>>AWS public sector summit here in person in Washington, D. C. For two days live. Finally a real event. I'm john for your host of the cube. Got a great guest Howard Levinson from data bricks, regional vice president and general manager of the federal team for data bricks. Uh Super unicorn. Is it a decade corn yet? It's uh, not yet public but welcome to the cube. >>I don't know what the next stage after unicorn is, but we're growing rapidly. >>Thank you. Our audience knows David bricks extremely well. Always been on the cube many times. Even back, we were covering them back when big data was big data. Now it's all data everything. So we watched your success. Congratulations. Thank you. Um, so there's no, you know, not a big bridge for us across to see you here at AWS public sector summit. Tell us what's going on inside the data bricks amazon relationship. >>Yeah. It's been a great relationship. You know, when the company got started some number of years ago we got a contract with the government to deliver the data brooks capability and they're classified cloud in amazon's classified cloud. So that was the start of a great federal relationship today. Virtually all of our businesses in AWS and we run in every single AWS environment from commercial cloud to Govcloud to secret top secret environments and we've got customers doing great things and experiencing great results from data bricks and amazon. >>The federal government's the classic, I call migration opportunity. Right? Because I mean, let's face it before the pandemic even five years ago, even 10 years ago. Glacier moving speed slow, slow and they had to get modernized with the pandemic forced really to do it. But you guys have already cleared the runway with your value problems. You've got lake house now you guys are really optimized for the cloud. >>Okay, hardcore. Yeah. We are, we only run in the cloud and we take advantage of every single go fast feature that amazon gives us. But you know john it's The Office of Management and Budget. Did a study a couple of years ago. I think there were 28,000 federal data centers, 28,000 federal data centers. Think about that for a minute and just think about like let's say in each one of those data centers you've got a handful of operational data stores of databases. The federal government is trying to take all of that data and make sense out of it. The first step to making sense out of it is bringing it all together, normalizing it. Fed aerating it and that's exactly what we do. And that's been a real win for our federal clients and it's been a real exciting opportunity to watch people succeed in that >>endeavour. We have another guest on. And she said those data center huggers tree huggers data center huggers, majority of term people won't let go. Yeah. So but they're slowly dying away and moving on to the cloud. So migrations huge. How are you guys migrating with your customers? Give us an example of how it's working. What are some of the use cases? >>So before I do that I want to tell you a quick story. I've I had the luxury of working with the Air Force Chief data officer Ailene vedrine and she is commonly quoted as saying just remember as as airmen it's not your data it's the Air Force's data. So people were data center huggers now their data huggers but all of that data belongs to the government at the end of the day. So how do we help in that? Well think about all this data sitting in all these operational data stores they're getting it's getting updated all the time. But you want to be able to Federated this data together and make some sense out of it. So for like an organization like uh us citizenship and immigration services they had I think 28 different data sources and they want to be able to pull that data basically in real time and bring it into a data lake. Well that means doing a change data capture off of those operational data stores transforming that data and normalizing it so that you can then enjoy it. And we've done that I think they're now up to 70 data sources that are continually ingested into their data lake. And from there they support thousands of users doing analysis and reports for the whole visa processing system for the United States, the whole naturalization environment And their efficiency has gone up I think by their metrics by 24 x. >>Yeah. I mean Sandy carter was just on the cube earlier. She's the Vice president partner ecosystem here at public sector. And I was coming to her that federal game has changed, it used to be hard to get into you know everybody and you navigate the trip wires and all the subtle hints and and the people who are friends and it was like cloak and dagger and so people were locked in on certain things databases and data because now has to be freely available. I know one of the things that you guys are passionate about and this is kind of hard core architectural thing is that you need horizontally scalable data to really make a I work right. Machine learning works when you have data. How far along are these guys in their thinking when you have a customer because we're seeing progress? How far along are we? >>Yeah, we still have a long way to go in the federal government. I mean, I tell everybody, I think the federal government's probably four or five years behind what data bricks top uh clients are doing. But there are clearly people in the federal government that have really ramped it up and are on a par were even exceeding some of the commercial clients, U. S. C. I. S CBP FBI or some of the clients that we work with that are pretty far ahead and I'll say I mentioned a lot about the operational data stores but there's all kinds of data that's coming in at U S. C. I. S. They do these naturalization interviews, those are captured in real text. So now you want to do natural language processing against them, make sure these interviews are of the highest quality control, We want to be able to predict which people are going to show up for interviews based on their geospatial location and the day of the week and other factors the weather perhaps. So they're using all of these data types uh imagery text and structure data all in the Lake House concept to make predictions about how they should run their >>business. So that's a really good point. I was talking with keith brooks earlier directive is development, go to market strategy for AWS public sector. He's been there from the beginning this the 10th year of Govcloud. Right, so we're kind of riffing but the jpl Nasa Jpl, they did production workloads out of the gate. Yeah. Full mission. So now fast forward today. Cloud Native really is available. So like how do you see the the agencies in the government handling Okay. Re platform and I get that but now to do the reef acting where you guys have the Lake House new things can happen with cloud Native technologies, what's the what's the what's the cross over point for that point. >>Yeah, I think our Lake House architecture is really a big breakthrough architecture. It used to be, people would take all of this data, they put it in a Hadoop data lake, they'd end up with a data swamp with really not good control or good data quality. And uh then they would take the data from the data swamp where the data lake and they curate it and go through an E. T. L. Process and put a second copy into their data warehouse. So now you have two copies of the data to governance models. Maybe two versions of the data. A lot to manage. A lot to control with our Lake House architecture. You can put all of that data in the data lake it with our delta format. It comes in a curated way. Uh there's a catalogue associated with the data. So you know what you've got. And now you can literally build an ephemeral data warehouse directly on top of that data and it exists only for the period of time that uh people need it. And so it's cloud Native. It's elastically scalable. It terminates when nobody's using it. We run the whole center for Medicaid Medicare services. The whole Medicaid repository for the United States runs in an ephemeral data warehouse built on Amazon S three. >>You know, that is a huge call out, I want to just unpack that for a second. What you just said to me puts the exclamation point on cloud value because it's not your grandfather's data warehouse, it's like okay we do data warehouse capability but we're using higher level cloud services, whether it's governance stuff for a I to actually make it work at scale for those environments. I mean that that to me is re factoring that's not re platform Ng. Just re platform that's re platform Ng in the cloud and then re factoring capability for on uh new >>advantages. It's really true. And now you know at CMS, they have one copy of the data so they do all of their reporting, they've got a lot of congressional reports that they need to do. But now they're leveraging that same data, not making a copy of it for uh the center for program integrity for fraud. And we know how many billions of dollars worth of fraud exist in the Medicaid system. And now we're applying artificial intelligence and machine learning on entity analytics to really get to the root of those problems. It's a game >>changer. And this is where the efficiency comes in at scale. Because you start to see, I mean we always talk on the cube about like how software is changed the old days you put on the shelf shelf where they called it. Uh that's our generation. And now you got the cloud, you didn't know if something is hot or not until the inventory is like we didn't sell through in the cloud. If you're not performing, you suck basically. So it's not working, >>it's an instant Mhm. >>Report card. So now when you go to the cloud, you think the data lake and uh the lake house what you guys do uh and others like snowflake and were optimized in the cloud, you can't deny it. And then when you compare it to like, okay, so I'm saving you millions and millions if you're just on one thing, never mind the top line opportunities. >>So so john you know, years ago people didn't believe the cloud was going to be what it is. Like pretty much today, the clouds inevitable. It's everywhere. I'm gonna make you another prediction. Um And you can say you heard it here first, the data warehouse is going away. The Lake house is clearly going to replace it. There's no need anymore for two separate copies, there's no need for a proprietary uh storage copy of your data and people want to be able to apply more than sequel to the data. Uh Data warehouses, just restrict. What about an ocean house? >>Yeah. Lake is kind of small. When you think about this lake, Michigan is pretty big now, I think it's I >>think it's going to go bigger than that. I think we're talking about Sky Computer, we've been a cloud computing, we're going to uh and we're going to do that because people aren't gonna put all of their data in one place, they're going to have, it spread across different amazon regions or or or amazon availability zones and you're going to want to share data and you know, we just introduced this delta sharing capability. I don't know if you're familiar with it but it allows you to share data without a sharing server directly from picking up basically the amazon, you RLS and sharing them with different organizations. So you're sharing in place. The data actually isn't moving. You've got great governance and great granularity of the data that you choose to share and data sharing is going to be the next uh >>next break. You know, I really loved the Lake House were fairly sing gateway. I totally see that. So I totally would align with that and say I bet with you on that one. The Sky net Skynet, the Sky computing. >>See you're taking it away man, >>I know Skynet got anything that was computing in the Sky is Skynet that's terminated So but that's real. I mean I think that's a concept where it's like, you know what services and functions does for servers, you don't have a data, >>you've got to be able to connect data, nobody lives in an island. You've got to be able to connect data and more data. We all know more data produces better results. So how do you get more data? You connect to more data sources, >>Howard great to have you on talk about the relationship real quick as we end up here with amazon, What are you guys doing together? How's the partnership? >>Yeah, I mean the partnership with amazon is amazing. We have, we work uh, I think probably 95% of our federal business is running in amazon's cloud today. As I mentioned, john we run across uh, AWS commercial AWS GovCloud secret environment. See to us and you know, we have better integration with amazon services than I'll say some of the amazon services if people want to integrate with glue or kinesis or Sagemaker, a red shift, we have complete integration with all of those and that's really, it's not just a partnership at the sales level. It's a partnership and integration at the engineering level. >>Well, I think I'm really impressed with you guys as a company. I think you're an example of the kind of business model that people might have been afraid of which is being in the cloud, you can have a moat, you have competitive advantage, you can build intellectual property >>and, and john don't forget, it's all based on open source, open data, like almost everything that we've done. We've made available to people, we get 30 million downloads of the data bricks technology just for people that want to use it for free. So no vendor lock in. I think that's really important to most of our federal clients into everybody. >>I've always said competitive advantage scale and choice. Right. That's a data bricks. Howard? Thanks for coming on the key, appreciate it. Thanks again. Alright. Cube coverage here in Washington from face to face physical event were on the ground. Of course, we're also streaming a digital for the hybrid event. This is the cubes coverage of a W. S. Public sector Summit will be right back after this short break.

(bright music) >> Welcome back to the CUBE presents the AWS Startup Showcase The Next Big Thing in cloud startups with AI security and life science tracks, 15 hottest growing startups are presented. And we had a great opening keynote with luminaries in the industry. And now our closing keynote is to get a deeper dive on cracking the code in the enterprise, how startups are changing the game and helping companies change. And they're also changing the game of open source. We have a great guest, Katie Drucker, Head of Business Development, Madrona Venture Group. Katie, thank you for coming on the CUBE for this special closing keynote. >> Thank you for having me, I appreciate it. >> So one of the topics we talked about with Soma from Madrona on the opening keynote, as well as Ali from Databricks is how startups are seeing success faster. So that's the theme of the Cloud speed, agility, but the game has changed in the enterprise. And I want to really discuss with you how growth changes and growth strategy specifically. They talk, go to market. We hear things like good sales to enterprise sales, organic, freemium, there's all kinds of different approaches, but at the end of the day, the most successful companies, the ones that might not be known that just come out of nowhere. So the economics are changing and the buyers are thinking differently. So let's explore that topic. So take us through your view 'cause you have a lot of experience. But first talk about your role at Madrona, what you do. >> Absolutely all great points. So my role at Madrona, I think I have personally one of the more enviable jobs and that my job is to... I get the privilege of working with all of these fantastic entrepreneurs in our portfolio and doing whatever we can as a firm to harness resources, knowledge, expertise, connections, to accelerate their growth. So my role in setting up business development is taking a look at all of those tools in the tool chest and partnering with the portfolio to make it so. And in our portfolio, we have a wide range of companies, some rely on enterprise sales, some have other go to markets. Some are direct to consumer, a wide range. >> Talk about the growth strategies that you see evolving because what's clear with the pandemic. And as we come out of it is that there are growth plays happening that don't look a little bit differently, more obvious now because of the Cloud scale, we're seeing companies like Databricks, like Snowflake, like other companies that have been built on the cloud or standalone. What are some of the new growth techniques, or I don't want to say growth hacking, that is a pejorative term, but like just a way for companies to quickly describe their value to an enterprise buyer who's moving away from the old RFP days of vendor selection. The game has changed. So take us through how you see secret key and unlocking that new equation of how to present value to an enterprise and how you see enterprises evaluating startups. >> Yes, absolutely. Well, and that's got a question, that's got a few components nestled in what I think are some bigger trends going on. AWS of course brought us the Cloud first. I think now the Cloud is more and more a utility. And so it's incumbent upon thinking about how an enterprise 'cause using the Cloud is going to go up the value stack and partner with its cloud provider and other service providers. I think also with that agility of operations, you have thinning, if you will, the systems of record and a lot of new entrance into this space that are saying things like, how can we harness AIML and other emerging trends to provide more value directly around work streams that were historically locked into those systems of record? And then I think you also have some price plans that are far more flexible around usage based as opposed to just flat subscription or even these big clunky annual or multi-year RFP type stuff. So all of those trends are really designed in ways that favor the emerging startup. And I think if done well, and in partnership with those underlying cloud providers, there can be some amazing benefits that the enterprise realizes an opportunity for those startups to grow. And I think that's what you're seeing. I think there's also this emergence of a buyer that's different than the CIO or the site the CISO. You have things with low code, no code. You've got other buyers in the organization, other line of business executives that are coming to the table, making software purchase decisions. And then you also have empowered developers that are these citizen builders and developer buyers and personas that really matter. So lots of inroads in places for a startup to reach in the enterprise to make a connection and to bring value. That's a great insight. I want to ask that just if you don't mind follow up on that, you mentioned personas. And what we're seeing is the shift happens. There's new roles that are emerging and new things that are being reconfigured or refactored if you will, whether it's human resources or AI, and you mentioned ML playing a role in automation. These are big parts of the new value proposition. How should companies posture to the customer? Because I don't want to say pivot 'cause that means it's not working but mostly extending our iterating around their positioning because as new things have not yet been realized, it might not be operationalized in a company or maybe new things need to be operationalized, it's a new solution for that. Positioning the value is super important and a lot of companies often struggle with that, but also if they get it right, that's the key. What's your feeling on startups in their positioning? So people will dismiss it like, "Oh, that's marketing." But maybe that's important. What's your thoughts on the great positioning question? >> I've been in this industry a long time. And I think there are some things that are just tried and true, and it is not unique to tech, which is, look, you have to tell a story and you have to reach the customer and you have to speak to the customer's need. And what that means is, AWS is a great example. They're famous for the whole concept of working back from the customer and thinking about what that customer's need is. I think any startup that is looking to partner or work alongside of AWS really has to embody that very, very customer centric way of thinking about things, even though, as we just talked about those personas are changing who that customer really is in the enterprise. And then speaking to that value proposition and meeting that customer and creating a dialogue with them that really helps to understand not only what their pain points are, but how you were offering solves those pain points. And sometimes the customer doesn't realize that that is their pain point and that's part of the education and part of the way in which you engage that dialogue. That doesn't change a lot, just generation to generation. I think the modality of how we have that dialogue, the methods in which we choose to convey that change, but that basic discussion is what makes us human. >> What's your... Great, great, great insight. I want to ask you on the value proposition question again, the question I often get, and it's hard to answer is am I competing on value or am I competing on commodity? And depending on where you're in the stack, there could be different things like, for example, land is getting faster, smaller, cheaper, as an example on Amazon. That's driving down to low cost high value, but it shifts up the stack. You start to see in companies this changing the criteria for how to evaluate. So an enterprise might be struggling. And I often hear enterprises say, "I don't know how to pick who I need. I buy tools, I don't buy many platforms." So they're constantly trying to look for that answer key, if you will, what's your thoughts on the changing requirements of an enterprise? And how to do vendor selection. >> Yeah, so obviously I don't think there's a single magic bullet. I always liked just philosophically to think about, I think it's always easier and frankly more exciting as a buyer to want to buy stuff that's going to help me make more revenue and build and grow as opposed to do things that save me money. And just in a binary way, I like to think which side of the fence are you sitting on as a product offering? And the best ways that you can articulate that, what opportunities are you unlocking for your customer? The problems that you're solving, what kind of growth and what impact is that going to lead to, even if you're one or two removed from that? And again, that's not a new concept. And I think that the companies that have that squarely in mind when they think about their go-to market strategy, when they think about the dialogue they're having, when they think about the problems that they're solving, find a much faster path. And I think that also speaks to why we're seeing so many explosion in the line of business, SAS apps that are out there. Again, that thinning of the systems of record, really thinking about what are the scenarios and work streams that we can have happened that are going to help with that revenue growth and unlocking those opportunities. >> What's the common startup challenge that you see when they're trying to do business development? Usually they build the product first, product led value, you hear that a lot. And then they go, "Okay, we're ready to sell, hire a sales guy." That seems to be shifting away because of the go to markets are changing. What are some of the challenges that startups have? What are some that you're seeing? >> Well, and I think the point that you're making about the changes are really almost a result of the trends that we're talking about. The sales organization itself is becoming... These work streams are becoming instrumented. Data is being collected, insights are being derived off of those things. So you see companies like Clary or Highspot or two examples or tutorial that are in our portfolio that are looking at that action and making the art of sales and marketing far more sophisticated overall, which then leads to the different growth hacking and the different insights that are driven. I think the common mistakes that I see across the board, especially with earlier stage startups, look you got to find product market fit. I think that's always... You start with a thesis or a belief and a passion that you're building something that you think the market needs. And it's a lot of dialogue you have to have to make sure that you do find that. I think once you find that another common problem that I see is leading with an explanation of technology. And again, not focusing on the buyer or the... Sorry, the buyer about solving a problem and focusing on that problem as opposed to focusing on how cool your technology is. Those are basic and really, really simple. And then I think setting a set of expectations, especially as it comes to business development and partnering with companies like AWS. The researching that you need to adequately meet the demand that can be turned on. And then I'm sure you heard about from Databricks, from an organization like AWS, you have to be pragmatic. >> Yeah, Databricks gone from zero a software sales a few years ago to over a billion. Now it looks like a Snowflake which came out of nowhere and they had a great product, but built on Amazon, they became the data cloud on top of Amazon. And now they're growing just whole new business models and new business development techniques. Katie, thank you for sharing your insight here. The CUBE's closing keynote. Thanks for coming on. >> Appreciate it, thank you. >> Okay, Katie Drucker, Head of Business Development at Madrona Venture Group. Premier VC in the Seattle area and beyond they're doing a lot of cloud action. And of course they know AWS very well and investing in the ecosystem. So great, great stuff there. Next up is Peter Wagner partner at Wing.VX. Love this URL first of all 'cause of the VC domain extension. But Peter is a long time venture capitalist. I've been following his career. He goes back to the old networking days, back when the internet was being connected during the OSI days, when the TCP IP open systems interconnect was really happening and created so much. Well, Peter, great to see you on the CUBE here and congratulations with success at Wing VC. >> Yeah, thanks, John. It's great to be here. I really appreciate you having me. >> Reason why I wanted to have you come on. First of all, you had a great track record in investing over many decades. You've seen many waves of innovation, startups. You've seen all the stories. You've seen the movie a few times, as I say. But now more than ever, enterprise wise it's probably the hottest I've ever seen. And you've got a confluence of many things on the stack. You were also an early seed investor in Snowflake, well-regarded as a huge success. So you've got your eye on some of these awesome deals. Got a great partner over there has got a network experience as well. What is the big aha moment here for the industry? Because it's not your classic enterprise startups anymore. They have multiple things going on and some of the winners are not even known. They come out of nowhere and they connect to enterprise and get the lucrative positions and can create a moat and value. Like out of nowhere, it's not the old way of like going to the airport and doing an RFP and going through the stringent requirements, and then you're in, you get to win the lucrative contract and you're in. Not anymore, that seems to have changed. What's your take on this 'cause people are trying to crack the code here and sometimes you don't have to be well-known. >> Yeah, well, thank goodness the game has changed 'cause that old thing was (indistinct) So I for one don't miss it. There was some modernization movement in the enterprise and the modern enterprise is built on data powered by AI infrastructure. That's an agile workplace. All three of those things are really transformational. There's big investments being made by enterprises, a lot of receptivity and openness to technology to enable all those agendas, and that translates to good prospects for startups. So I think as far as my career goes, I've never seen a more positive or fertile ground for startups in terms of penetrating enterprise, it doesn't mean it's easy to do, but you have a receptive audience on the other side and that hasn't necessarily always been the case. >> Yeah, I got to ask you, I know that you're a big sailor and your family and Franks Lubens also has a boat and sailing metaphor is always good to have 'cause you got to have a race that's being run and they have tactics. And this game that we're in now, you see the successes, there's investment thesises, and then there's also actually bets. And I want to get your thoughts on this because a lot of enterprises are trying to figure out how to evaluate startups and starts also can make the wrong bet. They could sail to the wrong continent and be in the wrong spot. So how do you pick the winners and how should enterprises understand how to pick winners too? >> Yeah, well, one of the real important things right now that enterprise is facing startups are learning how to do and so learning how to leverage product led growth dynamics in selling to the enterprise. And so product led growth has certainly always been important consumer facing companies. And then there's a few enterprise facing companies, early ones that cracked the code, as you said. And some of these examples are so old, if you think about, like the ones that people will want to talk about them and talk about Classy and want to talk about Twilio and these were of course are iconic companies that showed the way for others. But even before that, folks like Solar Winds, they'd go to market model, clearly product red, bottom stuff. Back then we didn't even have those words to talk about it. And then some of the examples are so enormous if think about them like the one right in front of your face, like AWS. (laughing) Pretty good PLG, (indistinct) but it targeted builders, it targeted developers and flipped over the way you think about enterprise infrastructure, as a result some how every company, even if they're harnessing relatively conventional sales and marketing motion, and you think about product led growth as a way to kick that motion off. And so it's not really an either word even more We might think OPLJ, that means there's no sales keep one company not true, but here's a way to set the table so that you can very efficiently use your sales and marketing resources, only have the most attractive targets and ones that are really (indistinct) >> I love the product led growth. I got to ask you because in the networking days, I remember the term inevitability was used being nested in a solution that they're just going to Cisco off router and a firewall is one you can unplug and replace with another vendor. Cisco you'd have to go through no switching costs were huge. So when you get it to the Cloud, how do you see the competitiveness? Because we were riffing on this with Ali, from Databricks where the lock-in might be value. The more value provider is the lock-in. Is their nestedness? Is their intimate ability as a competitive advantage for some of these starts? How do you look at that? Because startups, they're using open source. They want to have a land position in an enterprise, but how do they create that sustainable competitive advantage going forward? Because again, this is what you do. You bet on ones that you can see that could establish a model whatever we want to call it, but a competitive advantage and ongoing nested position. >> Sometimes it has to do with data, John, and so you mentioned Snowflake a couple of times here, a big part of Snowflake's strategy is what they now call the data cloud. And one of the reasons you go there is not to just be able to process data, to actually get access to it, exchange with the partners. And then that of course is a great reason for the customers to come to the Snowflake platform. And so the more data it gets more customers, it gets more data, the whole thing start spinning in the right direction. That's a really big example, but all of these startups that are using ML in a fundamental way, applying it in a novel way, the data modes are really important. So getting to the right data sources and training on it, and then putting it to work so that you can see that in this process better and doing this earlier on that scale. That's a big part of success. Another company that I work with is a good example that I call (indistinct) which works in sales technology space, really crushing it in terms of building better sales organizations both at performance level, in terms of the intelligence level, and just overall revenue attainment using ML, and using novel data sources, like the previously lost data or phone calls or Zoom calls as you already know. So I think the data advantages are really big. And smart startups are thinking through it early. >> It's interest-- >> And they're planning by the way, not to ramble on too much, but they're betting that PLG strategy. So their land option is designed not just to be an interesting way to gain usage, but it's also a way to gain access to data that then enables the expand in a component. >> That is a huge call-out point there, I was going to ask another question, but I think that is the key I see. It's a new go to market in a way. product led with that kind of approach gets you a beachhead and you get a little position, you get some data that is a cloud model, it means variable, whatever you want to call it variable value proposition, value proof, or whatever, getting that data and reiterating it. So it brings up the whole philosophical question of okay, product led growth, I love that with product led growth of data, I get that. Remember the old platform versus a tool? That's the way buyers used to think. How has that changed? 'Cause now almost, this conversation throws out the whole platform thing, but isn't like a platform. >> It looks like it's all. (laughs) you can if it is a platform, though to do that you can reveal that later, but you're looking for adoption, so if it's down stock product, you're looking for adoption by like developers or DevOps people or SOEs, and they're trying to solve a problem, and they want rapid gratification. So they don't want to have an architectural boomimg, placed in front of them. And if it's up stock product and application, then it's a user or the business or whatever that is, is adopting the application. And again, they're trying to solve a very specific problem. You need instant and immediate obvious time and value. And now you have a ticket to the dance and build on that and maybe a platform strategy can gradually take shape. But you know who's not in this conversation is the CIO, it's like, "I'm always the last to know." >> That's the CISO though. And they got him there on the firing lines. CISOs are buying tools like it's nobody's business. They need everything. They'll buy anything or you go meet with sand, they'll buy it. >> And you make it sound so easy. (laughing) We do a lot of security investment if only (indistinct) (laughing) >> I'm a little bit over the top, but CISOs are under a lot of pressure. I would talk to the CISO at Capital One and he was saying that he's on Amazon, now he's going to another cloud, not as a hedge, but he doesn't want to focus development teams. So he's making human resource decisions as well. Again, back to what IT used to be back in the old days where you made a vendor decision, you built around it. So again, clouds play that way. I see that happening. But the question is that I think you nailed this whole idea of cross hairs on the target persona, because you got to know who you are and then go to the market. So if you know you're a problem solving and the lower in the stack, do it and get a beachhead. That's a strategy, you can do that. You can't try to be the platform and then solve a problem at the same time. So you got to be careful. Is that what you were getting at? >> Well, I think you just understand what you're trying to achieve in that line of notion. And how those dynamics work and you just can't drag it out. And they could make it too difficult. Another company I work with is a very strategic cloud data platform. It's a (indistinct) on systems. We're not trying to foist that vision though (laughs) or not adopters today. We're solving some thorny problems with them in the short term, rapid time to value operational needs in scale. And then yeah, once they found success with (indistinct) there's would be an opportunity to be increasing the platform, and an obstacle for those customers. But we're not talking about that. >> Well, Peter, I appreciate you taking the time and coming out of a board meeting, I know that you're super busy and I really appreciate you making time for us. I know you've got an impressive partner in (indistinct) who's a former Sequoia, but Redback Networks part of that company over the years, you guys are doing extremely well, even a unique investment thesis. I'd like you to put the plug in for the firm. I think you guys have a good approach. I like what you guys are doing. You're humble, you don't brag a lot, but you make a lot of great investments. So could you take them in to explain what your investment thesis is and then how that relates to how an enterprise is making their investment thesis? >> Yeah, yeah, for sure. Well, the concept that I described earlier that the modern enterprise movement as a workplace built on data powered by AI. That's what we're trying to work with founders to enable. And also we're investing in companies that build the products and services that enable that modern enterprise to exist. And we do it from very early stages, but with a longterm outlook. So we'll be leading series and series, rounds of investment but staying deeply involved, both operationally financially throughout the whole life cycle of the company. And then we've done that a bunch of times, our goal is always the big independent public company and they don't always make it but enough for them to have it all be worthwhile. An interesting special case of this, and by the way, I think it intersects with some of startup showcase here is in the life sciences. And I know you were highlighting a lot of healthcare websites and deals, and that's a vertical where to disrupt tremendous impact of data both new data availability and new ways to put it to use. I know several of my partners are very focused on that. They call it bio-X data. It's a transformation all on its own. >> That's awesome. And I think that the reason why we're focusing on these verticals is if you have a cloud horizontal scale view and vertically specialized with machine learning, every vertical is impacted by data. It's so interesting that I think, first start, I was probably best time to be a cloud startup right now. I really am bullish on it. So I appreciate you taking the time Peter to come in again from your board meeting, popping out. Thanks for-- (indistinct) Go back in and approve those stock options for all the employees. Yeah, thanks for coming on. Appreciate it. >> All right, thank you John, it's a pleasure. >> Okay, Peter Wagner, Premier VC, very humble Wing.VC is a great firm. Really respect them. They do a lot of great investing investments, Snowflake, and we have Dave Vellante back who knows a lot about Snowflake's been covering like a blanket and Sarbjeet Johal. Cloud Influencer friend of the CUBE. Cloud commentator and cloud experience built clouds, runs clouds now invests. So V. Dave, thanks for coming back on. You heard Peter Wagner at Wing VC. These guys have their roots in networking, which networking back in the day was, V. Dave. You remember the internet Cisco days, remember Cisco, Wellfleet routers. I think Peter invested in Arrow Point, remember Arrow Point, that was about in the 495 belt where you were. >> Lynch's company. >> That was Chris Lynch's company. I think, was he a sales guy there? (indistinct) >> That was his first big hit I think. >> All right, well guys, let's wrap this up. We've got a great program here. Sarbjeet, thank you for coming on. >> No worries. Glad to be here todays. >> Hey, Sarbjeet. >> First of all, really appreciate the Twitter activity lately on the commentary, the observability piece on Jeremy Burton's launch, Dave was phenomenal, but Peter was talking about this dynamic and I think ties this cracking the code thing together, which is there's a product led strategy that feels like a platform, but it's also a tool. In other words, it's not mutually exclusive, the old methods thrown out the window. Land in an account, know what problem you're solving. If you're below the stack, nail it, get data and go from there. If you're a process improvement up the stack, you have to much more of a platform longer-term sale, more business oriented, different motions, different mechanics. What do you think about that? What's your reaction? >> Yeah, I was thinking about this when I was listening to some of the startups pitching, if you will, or talking about what they bring to the table in this cloud scale or cloud era, if you will. And there are tools, there are applications and then they're big monolithic platforms, if you will. And then they're part of the ecosystem. So I think the companies need to know where they play. A startup cannot be platform from the get-go I believe. Now many aspire to be, but they have to start with tooling. I believe in, especially in B2B side of things, and then go into the applications, one way is to go into the application area, if you will, like a very precise use cases for certain verticals and stuff like that. And other parties that are going into the platform, which is like horizontal play, if you will, in technology. So I think they have to understand their age, like how old they are, how new they are, how small they are, because when their size matter when you are procuring as a big business, procuring your technology vendors size matters and the economic viability matters and their proximity to other windows matter as well. So I think we'll jump into that in other discussions later, but I think that's key, as you said. >> I would agree with that. I would phrase it in my mind, somewhat differently from Sarbjeet which is you have product led growth, and that's your early phase and you get product market fit, you get product led growth, and then you expand and there are many, many examples of this, and that's when you... As part of your team expansion strategy, you're going to get into the platform discussion. There's so many examples of that. You take a look at Ali Ghodsi today with what's happening at Databricks, Snowflake is another good example. They've started with product led growth. And then now they're like, "Okay, we've got to expand the team." Okta is another example that just acquired zero. That's about building out the platform, versus more of a point product. And there's just many, many examples of that, but you cannot to your point, very hard to start with a platform. Arm did it, but that was like a one in a million chance. >> It's just harder, especially if it's new and it's not operationalized yet. So one of the things Dave that we've observed the Cloud is some of the best known successes where nobody's not known at all, database we've been covering from the beginning 'cause we were close to that movement when they came out of Berkeley. But they still were misunderstood and they just started generating revenue in only last year. So again, only a few years ago, zero software revenue, now they're approaching a billion dollars. So it's not easy to make these vendor selections anymore. And if you're new and you don't have someone to operate it or your there's no department and the departments changing, that's another problem. These are all like enterprisey problems. What's your thoughts on that, Dave? >> Well, I think there's a big discussion right now when you've been talking all day about how should enterprise think about startups and think about most of these startups they're software companies and software is very capital efficient business. At the same time, these companies are raising hundreds of millions, sometimes over a billion dollars before they go to IPO. Why is that? A lot of it's going to promotion. I look at it as... And there's a big discussion going on but well, maybe sales can be more efficient and more direct and so forth. I really think it comes down to the golden rule. Two things really mattered in the early days in the startup it's sales and engineering. And writers should probably say engineering and sales and start with engineering. And then you got to figure out your go to market. Everything else is peripheral to those two and you don't get those two things right, you struggle. And I think that's what some of these successful startups are proving. >> Sarbjeet, what's your take on that point? >> Could you repeat the point again? Sorry, I lost-- >> As cloud scale comes in this whole idea of competing, the roles are changing. So look at IOT, look at the Edge, for instance, you got all kinds of new use cases that no one actually knows is a problem to solve. It's just pure opportunity. So there's no one's operational I could have a product, but it don't know we can buy it yet. It's a problem. >> Yeah, I think the solutions have to be point solutions and the startups need to focus on the practitioners, number one, not the big buyers, not the IT, if you will, but the line of business, even within that sphere, like just focus on the practitioners who are going to use that technology. I talked to, I think it wasn't Fiddler, no, it was CoreLogics. I think that story was great today earlier in how they kind of struggle in the beginning, they were trying to do a big bang approach as a startup, but then they almost stumbled. And then they found their mojo, if you will. They went to Don the market, actually, that's a very classic theory of disruption, like what we study from Harvard School of Business that you go down the market, go to the non-consumers, because if you're trying to compete head to head with big guys. Because most of the big guys have lot of feature and functionality, especially at the platform level. And if you're trying to innovate in that space, you have to go to the practitioners and solve their core problems and then learn and expand kind of thing. So I think you have to focus on practitioners a lot more than the traditional oracle buyers. >> Sarbjeet, we had a great thread last night in Twitter, on observability that you started. And there's a couple of examples there. Chaos searches and relatively small company right now, they just raised them though. And they're part of this star showcase. And they could've said, "Hey, we're going to go after Splunk." But they chose not to. They said, "Okay, let's kind of disrupt the elk stack and simplify that." Another example is a company observed, you've mentioned Jeremy Burton's company, John. They're focused really on SAS companies. They're not going after initially these complicated enterprise deals because they got to get it right or else they'll get churn, and churn is that silent killer of software companies. >> The interesting other company that was on the showcase was Tetra Science. I don't know if you noticed that one in the life science track, and again, Peter Wagner pointed out the life science. That's an under recognized in the press vertical that's exploding. Certainly during the pandemic you saw it, Tetra science is an R&D cloud, Dave, R&D data cloud. So pharmaceuticals, they need to do their research. So the pandemic has brought to life, this now notion of tapping into data resources, not just data lakes, but like real deal. >> Yeah, you and Natalie and I were talking about that this morning and that's one of the opportunities for R&D and you have all these different data sources and yeah, it's not just about the data lake. It's about the ecosystem that you're building around them. And I see, it's really interesting to juxtapose what Databricks is doing and what Snowflake is doing. They've got different strategies, but they play a part there. You can see how ecosystems can build that system. It's not one company is going to solve all these problems. It's going to really have to be connections across these various companies. And that's what the Cloud enables and ecosystems have all this data flowing that can really drive new insights. >> And I want to call your attention to a tweet Sarbjeet you wrote about Splunk's earnings and they're data companies as well. They got Teresa Carlson there now AWS as the president, working with Doug, that should change the game a little bit more. But there was a thread of the neath there. Andy Thry says to replies to Dave you or Sarbjeet, you, if you're on AWS, they're a fine solution. The world doesn't just revolve around AWS, smiley face. Well, a lot of it does actually. So (laughing) nice point, Andy. But he brings up this thing and Ali brought it up too, Hybrid now is a new operating system for what now Edge does. So we got Mobile World Congress happening this month in person. This whole Telco 5G brings up a whole nother piece of the Cloud puzzle. Jeff Barr pointed out in his keynote, Dave. Guys, I want to get your reaction. The Edge now is... I'm calling it the super Edge because it's not just Edge as we know it before. You're going to have these pops, these points of presence that are going to have wavelength as your spectrum or whatever they have. I think that's the solution for Azure. So you're going to have all this new cloud power for low latency applications. Self-driving delivery VR, AR, gaming, Telemetry data from Teslas, you name it, it's happening. This is huge, what's your thoughts? Sarbjeet, we'll start with you. >> Yeah, I think Edge is like bound to happen. And for many reasons, the volume of data is increasing. Our use cases are also expanding if you will, with the democratization of computer analysis. Specialization of computer, actually Dave wrote extensively about how Intel and other chip players are gearing up for that future if you will. Most of the inference in the AI world will happen in the field close to the workloads if you will, that can be mobility, the self-driving car that can be AR, VR. It can be healthcare. It can be gaming, you name it. Those are the few use cases, which are in the forefront and what alarm or use cases will come into the play I believe. I've said this many times, Edge, I think it will be dominated by the hyperscalers, mainly because they're building their Metro data centers now. And with a very low latency in the Metro areas where the population is, we're serving the people still, not the machines yet, or the empty areas where there is no population. So wherever the population is, all these big players are putting their data centers there. And I think they will dominate the Edge. And I know some Edge lovers. (indistinct) >> Edge huggers. >> Edge huggers, yeah. They don't like the hyperscalers story, but I think that's the way were' going. Why would we go backwards? >> I think you're right, first of all, I agree with the hyperscale dying you look at the top three clouds right now. They're all in the Edge, Hardcore it's a huge competitive battleground, Dave. And I think the missing piece, that's going to be uncovered at Mobile Congress. Maybe they'll miss it this year, but it's the developer traction, whoever wins the developer market or wins the loyalty, winning over the market or having adoption. The applications will drive the Edge. >> And I would add the fourth cloud is Alibaba. Alibaba is actually bigger than Google and they're crushing it as well. But I would say this, first of all, it's popular to say, "Oh not everything's going to move into the Cloud, John, Dave, Sarbjeet." But the fact is that AWS they're trend setter. They are crushing it in terms of features. And you'd look at what they're doing in the plumbing with Annapurna. Everybody's following suit. So you can't just ignore that, number one. Second thing is what is the Edge? Well, the edge is... Where's the logical place to process the data? That's what the Edge is. And I think to your point, both Sarbjeet and John, the Edge is going to be won by developers. It's going to be one by programmability and it's going to be low cost and really super efficient. And most of the data is going to stay at the Edge. And so who is in the best position to actually create that? Is it going to be somebody who was taking an x86 box and throw it over the fence and give it a fancy name with the Edge in it and saying, "Here's our Edge box." No, that's not what's going to win the Edge. And so I think first of all it's huge, it's wide open. And I think where's the innovation coming from? I agree with you it's the hyperscalers. >> I think the developers as John said, developers are the kingmakers. They build the solutions. And in that context, I always talk about the skills gravity, a lot of people are educated in certain technologies and they will keep using those technologies. Their proximity to that technology is huge and they don't want to learn something new. So as humans we just tend to go what we know how to use it. So from that front, I usually talk with consumption economics of cloud and Edge. It has to focus on the practitioners. And in this case, practitioners are developers because you're just cooking up those solutions right now. We're not serving that in huge quantity right now, but-- >> Well, let's unpack that Sarbjeet, let's unpack that 'cause I think you're right on the money on that. The consumption of the tech and also the consumption of the application, the end use and end user. And I think the reason why hyperscalers will continue to dominate besides the fact that they have all the resource and they're going to bring that to the Edge, is that the developers are going to be driving the applications at the Edge. So if you're low latency Edge, that's going to open up new applications, not just the obvious ones I did mention, gaming, VR, AR, metaverse and other things that are obvious. There's going to be non-obvious things that are going to be huge that are going to come out from the developers. But the Cloud native aspect of the hyperscalers, to me is where the scales are tipping, let me explain. IT was built to build a supply resource to the businesses who were writing business applications. Mostly driven by IBM in the mainframe in the old days, Dave, and then IT became IT. Telcos have been OT closed, "This is our thing, that's it." Now they have to open up. And the Cloud native technologies is the fastest way to value. And I think that paths, Sarbjeet is going to be defined by this new developer and this new super Edge concept. So I think it's going to be wide open. I don't know what to say. I can't guess, but it's going to be creative. >> Let me ask you a question. You said years ago, data's new development kit, does low code and no code to Sarbjeet's point, change the equation? In other words, putting data in the hands of those OT professionals, those practitioners who have the context. Does low-code and no-code enable, more of those protocols? I know it's a bromide, but the citizen developer, and what impact does that have? And who's in the best position? >> Well, I think that anything that reduces friction to getting stuff out there that can be automated, will increase the value. And then the question is, that's not even a debate. That's just fact that's going to be like rent, massive rise. Then the issue comes down to who has the best asset? The software asset that's eating the world or the tower and the physical infrastructure. So if the physical infrastructure aka the Telcos, can't generate value fast enough, in my opinion, the private equity will come in and take it over, and then refactor that business model to take advantage of the over the top software model. That to me is the big stare down competition between the Telco world and this new cloud native, whichever one yields in valley is going to blink first, if you say. And I think the Cloud native wins this one hands down because the assets are valuable, but only if they enable the new model. If the old model tries to hang on to the old hog, the old model as the Edge hugger, as Sarbjeet says, they'll just going to slowly milk that cow dry. So it's like, it's over. So to me, they have to move. And I think this Mobile World Congress day, we will see, we will be looking for that. >> Yeah, I think that in the Mobile World Congress context, I think Telcos should partner with the hyperscalers very closely like everybody else has. And they have to cave in. (laughs) I usually say that to them, like the people came in IBM tried to fight and they cave in. Other second tier vendors tried to fight the big cloud vendors like top three or four. And then they cave in. okay, we will serve our stuff through your cloud. And that's where all the buyers are congregating. They're going to buy stuff along with the skills gravity, the feature proximity. I've got another term I'll turn a coin. It matters a lot when you're doing one thing and you want to do another thing when you're doing all this transactional stuff and regular stuff, and now you want to do data science, where do you go? You go next to it, wherever you have been. Your skills are in that same bucket. And then also you don't have to write a new contract with a new vendor, you just go there. So in order to serve, this is a lesson for startups as well. You need to prepare yourself for being in the Cloud marketplaces. You cannot go alone independently to fight. >> Cloud marketplace is going to replace procurement, for sure, we know that. And this brings up the point, Dave, we talked about years ago, remember on the CUBE. We said, there's going to be Tier two clouds. I used that word in quotes cause nothing... What does it even mean Tier two. And we were talking about like Amazon, versus Microsoft and Google. We set at the time and Alibaba but they're in China, put that aside for a second, but the big three. They're going to win it all. And they're all going to be successful to a relative terms, but whoever can enable that second tier. And it ended up happening, Snowflake is that example. As is Databricks as is others. So Google and Microsoft as fast as they can replicate the success of AWS by enabling someone to build their business on their cloud in a way that allows the customer to refactor their business will win. They will win most of the lion's share my opinion. So I think that applies to the Edge as well. So whoever can come in and say... Whichever cloud says, "I'm going to enable the next Snowflake, the next enterprise solution." I think takes it. >> Well, I think that it comes back... Every conversation coming back to the data. And if you think about the prevailing way in which we treated data with the exceptions of the two data driven companies in their quotes is as we've shoved all the data into some single repository and tried to come up with a single version of the truth and it's adjudicated by a centralized team, with hyper specialized roles. And then guess what? The line of business, there's no context for the business in that data architecture or data Corpus, if you will. And then the time it takes to go from idea for a data product or data service commoditization is way too long. And that's changing. And the winners are going to be the ones who are able to exploit this notion of leaving data where it is, the point about data gravity or courting a new term. I liked that, I think you said skills gravity. And then enabling the business lines to have access to their own data teams. That's exactly what Ali Ghodsi, he was saying this morning. And really having the ability to create their own data products without having to go bow down to an ivory tower. That is an emerging model. All right, well guys, I really appreciate the wrap up here, Dave and Sarbjeet. I'd love to get your final thoughts. I'll just start by saying that one of the highlights for me was the luminary guests size of 15 great companies, the luminary guests we had from our community on our keynotes today, but Ali Ghodsi said, "Don't listen to what everyone's saying in the press." That was his position. He says, "You got to figure out where the puck's going." He didn't say that, but I'm saying, I'm paraphrasing what he said. And I love how he brought up Sky Cloud. I call it Sky net. That's an interesting philosophy. And then he also brought up that machine learning auto ML has got to be table stakes. So I think to me, that's the highlight walk away. And the second one is this idea that the enterprises have to have a new way to procure and not just the consumption, but some vendor selection. I think it's going to be very interesting as value can be proved with data. So maybe the procurement process becomes, here's a beachhead, here's a little bit of data. Let me see what it can do. >> I would say... Again, I said it was this morning, that the big four have given... Last year they spent a hundred billion dollars more on CapEx. To me, that's a gift. In so many companies, especially focusing on trying to hang onto the legacy business. They're saying, "Well not everything's going to move to the Cloud." Whatever, the narrative should change to, "Hey, thank you for that gift. We're now going to build value on top of the Cloud." Ali Ghodsi laid that out, how Databricks is doing it. And it's clearly what Snowflake's new with the data cloud. It basically a layer that abstracts all that underlying complexity and add value on top. Eventually going out to the Edge. That's a value added model that's enabled by the hyperscalers. And that to me, if I have to evaluate where I'm going to place my bets as a CIO or IT practitioner, I'm going to look at who are the ones that are actually embracing that investment that's been made and adding value on top in a way that can drive my data-driven, my digital business or whatever buzzword you want to throw on. >> Yeah, I think we were talking about the startups in today's sessions. I think for startups, my advice is to be as close as you can be to hyperscalers and anybody who awards them, they will cave in at the end of the day, because that's where the whole span of gravity is. That's what the innovation gravity is, everybody's gravitating towards that. And I would say quite a few times in the last couple of years that the rate of innovation happening in a non-cloud companies, when I talk about non-cloud means are not public companies. I think it's like diminishing, if you will, as compared to in cloud, there's a lot of innovation. The Cloud companies are not paying by power people anymore. They have all sophisticated platforms and leverage those, and also leverage the marketplaces and leverage their buyers. And the key will be how you highlight yourself in that cloud market place if you will. It's like in a grocery store where your product is placed and you have to market around it, and you have to have a good story telling team in place as well after you do the product market fit. I think that's a key. I think just being close to the Cloud providers, that's the way to go for startups. >> Real, real quick. Each of you talk about what it takes to crack the code for the enterprise in the modern era now. Dave, we'll start with you. What's it take? (indistinct) >> You got to have it be solving a problem that is 10X better at one 10th a cost of anybody else, if you're a small company, that rule number one. Number two is you obviously got to get product market fit. You got to then figure out. And I think, and again, you're in your early phases, you have to be almost processed builders, figure out... Your KPIs should all be built around retention. How do I define customer success? How do I keep customers and how do I make them loyal so that I know that my cost of acquisition is going to be at least one-third or lower than my lifetime value of that customer? So you've got to nail that. And then once you nail that, you've got to codify that process in the next phase, which really probably gets into your platform discussion. And that's really where you can start to standardize and scale and figure out your go to market and the relationship between marketing spend and sales productivity. And then when you get that, then you got to move on to figure out your Mot. Your Mot might just be a brand. It might be some secret sauce, but more often than not though, it's going to be the relationship that you build. And I think you've got to think about those phases and in today's world, you got to move really fast. Sarbjeet, real quick. What's the secret to crack the code? >> I think the secret to crack the code is partnership and alliances. As a small company selling to the bigger enterprises, the vendors size will be one of the big objections. Even if they don't say it, it's on the back of their mind, "What if these guys disappear tomorrow what would we do if we pick this technology?" And another thing is like, if you're building on the left side, which is the developer side, not on the right side, which is the operations or production side, if you will, you have to understand the sales cycles are longer on the right side and left side is easier to get to, but that's why we see a lot more startups. And on the left side of your DevOps space, if you will, because it's easier to sell to practitioners and market to them and then show the value correctly. And also understand that on the left side, the developers are very know how hungry, on the right side people are very cost-conscious. So understanding the traits of these different personas, if you will buyers, it will, I think set you apart. And as Dave said, you have to solve a problem, focus on practitioners first, because you're small. You have to solve political problems very well. And then you can expand. >> Well, guys, I really appreciate the time. Dave, we're going to do more of these, Sarbjeet we're going to do more of these. We're going to add more community to it. We're going to add our community rooms next time. We're going to do these quarterly and try to do them as more frequently, we learned a lot and we still got a lot more to learn. There's a lot more contribution out in the community that we're going to tap into. Certainly the CUBE Club as we call it, Dave. We're going to build this actively around Cloud. This is another 20 years. The Edge brings us more life with Cloud, it's really exciting. And again, enterprise is no longer an enterprise, it's just the world now. So great companies here, the next Databricks, the next IPO. The next big thing is in this list, Dave. >> Hey, John, we'll see you in Barcelona. Looking forward to that. Sarbjeet, I know in a second half, we're going to run into each other. So (indistinct) thank you John. >> Trouble has started. Great talking to you guys today and have fun in Barcelona and keep us informed. >> Thanks for coming. I want to thank Natalie Erlich who's in Rome right now. She's probably well past her bedtime, but she kicked it off and emceeing and hosting with Dave and I for this AW startup showcase. This is batch two episode two day. What do we call this? It's like a release so that the next 15 startups are coming. So we'll figure it out. (laughs) Thanks for watching everyone. Thanks. (bright music)

>>Well, welcome everybody to our last session of the day. I want to introduce you to Sean O'Meara. Orfield Cto. Hey, Sean. >>Hey, Nick. Good afternoon. It's been a crazy day. It has. It's been a busy run up to today in a busy day with a lot of great things going on. You know, we've heard from Adrian on his strategy this morning. The great way the Marantz is moving forward. We announced our new product line. You know, we spoke about the new doctor Enterprise Container Cloud line, New future for Mirant. Us. We had a great lineup of customers share their story. We introduced lanes following on the lanes launch a couple of weeks ago. Andi, we're introducing new great projects like our mosque project. New way to deliver open stack going into the future on then in parallel sel. This We ran a great tutorial tracker teachers you all about how to use these new products, and hopefully you'll go and everyone had opportunity to go and look through guys. Yeah. What's next? What is next? Yeah, lots going on. A lot of new things that we're thinking about for the future. Obviously, a lot of work to do on what we have right now. A lot of great things coming. But, you know, we've had this opportunity to talk about all these cool things that are coming down the road. And everybody these days seems to be talking about topics like edge computing or hybrid cloud. Or, you know, hyper scale data centers, even things like disaster recovery is a service. Andi, you know, we talk a lot about things like hyper converged, but frankly, it's boring. It's one thing a little. Good morning. Uh, you know, you and I have been talking about these topics for a while now, and I think it's about time when we spoke about some of the cool things that we're thinking about for the future, not necessarily looking out for the road map, but ideas for the future. Things that may could have an impact on the way we do business going to. So today we're gonna talk a little bit about things like pervasive computing. A nick, what is pervasive computing. >>Well, basically pervasive computing is when everything that you interact with, for the most part, is computerized. So in some ways, we're already there in that You know your phone is a computer. Your refrigerator may have a computer in it. Um, your smart watch your car has a computer in it. And the the most obvious sign of that is this whole Internet of things where, you know your vacuum is, uh is connected to your phone and all of that. And so pervasive computing is this, uh is this sense of you don't even really think about it. You just kind of assume that everything is computerized. >>So how is that different from ubiquitous computing? >>Oh, God. You hit, You hit my hot button. Okay, so if you look, there are a lot of places that will say that pervasive computing and ubiquitous computing are the same thing, but not the same thing. Don't use them interchangeably. They're not the same thing. You big. What is computing is where you can do your computing virtually anywhere. So, for example, you know, I've got, uh I've got a document. I started it on my laptop. I can then go and finish it sitting on the beach on my phone. Or, you know, I can go and do it in a coffee shop or a library. or wherever. So the idea of ubiquitous computing is similar in that, yes, there's computing everywhere, but it's more about your data being universally accessible. So essentially it is cloud computing. That is what this whole ubiquitous computing thing is about. >>Okay on that then differs from pervasive computing in the fact that pervasive is the devices that we have all around us versus the access to those devices. >>Exactly. It's it's really it's more about the data. So ubiquitous computing is more about My data is stored in some central place, and I could hit it from anywhere. There is a device, whereas pervasive computing is there is a device almost everywhere. Okay, so yeah, >>So why Why do we as Moran takes care about the vice of computing? >>Well, pervasive computing brings up a whole lot of new issues, and it's coming up really fast. I mean, you last night I was watching, you know, commercial where you know, somebody a woman's coming out and starting her car with her phone. Um, which sounds really cool. Um, but you know what they say Anything that you can access, you know, with your computer is hackable. So, you know, there are security issues that need to be considered when it comes to all of this, but that's that's the downside. But there's just this huge upside on pervasive computing that it's so exciting when you think about this. I mean, think about a world where remember I said your refrigerator might be attached to the network. Well, what if you could rent out space on your refrigerator to somebody someplace else in a secure way? Of course. You know what? If you could define your personal network as all of these devices that you own and it doesn't matter where your workloads run or, you know, you could define all of this stuff in such a way that the connectivity between objects is really huge. Um, so you know, I mean, you look at things like, you know, I f t t you know, it's like get a notification when the International space station passes over your house. Okay? I don't know why I would need that. Um, but it's the kind of thing that people >>would have a nine year old. You can run him outside and show Z. Oh, >>there you go. There you go. So I mean, that kind of level of connectivity between objects is really really it gives us this new level off. Uh, this new level of functionality that we would never even considered even 10 years ago. Um, it also extends the life of objects that we already have. So, you know, maybe you've got that, uh, that computerized vacuum cleaner, and you don't like the way that it you don't like the pattern that uses in your house. So you re program it or, uh Or I watched. I watched a guy decide that he didn't want to buy multiple vacuums for his house. So he programmed his programa will act Hume to fly between floors. It was actually pretty funny. Um, I it's some people just have too much time. >>It's driving the whole world of programmable at all levels. Really? Like the projects coming out of the car industry of creating a programmable car would fit into that category. Then, I >>suppose absolutely, absolutely needs developer tool kits. Um, that make it possible for anybody to re program these devices that you never would have thought of reprogramming before. So it's important. So do >>we want to talk about the questions. We would love people to give us some feedback on at this stage. >>I would love to talk about these questions. So what we did is we put together, uh, we put together a place for you to answer questions. If you're not watching this live. If you're watching this live, please go ahead. Drop your ideas in the chat. We would love to discuss them, you know. Do you want to see more of this? Or does it? Conversely, Does it scare you, Sean? You What? >>What do you >>think about these questions? >>Well, I mean, for me, the idea of the connected world at one level, the engineering me loves the idea. Another level. It comes to these questions of privacy. Vegas questions off. How do I control this going into the future? What prevents somebody from taking over my flying vacuum cleaner? I'm using it, you know? So it's an interesting question. I think there's a lot of cool, cool ideas. Yeah, and a lot of work to be done. I really want to hear other people's ideas as well and see how we can take this into the future. >>Definitely, definitely. I mean, look I mean, we're joking about it, but, you know, when somebody hacks into your grandmother's insulin pump, maybe not so funny. >>Yeah, a very real risk. >>A very real risk. A very real risk. But yeah, I mean, we'd love Thio. We'd love to hear how you'd like to see this used. So that's that's my That's kind of what I've been thinking about thes days. Um, but, you know, Sean, uh, now, you I know you are really concerned about this whole issue of developers and how they feel about infrastructure. So I would love to hear what you've got to say on that. >>Yeah, I'd like to sex, but a bit about that. You know, we we've done a lot of work over the last few years looking at how developing our history has been very focused on operations, but without big drive towards supporting developers providing better infrastructure for developers. One of the interesting things that keeps coming up to the four on Do you know, the way the world is changing is that big question is, do developers actually give a damn about infrastructure in any way, shape or form? Um, you know, ultimately more and more development languages and tools abstract that underlying infrastructure. What communities does is basically abstract. The infrastructure away, Um, mawr and more options. They're coming to market, which you can quite literally creating application without out of a writing a line of code. Um, so this morning, way Dio, we're doing it all the time, sometimes without even realizing it on. I think the definition of what a developer is is also changing to a certain extent. So you know the big question, which I have on which I'd like to understand Maureen, from talking to low developers is due. Developers care about infra What is it that you expect from infrastructure? What do they want going into the future? How are they going to interact with that infrastructure? I My personal opinion is that they don't really care about infrastructure, that they're going to find more ways to completely abstract away from that. And they just want to focus on delivering applications faster and getting value to market. But I might be wrong, and I'd really like to hear people's impact ideas and thoughts on that >>on. And that's exactly and that's why we're asking this question. Developers out there. Do you care? Or do you just want the whole thing completely abstracted away from you >>on? If you do care why, If you don't, what would you like to see? Another. It's a couple of questions to ask, but really like to hear those opinions on bond. You know, Do you just want the operations guys to live with it? You never want to hear about it again, just fine. It's actually good to say that we'll work it out. >>Yes, and that there's nothing. There's nothing wrong with pushing that up stack >>pretty much what we're trying to do here. >>Well, it is what we're trying to do. But at the same time, we want to do what's good for developers. And if you developers or like No, don't don't do that. Well, we want to know because, you know, we don't wanna work away here and some ivory tower and wind up with something that's not good for >>you after school. So cool. So, yeah, there are some other interesting things we're talking about. >>I know, I know. This is This is one of my favorites. This is one of my favorites. >>Zoo this? Yes. While >>we're on the subject of not getting involved with the infrastructure. Go ahead, Sean. Tell us about it. >>Thing is a pet topic of mine and something that that we've spoken about a lot. And thanks something that we we have spent many nights talking about. The idea is AI ops using artificial intelligence to drive operations within our infrastructure. And so a lot of people ask me, You know why? Um, essentially, What the hell is a I out on? I have answered this question many times, and it does often seem that we all take this AI ops thing for granted or look at it in a different way. To me, it is essentially, it's it's automation on steroids. That's what it boils down Thio. It's using intelligence systems that to replace the human cerebellum. I mean, let's just be blunt about this. We're trying to replace humans. Onda reason for that is we humans less meat sacks are airplane. We make mistakes all the time and compared to computers were incredibly slow. Um, you know, that's really the simplest point with the scale of modern infrastructure that we're dealing with the sheer volume. I mean, we've gone from, you know, thousands to tens of thousands of the EMS to now hundreds and thousands of containers spread across multiple time zones. Multiple places. We need to come up with better ways of managing this on the old fashioned stick through mechanism of automation. It's just too limited for that. Right >>when we say we want to replace meat sacks, we mean in a good way. >>We mean in a good way. I know it's a bit of a harsh way of putting it. Um, ultimately, humans have ability for creativity that machines just don't have. But machines can do other things, and they could do analysis of data a lot faster than we can. Quite often, we have to present that data to humans to have invalidate that information. But, you know, one of the options for us is to use artificial intelligence, quantified data, um, correlated, you know, look for root cause and then provide that information to us in such a way that we can make valid decisions based on that information a lot faster than we could otherwise, >>right? So what are the what are the implications? What are the practical implications of doing this so >>practically we can analyze massive amounts of data a lot faster than a single human. Could we even just a normal type system that's searching? We We have the tools to learn by looking at data and have machines do it a lot faster than we can. We can take action faster based on that data, because we get the data foster. We can take action and much more complex action that involves maybe many different layers of tasking much, much faster. Um, on we could start to do maintenance operations and maintenance tasks without having to wait for human beings to wake up or get to an office. But more importantly, we could start making tasks happen very complex tasks in a very specific orders, with much less potential for error. And those are the kinds of areas we're looking at. >>That's that's true. So how do you kind of see this moving forward? I mean, obviously, we're not gonna go from nothing to Skynet, and hopefully we never get to Skynet. Well, >>depends if you are in control of Skynet or not. Ultimately, Dionysus little computer. Um, practically speaking, we have a few things Thio hoops to jump through our suppose before we can look at where else is going to be really effective on the first one is a trust issue. We have to learn to trust it. And to do that, we have to put in a position where it can learn and start providing us that data analysis on that inference and then having humans validated. That's practically the very first step. No, it's a trust issue. You know, we've seen been watching sci fi for the last 30 years. Class on. Do you know the computers take over? Well, ultimately, is that real or not? Um, if we look at how we gonna get there? Probably midterm. Adaptive maintenance, maybe infrastructure orchestration. Smart allocation of resource is across cloud services. Well, >>we can talk for a minute About what that would would actually look like. So, I mean, we could talk about, you know, abs, midterms. I mean, in a practical sense, how would that actually work? >>Yeah, Okay. It's a great question. So, practically speaking, the first thing we're gonna do is we're going to start to collect all this data. We're gonna find all this data. I mean, the modern computer systems that we have infrastructure systems. We are producing many hundreds of gigabytes, sometimes terabytes of logging data every day. The majority of it gives far 13. I mean, we throw the majority of their logging information away or if it's not thrown away, it's stored some way for security purposes and never analyzed. So let's start by taking their data and actually analyzing it. To do that, we have laid and correlated, >>so we >>gotta put it all together. We've got a match it and we've got to start building patents. We're going to start looking for the patterns. This is where I is particularly good at starting to help us. Bold patterns start to look for those patterns. Initially, humans will have to do some training. Um, once we have that patent, once we've got that working, we can now start having the AI systems start to do some affairs. E, here's the recalls. So we the system can tell us based on the data based on the patterns we've been learning. We know from the past debt. If those three network links get full bad example, we're gonna have a failure in Region X, right. So start telling us while those network links of filling up tell us before they fall rather than after their full always they're falling up as we see trending information now seems like a simple I could do trending information with just normal monitoring systems. But if I can start to correlate that with greater users in, you know, Beijing Office versus Users in California office filling up those links and different times of the day, I can now start to make much more clever decisions, which is a human on its own, to try and correlate that information, which is be insane once you've done that way to go to the next stage, which is not to have the system act do actions for us. Based on that information right now, we're starting to get close to the scan it. Speaking of this doesn't have to be a big, complex pile of change. Smart ai solution. I have data on that AI solution is talking to my existing automation solutions to action. That change. That's how I see this moving forward, >>right? So essentially you, instead of saying, you know, deploy this too. Uh, this workload to AWS, you would say deploy this. Yeah, And then the system would look and go. Okay, It's this kind of workload. At this time of day at this size, it's gonna interact with this and this and this. And so it's gonna be best off in this region of this cloud provider on then. Uh, you know, two days from now, when the prices drop, we're gonna put it over there, >>even taking a different different. Spoken exactly that it could be. The Beijing office is coming online. Let's move the majority of the workload to a cloud that's closer to them. Reducing the network bandwidth. Yeah, and inference. Andi Also reducing the impact on international lines as Beijing winds down for the day, I can just move the majority of the workload into California on board Europe. In between, it's very simple examples, but have humans do that would be very complex and very time consuming >>exactly. And end. Just having humans notice those patterns would be difficult. But once you have the system noticing those patterns, then the humans could start to think, How can I take advantage of this, you know, So as you are talking about much longer term in the actual applicant patients themselves. So you know, everything can be optimized that way so >>everybody may optimized way can optimize down to the way we even potentially write applications in the future. Humans were still deciding the base logic. Humans were still deciding the creative components of that. Right as we as we build things, we can start to optimize them, breaking down into smaller and smaller units that are much more specific. But the complexity goes up. When we do that right. I want to use AI and AI solutions to start to manage that complexity across multiple spaces. Multiple time zones, etcetera. >>Exactly. Exactly. So. So that's the question, you know. What do you guys think? You know, we really want to know >>on Dhere again. You know, we mentioned this around the beginning, but do you think you could trust in a iob sedition? What would it take for you to trust in our absolution? And where do you practically see it being used in the short term? >>Yeah, that's that's the big question is where do you see it being used? Where would you like it to be used, you know? Is there something that you don't think would be possible, but you would like to see it, you know. But the main thing is, on a practical matter, what would you like to see? >>Let me ask. The question is like a different way. Do you have a problem that we could solve within a isolation today? E, They're really well >>right. A re a world problem. And And assuming that, you know, we are not gonna, you know, take over the world. >>Yeah. Important. My evil plan is to take over the world with >>man. I'm so sorry. First >>had to let that draw. >>I did. I did. I'm so sorry. Okay, Alright. So that's so That's a I ops. And we like I said, if you're watching this live, throw in the chat. We want to hear your ideas. If your, um if you're doing this, if you're watching this on the replay, go to the survey because we way, we really want to hear your ideas and your opinions. All right, So moving right along. All right. What the heck are you know, kernels? >>Uh, lovely questions. So, you know, the whole world is talking about containers today way we're talking about containers today. But containers like VMS or just one way to handle compute Andi. They're more and more ideas that are out there today, and people have been trying different ways off, shrinking the size of the compute environment. COMPUTER Paxil Another cool way of looking at this and saying That's been around for a little while. But it's getting your attraction to learn to sing called unique kernels, and what they are is they're basically highly optimized. Execute a bles that include the operating system, Um, there on OS settle libraries, um, and some very simplified application code all mixed into a very, very tiny package. Easiest way to describe them. They're super simplified. And I were talking about in the eye ops discussion this idea off taking everything into smaller and smaller individual functions but creating a certain level of complexity. Well, if we look at uni kernels, those are those smaller and smaller bundles and functions. They interact directly with the hardware or through a hyper visor. Um, so actually, no overhead. I mean the overhead If you just look at what a modern you clinics operating system is made up of these days, there are so many different parts and components. Even just the colonel has got anything from, you know, 5 to 7 different parts to it. Plus, of course, drivers and a boot loader. Then we look at the system libraries that set on top of that, you know? And then they're demons and utilities and shells and scream components and, you know, additional colonel stacks that go on top of that for hyper visors. What we're trying to say is, what, This text of space, I'm >>getting tired. Just listening, >>Thio. I'm tired talking about it. You know that the unique colonel, really, it just takes over their complexity. It puts the application the OS on the basic libraries necessary. That application in tow, one really tiny package. Um, yeah. Give you an idea what we're talking about here. We're talking about memory footprints or time package footprints in the kilobytes. You know, a small container is considered 100 make plus, we're talking kilobytes. We're talking memory utilization in the kilobyte two megabytes space because there's no no fact, no fluff, no unnecessary components. And then only the CPU that it needs. >>So Bill Gates was right 6. 40 k is all anybody will ever need >>Potentially. Yeah, right. E, there was there was an IBM CEO who said even less at some point. So we'll see >>how that go. What goes around comes around. >>But one of the really interesting things about this small size, which is really critical, is how fast they can boot. Yeah, we're talking boot times measured in 30 seconds. Wow, We're talking the ability to spin up specific functions only when you need them. Now, if we look at the knock on effect of that, we're looking at power saving. Who knew? Run the app when I need it because there's no Leighton. See to start it up. The app is tiny so I can pack a lot mawr into a lot less space game power seconds. But when I start looking at where you were talking about earlier, which the basic compute idea in the world all of a sudden that tiny little arm chips it in my raspberry pi that's running my fridge, My raspberry pi equivalent that's running my fridge no longer has a fact operating system around it. I can run tens thousands, potentially off these very tiny specific devices when I need them. Wow, I'm kind of excited about it. I'm excited by the idea. You >>can hear that >>I'm a hardware geek from from many, many moons ago on DSO. I kind of like the idea of being able to better utilize along this very low powered hardware that we have lying around and really take it into the future. Well, that's good. Yeah. So I'm not going to kill, not going to kill containers. But it is a parallel technology that I'm very interested in >>that that is true. Now what does it I mean in terms of, like, attack surface. That means it's got a much smaller attack surface, though, right? >>Yeah. Great. Great point. I mean, there's no there's no fluff. There's no extra components in the system. Therefore, the attack surface is very, very small. Um, you know, and because they're so small and can be distributed much, much faster and much more easily updating and upgrading them as much easier way can we can upgrade a 60 k b file across a GPRS connection on which I certainly can't do with 600 make, uh, four gig VM 600 made container. You know, just unrealistic. Um, e >>I was just going to say so. So now these. You know, kernels, they're they're so small. And they have on Lee what they absolutely need. Now, how do you access the hardware? >>So the hardware is accessed via hyper visor. So you have to have some kind of hyper visor running on top of the hard way. But because Because we need very little from their type adviser, we don't actually need to interact with that very much. It could be a very cut down operating system. Very, very simplified operating system. We're also not trying to run another layer on top of that. We're not We're not ending up with multiple potential VMS or something underneath it were completely removed. That layer, um, the the drivers, the necessary drivers are built into that particular colonel device. >>Oh, okay. That makes sense. >>Tiny footprint easily distributed, um, and once again, very specialized, >>right? Right. Well, that makes sense. Okay. So, yeah, I mean, I guess so. These these individual stacks, you know, comparing virtual machines to containers to unit colonels, there just a completely different architecture. But I can see how that would How That would work where you have the hi perverse. A little hyper buys are on top of rented teeth. OK, so moving right along certain. Where do we see these being used? >>Um, it's early days, although there are some very good practical applications out there. There's a big, big ecosystem of people trying different ways for this I o ts off the obvious immediate place. I i o t s a quick, easy place for something very specialized. Um, what's interesting to me? And you mentioned this earlier. You know, we're talking about medical devices. We're talking about potentially disposable medical devices. Now, if I can keep those devices to run on really low power very, very cheap, um, CPUs and all of a sudden I've got a device that is available to a lot more people. I don't need a massive, powerful CPU. I just need saying that runs a very specific function really fast, A very small scale. I could do well disposable devices. I can build medical devices that are so small we can potentially swallow them and other areas which are really interesting. And I spoke a little bit about it, but it's energy efficiency. Where We need to be very, very energy efficient. No. And that can also impact on massively scalable systems where I want to deal with tens of thousands of potential transactions from users going into a system. I can spin them up only when I need them. I don't need to keep them running all the time again. It comes back to that low latency on then. Anyway, that an incredibly fast food time is valuable. Um, a car, you know, Think about it. If if my if my electric car is constantly draining that battery when it's parked in the garage and I'm traveling or if it takes 20 minutes from my car to boot up its clinics. Colonel, when I wanted, I'm going to get very irritated. Well, >>that and if you have a specific function, you know, like, identify that thing, Yeah, it would be good if you haven't smashed into it before. Identified it as a baby carriage e dark today. Yes. >>So, Nick, you know, these is all really interesting topics. Um, yeah. We spoke about air ops. We spoke about the impact is gonna have on humans. Um, all of these changes to the world that we're living in from computer systems, the impact it's having on our lives biggest. An interesting question about the ethics of all of this >>ethics of all of this. Yes, because let's be let's be realistic. There are actual riel concerns when it comes to privacy, when it comes to how corporations operate, when it comes to how governments operate. Um, there are areas of the world's where, how all of this has has moved, it's absolutely I'll be honest, absolutely terrifying the economic disparity. Um, but when you really come right down to it, um, it's all about the human control over the technology because all of these ethical issues are are in our hands. Okay, we could joke about Sky Net. We can joke about things like that, but this is one place that technology can't help us. We have to do this. We have to be aware of what's going on. We have to be aware. Are they using facial recognition? Uh, you know, when you go to X y Z, are they using recidivism algorithms in sentencing? And how is that? How is that going? Is it? Are those algorithms fair? Certain groups get longer sentences because historical data, uh, is skewed. Be educated. Know how this works? Don't be afraid of any of this. None of this is, uh, none of this is rocket science. Really? Come right down to it. I mean, it's it's not simple, but you can learn this. You can do it. >>Ask good questions. Be interested to be part of the part of the discussion. Not just a passive bystander. >>Exactly. Don't just complain about what you think is going on. Learn about what is actually going on and be active, where you see something that needs to be fixed. So that's what that's what we can do about it. We need to be aware that there's an issue or potential issues, and we need to step in and fix it. So that z myself box, I'll step down zone >>important topic. And it's one that we all can have influence on on bits one. Those who are us who are actually involved in building these systems for the future. We can help make sure that the rules are there. That's right. Systems are built correctly on that. We have open dialogues and discussions around these points and topics and on going away, was she? I think we're coming to the end of the time on hopefully we've kept everybody interested in some of the things that we think are cool for the future. And we're putting our efforts into E O. But I think we need to wrap this up now. So, Nick, great chatting to you is always >>always, always a pleasure, Sean. >>It's been an amazing week. Um, been amazing. Couple of weeks, everybody leading up to this event on bond. No, thank you, everybody for listening to us. Please go and download and try. Dr. Enterprise, Uh, the container card is available. Will post the links here to better understand what we've been doing. Go and have a look through the tutorial track. You'll hear my voice. I'm sure you'll hear next voice and make other people's voices through those tutorials. Hopefully, we keep you all interested and then going download and try lens, Please. Finally, we want your feedback. We're interested to hear what you think would be the great ideas. Good, Bad. Otherwise let us know what you think about products. We are striving to make them better all the time. >>Absolutely. And we want your involvement. Was it all right? Thank you all. Bye bye. Yeah,