask you, won't the SAS vendors, won't they be able to just put, inject AI into their platforms, and I guess I'm inferring saying, yeah, but a lot of the problems that they're solving, are going to go away because of AI, is that right? And automation and RPA and things of that nature, is that right? >> Yes and no. So I'll tell you what, sorry, you have asked a very good question, let's answer, let me rephrase that question. What you're saying is, "Why can't the existing SAS vendors do the AI?" >> Yes, right. >> Right, >> And there's a reason they can't do it is their pricing model is by number of seats. So I'm not going to come to Dave, and say, come on, come pay me less money. It's the same reason why a board and general lover build an electric car. They're selling 10 million gasoline cars. There's no incentive for me, I'm not going to do any AI, I'm going to put, I'm not going to come to you and say, hey, buy me a hundred less license next year from it. So that is one reason why AI, even though these guys do any AI, it's going to be just so I call it, they're going to, what do you call it, a whitewash, kind of like you put some paint brush on it, trying to show you some AI you did from a marketing dynamics. But at the core, if you really implement the AI with you take the driver out, how are you going to change the pricing model? And being a public company, you got to take a hit on the pricing model and the price, and it's going to have a stocking part. So that, to your earlier question, will somebody disrupt them? The person who is going to disrupt them, will disrupt them on the pricing model. >> Right. So I want to ask you about that, because we saw a Snowflake, and it's IPO, we were able to pour through its S-1, and they have a different pricing model. It's a true Cloud consumption model, Whereas of course, most SAS companies, they're going to lock you in for at least one year term, maybe more, and then, you buy the license, you got to pay X. If you, don't use it, you still got to pay for it. Snowflake's different, actually they have a different problem, that people are using it too much and the sea is driving the CFO crazy because the bill is going up and up and up, but to me, that's the right model, It's just like the Amazon model, if you can justify it, so how do you see the pricing, that consumption model is actually, you're seeing some of the On-prem guys at HPE, Dell, they're doing as a service. They're kind of taking a page out of the last decade SAS model, so I think pricing is a real tricky one, isn't it? >> No, you nailed it, you nailed it. So I think the way in which the Snowflake there, how the disruptors are data warehouse, that disrupted the open source vendors too. Snowflake distributed, imagine the playbook, you disrupted something as the $ 0, right? It's an open source with Cloudera, Hortonworks, Mapper, that whole big data that you want me to, or that market is this, that disrupting data warehouses like Netezza, Teradata, and the charging more money, they're making more money and disrupting at $0, because the pricing models by consumption that you talked about. CMT is going to happen in the service now, Zen Desk, well, 'cause their pricing one is by number of seats. People are going to say, "How are my users are going to ask?" right? If you're an employee help desk, you're back to your original health collaborative. I may be on Slack, I could be on zoom, I'll maybe on MS Teams, I'm going to ask by using usage model on Slack, tools by employees to service now is the pricing model that people want to pay for. The more my employees use it, the more value I get. But I don't want to pay by number of seats, so the vendor, who's going to figure that out, and that's where I look, if you know me, I'm right over as I started, that's what I've tried to push that model look, I love that because that's the core of how you want to change the new game. >> I agree. I say, kill me with that problem, I mean, some people are trying to make it a criticism, but you hit on the point. If you pay more, it's only because you're getting more value out of it. So I wanted to flip the switch here a little bit and take a customer angle. Something that you've been on all sides. And I want to talk a little bit about strategies, you've been a strategist, I guess, once a strategist, always a strategist. How should organizations be thinking about their approach to Cloud, it's cost different for different industries, but, back when the cube started, financial services Cloud was a four-letter word. But of course the age of company is going to matter, but what's the framework for figuring out your Cloud strategy to get to your 70% and really take advantage of the economics? Should I be Mono Cloud, Multi-Cloud, Multi-vendor, what would you advise? >> Yeah, no, I mean, I mean, I actually call it the tech stack. Actually you and John taught me that what was the tech stack, like the lamp stack, I think there is a new Cloud stack needs to come, and that I think the bottomline there should be... First of all, anything with storage should be in the Cloud. I mean, if you want to start, whether you are, financial, doesn't matter, there's no way. I come from cybersecurity side, I've seen it. Your attackers will be more with insiders than being on the Cloud, so storage has to be in the Cloud and encompass compute whoever it is. If you really want to use containers and Kubernetes, it has to be in the public Cloud, leverage that have the computer on their databases. That's where it can be like if your data is so strong, maybe run it On-prem, maybe have it on a hosted model for when it comes to database, but there you have a choice between hybrid Cloud and public Cloud choice. Then on top when it comes to App, the app itself, you can run locally or anywhere, the App and database. Now the areas that you really want to go after to migrate is look at anything that's an enterprise workload that you don't need people to manage it. You want your own team to move up in the career. You don't want thousand people looking at... you don't want to have a, for example, IT administrators to call central people to the people to manage your compute storage. That workload should be more, right? You already saw Sierra moved out to Salesforce. We saw collaboration already moved out. Zoom is not running locally. You already saw SharePoint with knowledge management mode up, right? With a box, drawbacks, you name anything. The next global mode is a SAS workloads, right? I think Workday service running there, but work data will go into the Cloud. I bet at some point Zendesk, ServiceNow, then either they put it on the public Cloud, or they have to create a product and public Cloud. To your point, these public Cloud vendors are at $2 trillion market cap. They're they're bigger than the... I call them nation States. >> Yeah, >> So I'm servicing though. I mean, there's a 2 trillion market gap between Amazon and Azure, I'm not going to compete with them. So I want to take this workload to run it there. So all these vendors, if you see that's where Shandra from Adobe is pushing this right, Adobe, Workday, Anaplan, all the SAS vendors we'll move them into the public Cloud within these vendors. So those workloads need to move out, right? So that all those things will start, then you'll start migrating, but I call your procurement. That's where the RPA comes in. The other thing that we didn't talk about, back to your first question, what is the next 10 years of Cloud will be RPA? That third piece to Cloud is RPA because if you have your systems On-prem, I can't automate them. I have to do a VPN into your house there and then try to automate your systems, or your procurement, et cetera. So all these RPA vendors are still running On-prem, most of them, whether it's UI path automation anywhere. So the Cloud should be where the brain should be. That's what I call them like the octopus analogy, the brain is in the Cloud, the tentacles are everywhere, they should manage it. But if my tentacles have to do a VPN with your house to manage it, I'm always will have failures. So if you look at the why RPA did not have the growth, like the Snowflake, like the Cloud, because they are running it On-prem, most of them still. 80% of the RP revenue is On-prem, running On-prem, that needs to be called clarified. So AI, RPA and the SAS, are the three reasons Cloud will take off. >> Awesome. Thank you for that. Now I want to flip the switch again. You're an investor or a multi-tool player here, but so if you're, let's say you're an ecosystem player, and you're kind of looking at the landscape as you're in an investor, of course you've invested in the Cloud, because the Cloud is where it's at, but you got to be careful as an ecosystem player to pick a spot that both provides growth, but allows you to have a moat as, I mean, that's why I'm really curious to see how Snowflake's going to compete because they're competing with AWS, Microsoft, and Google, unlike, Frank, when he was at service now, he was competing with BMC and with on-prem and he crushed it, but the competitors are much more capable here, but it seems like they've got, maybe they've got a moat with MultiCloud, and that whole data sharing thing, we'll see. But, what about that? Where are the opportunities? Where's that white space? And I know there's a lot of white space, but what's the framework to look at, from an investor standpoint, or even a CEO standpoint, where you want to put place your bets. >> No, very good question, so look, I did something. We talk as an investor in the board with many companies, right? So one thing that says as an investor, if you come back and say, I want to create a next generation Docker or a computer, there's no way nobody's going to invest. So that we can motor off, even if you want to do object storage or a block storage, I mean, I've been an investor board member of so many storage companies, there's no way as an industry, I'll write a check for a compute or storage, right? If you want to create a next generation network, like either NetSuite, or restart Juniper, Cisco, there is no way. But if you come back and say, I want to create a next generation Viper for remote working environments, where AI is at the core, I'm interested in that, right? So if you look at how the packets are dropped, there's no intelligence in either not switching today. The packets come, I do it. The intelligence is not built into the network with AI level. So if somebody comes with an AI, what good is all this NVD, our GPS, et cetera, if you cannot do wire speed, packet inspection, looking at the content and then route the traffic. If I see if it's a video package, but in UN Boston, there's high interview day of they should be loading our package faster, because you are a premium ISP. That intelligence has not gone there. So you will see, and that will be a bad people will happen in the network, switching, et cetera, right? So that is still an angle. But if you work and it comes to platform services, remember when I was at Pivotal and VMware, all models was my boss, that would, yes, as a platform, service is a game already won by the Cloud guys. >> Right. (indistinct) >> Silicon Valley Investors, I don't think you want to invest in past services, right? I mean, you might come with some lecture edition database to do some updates, there could be some game, let's say we want to do a time series database, or some metrics database, there's always some small angle, but the opportunity to go create a national database there it's very few. So I'm kind of eliminating all the black spaces, right? >> Yeah. >> We have the white spaces that comes in is the SAS level. Now to your point, if I'm Amazon, I'm going to compete with Snowflake, I have Redshift. So this is where at some point, these Cloud platforms, I call them aircraft carriers. They're not going to stay on the aircraft carriers, they're going to own the land as well. So they're going to move up to the SAS space. The question is you want to create a SAS service like CRM. They are not going to create a CRM like service, they may not create a sales force and service now, but if you're going to add a data warehouse, I can very well see Azure, Google, and AWS, going to create something to compute a Snowflake. Why would I not? It's so close to my database and data warehouse, I already have Redshift. So that's going to be nightlights, same reason, If you look at Netflix, you have a Netflix and you have Amazon prime. Netflix runs on Amazon, but you have Amazon prime. So you have the same model, you have Snowflake, and you'll have Redshift. The both will help each other, there'll be a... What do you call it? Coexistence will happen. But if you really want to invest, you want to invest in SAS companies. You do not want to be investing in a compliment players. You don't want to a feature. >> Yeah, that's great, I appreciate that perspective. And I wonder, so obviously Microsoft play in SAS, Google's got G suite. And I wonder if people often ask the Andy Jassy, you're going to move up the stack, you got to be an application, a SAS vendor, and you never say never with Atavist, But I wonder, and we were talking to Jerry Chen about this, years ago on theCube, and his angle was that Amazon will play, but they'll play through developers. They'll enable developers, and they'll participate, they'll take their, lick off the cone. So it's going to be interesting to see how directly Amazon plays, but at some point you got Tam expansion, you got to play in that space. >> Yeah, I'll give you an example of knowing, I got acquired by a couple of times by EMC. So I learned a lot from Joe Tucci and Paul Merage over the years. see Paul and Joe, what they did is to look at how 20 years, and they are very close to Boston in your area, Joe, what games did is they used to sell storage, but you know what he did, he went and bought the Apps to drive them. He bought like Legato, he bought Documentum, he bought Captiva, if you remember how he acquired all these companies as a services, he bought VMware to drive that. So I think the good angle that Microsoft has is, I'm a SAS player, I have dynamics, I have CRM, I have SharePoint, I have Collaboration, I have Office 365, MS Teams for users, and then I have the platform as Azure. So I think if I'm Amazon, (indistinct). I got to own the apps so that I can drive this workforce on my platform. >> Interesting. >> Just going to developers, like I know Jerry Chan, he was my peer a BMF. I don't think just literally to developers and that model works in open source, but the open source game is pretty much gone, and not too many companies made money. >> Well, >> Most companies pretty much gone. >> Yeah, he's right. Red hats not bad idea. But it's very interesting what you're saying there. And so, hey, its why Oracle wants to have Tiktok, running on their platform, right? I mean, it's going to. (laughing) It's going to drive that further integration. I wanted to ask you something, you were talking about, you wouldn't invest in storage or compute, but I wonder, and you mentioned some commentary about GPU's. Of course the videos has been going crazy, but they're now saying, okay, how do we expand our Team, they make the acquisition of arm, et cetera. What about this DPU thing, if you follow that, that data processing unit where they're like hyper dis-aggregation and then they reaggregate, and as an offload and really to drive data centric workloads. Have you looked at that at all? >> I did, I think, and that's a good angle. So I think, look, it's like, it goes through it. I don't know if you remember in your career, we have seen it. I used to get Silicon graphics. I saw the first graphic GPU, right? That time GPU was more graphic processor unit, >> Right, yeah, work stations. >> So then become NPUs at work processing units, right? There was a TCP/IP office offloading, if you remember right, there was like vector processing unit. So I think every once in a while the industry, recreated this separate unit, as a co-processor to the main CPU, because main CPU's inefficient, and it makes sense. And then Google created TPU's and then we have the new world of the media GPU's, now we have DPS all these are good, but what's happening is, all these are driving for machine learning, AI for the training period there. Training period Sometimes it's so long with the workloads, if you can cut down, it makes sense. >> Yeah. >> Because, but the question is, these aren't so specialized in nature. I can't use it for everything. >> Yup. >> I want Ideally, algorithms to be paralyzed, I want the training to be paralyzed, I want so having deep use and GPS are important, I think where I want to see them as more, the algorithm, there should be more investment from the NVIDIA's and these guys, taking the algorithm to be highly paralyzed them. (indistinct) And I think that still has not happened in industry yet. >> All right, so we're pretty much out of time, but what are you doing these days? Where are you spending your time, are you still in Stealth, give us a little glimpse. >> Yeah, no, I'm out of the Stealth, I'm actually the CEO of Aisera now, Aisera, obviously I invested with them, but I'm the CEO of Aisero. It's funded by Menlo ventures, Norwest, True, along with Khosla ventures and Ram Shriram is a big investor. Robin's on the board of Google, so these guys, look, we are going out to the collaboration game. How do you automate customer service and support for employees and then users, right? In this whole game, we talked about the Zoom, Slack and MS Teams, that's what I'm spending time, I want to create next generation service now. >> Fantastic. Muddu, I always love having you on you, pull punches, you tell it like it is, that you're a great visionary technologist. Thanks so much for coming on theCube, and participating in our program. >> Dave, it's always a pleasure speaking to you sir. Thank you. >> Okay. Keep it right there, there's more coming from Cuba and Cloud right after this break. (slow music)