AWS re:Invent 2022 | Justin Borgman | Starburst & Ashwin Patil | Deloitte

News
More from AWS re:Invent 2022

Ashwin Patil

Principal, AI & Data Engineering

Deloitte

Data as a product: using data products to successfully drive business outcomes

With a data mesh, enterprises can revolutionize the way they approach data-based architecture and management.As data grows in value and brings new ways to generate revenue and insights, businesses are adopting a myriad of data products and solutions to manage the massive amount of data produced.“Increasingly in today’s world, people are realizing the data will always be decentralized to some degree,” said Justin Borgman (pictured, left), chairman and chief executive officer of Starburst Data Inc

Cloud-native application security firm Apiiro raises $100M round

Israeli application security startup Apiiro LLC today announced an impressive Series B round of $100 million, led by General Catalyst with participation by Greylock and Kleiner Perkins.The company, which was founded by former employees of Microsoft Corp. and has offices in New York and Tel Aviv, raised $35 million in funding in 2020 from Kleiner Perkins and Greylock. Today’s round shows significant trust at a time when investors are increasingly keeping a tight hold on their purse strings.Apiiro offers software to address security issues before developers release code to the cloud. That should ensure that vulnerable code is discovered before hackers have a chance to exploit it

Investors hammer Atlassian on earnings miss and weak outlook

Investors hammered Atlassian Corp. stock in after-hours trading after the Australian software company missed earnings forecasts and gave a lower outlook than had been expected by analysts.For its fiscal quarter ended Sept. 30, Atlassian reported net income of $92.5 million, or 36 cents per diluted share, down from $94.4 million, or 37 cents in the same quarter of last year. Revenue rose 31%, to $807.4 million. Analysts had been expecting earnings per share of 38 cents on revenue of $806.4 million.Atlassian added 6,550 new customers in the quarter, its 10th quarter of customer growth, bringing its total number of customers to 249,173

Datadog beats Wall Street’s expectations as its revenue grows by more than 60%

Application monitoring software firm Datadog Inc. delivered strong third-quarter financial results today, beating expectations on earnings and revenue, and offered solid guidance for the coming quarter too, but its stock barely moved.The company reported a net loss of $26 million for the quarter, with earnings before certain costs such as stock compensation coming to 23 cents per share. Revenue for the period rose by an impressive 61% from a year earlier to $436.5 million. The numbers came in comfortably ahead of Wall Street’s forecasts. Analysts were modeling earnings of just 15 cents per share on revenue of $412.2 million

Despite smashing earnings forecasts, Cloudflare sees shares plunge on growth concerns

Shares in content delivery network company Cloudflare Inc. plunged in late trading even though it beat all expected metrics in its latest earnings report as investors were reportedly spooked by figures suggesting moderating growth.For its third quarter that ended Sept. 30, Cloudflare reported a profit before costs such as stock compensation of $19.1 million, or six cents per share, up from $1.4 million or breakeven in the same quarter of 2021. Revenue in the quarter jumped 47%, to $253.9 million. Analysts had expected breakeven earnings per share on revenue of $250.63 million.Adjusted gross profit was $198.4 million or 78.1% gross margin, up from $136.6 million or 79

Twilio’s stock crumbles on poor revenue guidance

Shares of the cloud communications firm Twilio Inc. nosedived in extended trading today after it issued a soft earnings and revenue outlook for the coming quarter.The company did OK in the quarter just gone, reporting a third-quarter loss before certain costs such as stock compensation of 27 cents per share, ahead of Wall Street’s forecast of a 36-cent loss. Revenue for the period came to $983 million, beating the consensus estimate of $972.2 million. Overall, Twilio reported a net loss of $482.2 million, much bigger than the $224.1 million loss it delivered in the same period last year

Block shares surge on strong earnings beat

Shares in Block Inc. surged in late trading after the payments company delivered a strong earnings beat in its latest quarterly report.For its fiscal third quarter, Block reported net income before costs such as stock compensation of 42 cents a share, up from 25 cents a share in the same quarter of last year. Revenue rose 18% from a year ago, to $4.52 billion. Analysts had been expecting earnings per share of 23 cents on revenue of $4.49 billion.Gross profit in the quarter rose 38%, to $157 billion, marking the fifth straight quarter of increasing profit

Dropbox delivers on earnings and revenue targets with slow and steady growth

Dropbox Inc. delivered strong third-quarter financial results today, beating Wall Street’s forecasts on earnings and revenue and offering guidance above expectations.The file-sharing company reported earnings before certain costs such as stock compensation of 43 cents per share, with revenue coming to $591 million in the period, up 7.4% from a year earlier. Altogether, the company netted a profit of $83.2 million.The results were good, since Wall Street analysts were modeling lower earnings of just 37 cents per share on sales of $585.5 million. Dropbox’s share price stayed flat in after-hours trading

Higher-than-expected user numbers elevate Coinbase shares

Shares in Coinbase Global Inc. rose slightly in late trading after the cryptocurrency exchange reported better-than-expected user numbers amid misses on its latest quarterly earnings and revenue.For its fiscal third quarter that ended Sept. 30, Coinbase reported a loss before costs such as stock compensation of $2.43 a share on revenue of $590 million. Analysts had expected an adjusted loss of $2.40 a share on revenue of $654 million.Monthly transacting users in the quarter came in at 8.5 million, down from 9 million in the previous quarter but up from 7.3 million in the third quarter of last year. Analysts had been expecting 7.84 million

Growing tech landscape shines spotlight on need for enhanced data culture

The data revolution is not just about technology innovation, but also cultural and community transformation.Since a data citizen is anyone who uses data to do their job, enterprises have to cultivate a conducive data culture so that employees can hone their skills and competencies in this space, according to Stijn “Stan” Christiaens (pictured), co-founder and chief data citizen at Collibra NV, who believes that in today’s organizations, most employees will be data citizens

Cisco Partner Summit 2022 kicks off the Age of the Partner

Cisco Systems Inc. held its annual Partner Summit this week in Las Vegas, bringing together thousands of global partners to exchange ideas and learn about the networking giant’s most recent innovations. Cisco has one of the largest and most diverse partner ecosystems in the information technology industry, with about 90% of the company’s revenue flowing through its partners from more than 150 countries.The event returned to an in-person format for the first time since 2019, which is crucial for this event. Some events are fine done virtually but partner relationships for all vendors, including Cisco, requires a high level of engagement and that can’t be simulated virtually

Info

Justin Borgman | Starburst & Ashwin Patil | Deloitte

Ashwin Patil

Principal, AI & Data Engineering Deloitte

search

(electronic music) (graphics whoosh) (graphics tinkle)

>> Welcome to Las Vegas! It's theCUBE live at AWS re:Invent '22. Lisa Martin here with Dave Vellante. Dave, it is not only great to be back, but this re:Invent seems to be bigger than last year for sure.

>> Oh, definitely. I'd say it's double last year. I'd say it's comparable to 2019. Maybe even a little bigger, I've heard it's the largest re:Invent ever. And we're going to talk data, one of our favorite topics.

>> We're going to talk data products. We have some great guests. One of them is an alumni who's back with us. Justin Borgman, the CEO of Starburst, and Ashwin Patil also joins us, Principal AI and Data Engineering at Deloitte. Guys, welcome to the program.

>> Thank you.

>> Together: Thank you.

>> Justin, define data products. Give us the scoop, what's goin' on with Starburst. But define data products and the value in it for organizations of productizing data.

>> Mm-hmm. So, data products are curated data sets that are able to span across multiple data sets. And I think that's what's makes it particularly unique, is you can span across multiple data sources to create federated data products that allow you to really bring together the business value that you're seeking. And I think ultimately, what's driving the interest in data products is a desire to ultimately facilitate self-service consumption within the enterprise. I think that's the holy grail that we've all been building towards. And data products represents a framework for sort of how you would do that.

>> So, monetization is not necessarily a criterion?

>> Not necessarily. (Dave's voice drowns)

>> But it could be.

>> It could be. It can be internal data products or external data products. And in either case, it's really intended to facilitate easier discovery and consumption of data.

>> Ashwin, bringing you into the conversation, talk about some of the revenue drivers that data products can help organizations to unlock.

>> Sure. Like Justin said, there are internal and external revenue drivers. So internally, a lot of clients are focused around, hey, how do I make the most out of my modernization platform? So, a lot of them are thinking about what AI, what analytics, what can they run to drive consumption? And when you think about consumption, consumption typically requires data from across the enterprise, right? And data from the enterprise is sometimes fragmented in pieces, in places. So, we've gone from being data in too many places to now, data products, helping bring all of that together, and really aid, drive business decisions faster with more data and more accuracy, right? Externally, a lot of that has got to do with how the ecosystems are evolving for data products that use not only company data, but also the ecosystem data that includes customers, that include suppliers and vendors.

>> I mean, conceptually, data products, you could say have been around a long time. When I think of financial services, I think that's always been a data product in a sense. But suddenly, there's a lot more conversation about it. There's data mesh, there's data fabric, we could talk about that too, but why do you think now it's coming to the fore again?

>> Yeah, I mean, I think it's because historically, there's always been this disconnect between the people that understand data infrastructure, and the people who know the right questions to ask of the data. Generally, these have been two very distinct groups. And so, the interest in data mesh as you mentioned, and data products as a foundational element of it, is really centered around how do we bring these groups together? How do we get the people who know the data the best to participate in the process of creating data to be consumed? Ultimately, again, trying to facilitate greater self-service consumption. And I think that's the real beauty behind it. And I think increasingly, in today's world, people are realizing the data will always be decentralized to some degree. That notion of bringing everything together into one single database has never really been successfully achieved, and is probably even further from the truth at this point in time, given you've got data on-prem and multiple clouds, and multiple different systems. And so, data products and data mesh represents, again, a framework for you to sort of think about data that lives everywhere.

>> We did a session this summer with (chuckles) Justin and I, and some others on the data lies. And that was one of the good ol' lies, right? There's a single source of truth.

>> Justin: Right.

>> And all that is, we've probably never been further from the single source of truth. But actually, you're suggesting that there's maybe multiple truths that the same data can support. Is that a right way to think about it?

>> Yeah, exactly. And I think ultimately, you want a single point of access that gives you, at your fingertips, everything that your organization knows about its business today. And that's really what data products aims to do, is sort of curate that for you, and provide high quality data sets that you can trust, that you can now self-service to answer your business question.

>> One of the things that, oh, go ahead.

>> No, no, I was just going to say, I mean, if you pivot it from the way the usage of data has changed, right? Traditionally, IT has been in the business of providing data to the business users. Today, with more self-service being driven, we want business users to be the drivers of consumption, right? So if you take that backwards one step, it's basically saying, what data do I need to support my business needs, such that IT doesn't always have to get involved in providing that data, or providing the reports on top of that data? So, the data products concept, I think supports that thinking of business-led technology-enabled, or IT-enabled really well.

>> Business led. One of the things that Adam Zelinsky talked with John Furrier about just a week or so ago in their pre re:Invent interview, was talking about the role of the data analyst going away. That everybody in an organization, regardless of function, will be able to eventually be a data analyst, and need to evaluate and analyze data for their roles. Talk about data products as a facilitator of that democratization.

>> Yeah. We are seeing more and more the concept of citizen data scientists. We are seeing more and more citizens AI. What we are seeing is a general trend, as we move towards self-service, there is going to be a need for business users to be able to access data when they want, how they want, and merge data across the enterprise in ways that they haven't done before, right? Technology today, through products like data products, right, provides you the access to do that. And that's why we are going to see this movement of people of seeing people become more and more self-service oriented, where you're going to democratize the use of AI and analytics into the business users.

>> Do you think, when you talk to a data analyst, by the way, about that, he or she will be like, yeah, mm, maybe, good luck with that. So, do ya think maybe there's a sort of an interim step? Because we've had these highly, ZeMac lays this out very well. We've had these highly-centralized, highly-specialized teams. The premise being, oh, that's less expensive. Perhaps data analysts, like functions, get put into the line of business. Do you see that as a bridge or a stepping stone? Because it feels like it's quite a distance between what a data analyst does today, and this nirvana that we talk about. What are your thoughts on that?

>> Yeah, I mean, I think there's possibly a new role around a data product manager. Much the way you have product managers in the products you actually build to sell, you might need data product managers to help facilitate and curate the high quality data products that others can consume. And I think that becomes an interesting and important, a skill set. Much the way that data scientist was created as a occupation, if you will, maybe 10 years ago, when previously, those were statisticians, or other names.

>> Right. A big risk that many clients are seeing around data products is, how do you drive governance? And to that, to the point that Justin's making, we are going to see that role evolve where governance in the world, where data products are getting democratized is going to become increasingly important in terms of how are data products being generated, how is the propensity of data products towards a more governed environment being managed? And that's going to continue to play an important role as data products evolve.

>> Okay, so how do you guys fit, because you take ZeMac's four principles, domain ownership, data as product. And that creates two problems. Governance. (chuckles) Right? How do you automate, and self-service, infrastructure and automated governance.

>> Yep.

>> Tell us what role Starburst plays in solving all of those, but the latter two in particular.

>> Yeah. Well, we're working on all four of those dimensions to some degree, but I think ultimately, what we're focused today is the governance piece, providing fine-grained access controls, which is so important, if you're going to have a single point of access, you better have a way of controlling who has access to what. But secondly, data products allows you to really abstract away or decouple where the data is stored from the business meaning of the data. And I think that's what's so key here is, if we're going to ultimately democratize data as we've talked about, we need to change the conversation from a very storage-centric world, like, oh, that table lives in this system or that system, or that system. And make it much more about the data, and the value that it represents. And I think that's what data products aims to do.

>> What about data fabric? I have to say, I'm confused by data fabric. I read this, I feel like Gartner just threw it in there to muck it up. And say, no, no, we get to make up the terms, but I've read data mesh versus data fabric, is data fabric just more sort of the physical infrastructure? And data mesh is more of an organizational construct, or how do you see it?

>> Yeah, I'm happy to take that or. So, I mean, to me, it's a little bit of potato potato. I think there are some subtle differences. Data fabric is a little bit more about data movement. Whereas, I think data mesh is a little bit more about accessing the data where it lies. But they're both trying to solve the similar problem, which is that we have data in a wide variety of different data sets. And for us to actually analyze it, we need to have a single view.

>> Because Gartner hype cycle says data mesh

is DOA- >> Justin: I know.

>> Which I think is complete BS, I think is real. You talk to customers that are doing it, they're doing it on AWS, they're trying to extend it across clouds, I mean, it's a real trend. I mean, anyway, that's how I see it.

>> Yeah. I feel the word data fabric many a times gets misused. Because when you think about the digitization movement that happened, started almost a decade ago, many companies tried to digitize or create digital twins of their systems into the data world, right? So, everything has an underlying data fabric that replicates what's happening transactionally, or otherwise in the real world. What data mesh does is creates structure that works complimentary to the data fabric, that then lends itself to data products, right? So to me, data products becomes a medium, which drives the connection between data mesh and data fabric into the real world for usage and consumption.

>> You should write for Gartner. (all laugh) That's the best explanation I've heard. That made sense!

>> That really did. That was excellent. So, when we think about any company these days has to be a data company, whether it's your grocery store, a gas station, a car dealer, what can companies do to start productizing their data, so that they can actually unlock new revenue streams, new routes to market? What are some steps and recommendations that you have? Justin, we'll start with you.

>> Sure. I would say the first thing is find data that is ultimately valuable to the consumers within your business, and create a product of it. And the way you do that at Starburst is allow you to essentially create a view of your data that can span multiple data sources. So again, we're decoupling where the data lives. That might be a table that lives in a traditional data warehouse, a table that lives in an operational system like Mongo, a table that lives in a data lake. And you can actually join those together, and represent it as a view, and now make it easily consumable. And so, the end user doesn't need to know, did that live in a data warehouse, an operational database, or a data lake? I'm just accessing that. And I think that's a great, easy way to start in your journey. Because I think if you absorb all the elements of data mesh at once, it can feel overwhelming. And I think that's a great way to start.

>> Irrespective of physical location.

>> Yes.

>> Right?

>> Precisely. Yep, multicloud, hybrid cloud, you name it.

>> And when you think about the broader landscape, right? For the traditionally, companies that only looked at internal data as a way of driving business decisions. More and more, as things evolve into industry, clouds, or ecosystem data, and companies start going beyond their four walls in terms of the data that they manage or the data that they use to make decisions, I think data products are going to play more and more an important part in that construct where you don't govern all the data that our entities within that ecosystem will govern parts of their data, but that data lives together in the form of data products that are governed somewhat centrally. I mean, kind of like a blockchain system, but not really.

>> Justin, for our folks here, as we kind of wrap the segment here, what's the bumper sticker for Starburst, and how you're helping organizations to really be able to build data products that add value to their organization?

>> I would say analytics anywhere. Our core ethos is, we want to give you the ability to access data wherever it lives, and understand your business holistically. And our query engine allows you to do that from a query perspective, and data products allows you to bring that up a level and make it consumable.

>> Make it consumable. Ashwin, last question for you, here we are, day one of re:Invent, loads of people behind us. Tomorrow all the great keynotes kick up. What are you hoping to take away from re:Invent '22?

>> Well, I'm hoping to understand how all of these different entities that are represented here connect with each other, right? And to me, Starburst is an important player in terms of how do you drive connectivity. And to me, as we help plans from a Deloitte perspective, drive that business value, connectivity across all of the technology players is extremely important part. So, integration across those technology players is what I'm trying to get from re:Invent here.

>> And so, you guys do, you're dot connectors. (Ashwin chuckles)

>> Exactly, excellent. Guys, thank you so much for joining David and me on the program tonight. We appreciate your insights, your time, and probably the best explanation of data fabric versus data mesh. (Justin chuckles) And data products that we've maybe ever had on the show! We appreciate your time, thank you.

>> Together: Thank you- >> Thanks, guys.

>> All right. For our guests and Dave Vellante, I'm Lisa Martin, you're watching theCUBE, the leader in enterprise and emerging tech coverage. (electronic music)

Justin Borgman | Starburst & Ashwin Patil | Deloitte

search

(electronic music) (graphics whoosh) (graphics tinkle)

>> Welcome to Las Vegas! It's theCUBE live at AWS re:Invent '22. Lisa Martin here with Dave Vellante. Dave, it is not only great to be back, but this re:Invent seems to be bigger than last year for sure.

>> Oh, definitely. I'd say it's double last year. I'd say it's comparable to 2019. Maybe even a little bigger, I've heard it's the largest re:Invent ever. And we're going to talk data, one of our favorite topics.

>> We're going to talk data products. We have some great guests. One of them is an alumni who's back with us. Justin Borgman, the CEO of Starburst, and Ashwin Patil also joins us, Principal AI and Data Engineering at Deloitte. Guys, welcome to the program.

>> Thank you.

>> Together: Thank you.

>> Justin, define data products. Give us the scoop, what's goin' on with Starburst. But define data products and the value in it for organizations of productizing data.

>> Mm-hmm. So, data products are curated data sets that are able to span across multiple data sets. And I think that's what's makes it particularly unique, is you can span across multiple data sources to create federated data products that allow you to really bring together the business value that you're seeking. And I think ultimately, what's driving the interest in data products is a desire to ultimately facilitate self-service consumption within the enterprise. I think that's the holy grail that we've all been building towards. And data products represents a framework for sort of how you would do that.

>> So, monetization is not necessarily a criterion?

>> Not necessarily. (Dave's voice drowns)

>> But it could be.

>> It could be. It can be internal data products or external data products. And in either case, it's really intended to facilitate easier discovery and consumption of data.

>> Ashwin, bringing you into the conversation, talk about some of the revenue drivers that data products can help organizations to unlock.

>> Sure. Like Justin said, there are internal and external revenue drivers. So internally, a lot of clients are focused around, hey, how do I make the most out of my modernization platform? So, a lot of them are thinking about what AI, what analytics, what can they run to drive consumption? And when you think about consumption, consumption typically requires data from across the enterprise, right? And data from the enterprise is sometimes fragmented in pieces, in places. So, we've gone from being data in too many places to now, data products, helping bring all of that together, and really aid, drive business decisions faster with more data and more accuracy, right? Externally, a lot of that has got to do with how the ecosystems are evolving for data products that use not only company data, but also the ecosystem data that includes customers, that include suppliers and vendors.

>> I mean, conceptually, data products, you could say have been around a long time. When I think of financial services, I think that's always been a data product in a sense. But suddenly, there's a lot more conversation about it. There's data mesh, there's data fabric, we could talk about that too, but why do you think now it's coming to the fore again?

>> Yeah, I mean, I think it's because historically, there's always been this disconnect between the people that understand data infrastructure, and the people who know the right questions to ask of the data. Generally, these have been two very distinct groups. And so, the interest in data mesh as you mentioned, and data products as a foundational element of it, is really centered around how do we bring these groups together? How do we get the people who know the data the best to participate in the process of creating data to be consumed? Ultimately, again, trying to facilitate greater self-service consumption. And I think that's the real beauty behind it. And I think increasingly, in today's world, people are realizing the data will always be decentralized to some degree. That notion of bringing everything together into one single database has never really been successfully achieved, and is probably even further from the truth at this point in time, given you've got data on-prem and multiple clouds, and multiple different systems. And so, data products and data mesh represents, again, a framework for you to sort of think about data that lives everywhere.

>> We did a session this summer with (chuckles) Justin and I, and some others on the data lies. And that was one of the good ol' lies, right? There's a single source of truth.

>> Justin: Right.

>> And all that is, we've probably never been further from the single source of truth. But actually, you're suggesting that there's maybe multiple truths that the same data can support. Is that a right way to think about it?

>> Yeah, exactly. And I think ultimately, you want a single point of access that gives you, at your fingertips, everything that your organization knows about its business today. And that's really what data products aims to do, is sort of curate that for you, and provide high quality data sets that you can trust, that you can now self-service to answer your business question.

>> One of the things that, oh, go ahead.

>> No, no, I was just going to say, I mean, if you pivot it from the way the usage of data has changed, right? Traditionally, IT has been in the business of providing data to the business users. Today, with more self-service being driven, we want business users to be the drivers of consumption, right? So if you take that backwards one step, it's basically saying, what data do I need to support my business needs, such that IT doesn't always have to get involved in providing that data, or providing the reports on top of that data? So, the data products concept, I think supports that thinking of business-led technology-enabled, or IT-enabled really well.

>> Business led. One of the things that Adam Zelinsky talked with John Furrier about just a week or so ago in their pre re:Invent interview, was talking about the role of the data analyst going away. That everybody in an organization, regardless of function, will be able to eventually be a data analyst, and need to evaluate and analyze data for their roles. Talk about data products as a facilitator of that democratization.

>> Yeah. We are seeing more and more the concept of citizen data scientists. We are seeing more and more citizens AI. What we are seeing is a general trend, as we move towards self-service, there is going to be a need for business users to be able to access data when they want, how they want, and merge data across the enterprise in ways that they haven't done before, right? Technology today, through products like data products, right, provides you the access to do that. And that's why we are going to see this movement of people of seeing people become more and more self-service oriented, where you're going to democratize the use of AI and analytics into the business users.

>> Do you think, when you talk to a data analyst, by the way, about that, he or she will be like, yeah, mm, maybe, good luck with that. So, do ya think maybe there's a sort of an interim step? Because we've had these highly, ZeMac lays this out very well. We've had these highly-centralized, highly-specialized teams. The premise being, oh, that's less expensive. Perhaps data analysts, like functions, get put into the line of business. Do you see that as a bridge or a stepping stone? Because it feels like it's quite a distance between what a data analyst does today, and this nirvana that we talk about. What are your thoughts on that?

>> Yeah, I mean, I think there's possibly a new role around a data product manager. Much the way you have product managers in the products you actually build to sell, you might need data product managers to help facilitate and curate the high quality data products that others can consume. And I think that becomes an interesting and important, a skill set. Much the way that data scientist was created as a occupation, if you will, maybe 10 years ago, when previously, those were statisticians, or other names.

>> Right. A big risk that many clients are seeing around data products is, how do you drive governance? And to that, to the point that Justin's making, we are going to see that role evolve where governance in the world, where data products are getting democratized is going to become increasingly important in terms of how are data products being generated, how is the propensity of data products towards a more governed environment being managed? And that's going to continue to play an important role as data products evolve.

>> Okay, so how do you guys fit, because you take ZeMac's four principles, domain ownership, data as product. And that creates two problems. Governance. (chuckles) Right? How do you automate, and self-service, infrastructure and automated governance.

>> Yep.

>> Tell us what role Starburst plays in solving all of those, but the latter two in particular.

>> Yeah. Well, we're working on all four of those dimensions to some degree, but I think ultimately, what we're focused today is the governance piece, providing fine-grained access controls, which is so important, if you're going to have a single point of access, you better have a way of controlling who has access to what. But secondly, data products allows you to really abstract away or decouple where the data is stored from the business meaning of the data. And I think that's what's so key here is, if we're going to ultimately democratize data as we've talked about, we need to change the conversation from a very storage-centric world, like, oh, that table lives in this system or that system, or that system. And make it much more about the data, and the value that it represents. And I think that's what data products aims to do.

>> What about data fabric? I have to say, I'm confused by data fabric. I read this, I feel like Gartner just threw it in there to muck it up. And say, no, no, we get to make up the terms, but I've read data mesh versus data fabric, is data fabric just more sort of the physical infrastructure? And data mesh is more of an organizational construct, or how do you see it?

>> Yeah, I'm happy to take that or. So, I mean, to me, it's a little bit of potato potato. I think there are some subtle differences. Data fabric is a little bit more about data movement. Whereas, I think data mesh is a little bit more about accessing the data where it lies. But they're both trying to solve the similar problem, which is that we have data in a wide variety of different data sets. And for us to actually analyze it, we need to have a single view.

>> Because Gartner hype cycle says data mesh

is DOA- >> Justin: I know.

>> Which I think is complete BS, I think is real. You talk to customers that are doing it, they're doing it on AWS, they're trying to extend it across clouds, I mean, it's a real trend. I mean, anyway, that's how I see it.

>> Yeah. I feel the word data fabric many a times gets misused. Because when you think about the digitization movement that happened, started almost a decade ago, many companies tried to digitize or create digital twins of their systems into the data world, right? So, everything has an underlying data fabric that replicates what's happening transactionally, or otherwise in the real world. What data mesh does is creates structure that works complimentary to the data fabric, that then lends itself to data products, right? So to me, data products becomes a medium, which drives the connection between data mesh and data fabric into the real world for usage and consumption.

>> You should write for Gartner. (all laugh) That's the best explanation I've heard. That made sense!

>> That really did. That was excellent. So, when we think about any company these days has to be a data company, whether it's your grocery store, a gas station, a car dealer, what can companies do to start productizing their data, so that they can actually unlock new revenue streams, new routes to market? What are some steps and recommendations that you have? Justin, we'll start with you.

>> Sure. I would say the first thing is find data that is ultimately valuable to the consumers within your business, and create a product of it. And the way you do that at Starburst is allow you to essentially create a view of your data that can span multiple data sources. So again, we're decoupling where the data lives. That might be a table that lives in a traditional data warehouse, a table that lives in an operational system like Mongo, a table that lives in a data lake. And you can actually join those together, and represent it as a view, and now make it easily consumable. And so, the end user doesn't need to know, did that live in a data warehouse, an operational database, or a data lake? I'm just accessing that. And I think that's a great, easy way to start in your journey. Because I think if you absorb all the elements of data mesh at once, it can feel overwhelming. And I think that's a great way to start.

>> Irrespective of physical location.

>> Yes.

>> Right?

>> Precisely. Yep, multicloud, hybrid cloud, you name it.

>> And when you think about the broader landscape, right? For the traditionally, companies that only looked at internal data as a way of driving business decisions. More and more, as things evolve into industry, clouds, or ecosystem data, and companies start going beyond their four walls in terms of the data that they manage or the data that they use to make decisions, I think data products are going to play more and more an important part in that construct where you don't govern all the data that our entities within that ecosystem will govern parts of their data, but that data lives together in the form of data products that are governed somewhat centrally. I mean, kind of like a blockchain system, but not really.

>> Justin, for our folks here, as we kind of wrap the segment here, what's the bumper sticker for Starburst, and how you're helping organizations to really be able to build data products that add value to their organization?

>> I would say analytics anywhere. Our core ethos is, we want to give you the ability to access data wherever it lives, and understand your business holistically. And our query engine allows you to do that from a query perspective, and data products allows you to bring that up a level and make it consumable.

>> Make it consumable. Ashwin, last question for you, here we are, day one of re:Invent, loads of people behind us. Tomorrow all the great keynotes kick up. What are you hoping to take away from re:Invent '22?

>> Well, I'm hoping to understand how all of these different entities that are represented here connect with each other, right? And to me, Starburst is an important player in terms of how do you drive connectivity. And to me, as we help plans from a Deloitte perspective, drive that business value, connectivity across all of the technology players is extremely important part. So, integration across those technology players is what I'm trying to get from re:Invent here.

>> And so, you guys do, you're dot connectors. (Ashwin chuckles)

>> Exactly, excellent. Guys, thank you so much for joining David and me on the program tonight. We appreciate your insights, your time, and probably the best explanation of data fabric versus data mesh. (Justin chuckles) And data products that we've maybe ever had on the show! We appreciate your time, thank you.

>> Together: Thank you- >> Thanks, guys.

>> All right. For our guests and Dave Vellante, I'm Lisa Martin, you're watching theCUBE, the leader in enterprise and emerging tech coverage. (electronic music)