[Music] in the 2010s organizations became keenly aware that data would become the key ingredient in driving competitive advantage differentiation and growth but to this day putting data to work remains a difficult challenge for many if not most organizations now as the cloud matures it has become a game changer for data practitioners by making cheap storage and massive processing power readily accessible we've also seen better tooling in the form of data workflows streaming machine intelligence ai developer tools security observability automation new databases and the like these innovations they accelerate data proficiency but at the same time they had complexity for practitioners data lakes data hubs data warehouses data marts data fabrics data meshes data catalogs data oceans are forming they're evolving and exploding onto the scene so in an effort to bring perspective to the sea of optionality we've brought together the brightest minds in the data analyst community to discuss how data management is morphing and what practitioners should expect in 2022 and beyond hello everyone my name is dave vellante with the cube and i'd like to welcome you to a special cube presentation analyst predictions 2022 the future of data management we've gathered six of the best analysts in data and data management who are going to present and discuss their top predictions and trends for 2022 in the first half of this decade let me introduce our six power panelists sanjeev mohan is former gartner analyst and principal at sanjamo tony bear is principal at db insight carl olufsen is well-known research vice president with idc dave meninger is senior vice president and research director at ventana research brad shimon chief analyst at ai platforms analytics and data management at omnia and doug henschen vice president and principal analyst at constellation research gentlemen welcome to the program and thanks for coming on thecube today great to be here thank you all right here's the format we're going to use i as moderator are going to call on each analyst separately who then will deliver their prediction or mega trend and then in the interest of time management and pace two analysts will have the opportunity to comment if we have more time we'll elongate it but let's get started right away sanjeev mohan please kick it off you want to talk about governance go ahead sir thank you dave i i believe that data governance which we've been talking about for many years is now not only going to be mainstream it's going to be table stakes and all the things that you mentioned you know with data oceans data lakes lake houses data fabric meshes the common glue is metadata if we don't understand what data we have and we are governing it there is no way we can manage it so we saw informatica when public last year after a hiatus of six years i've i'm predicting that this year we see some more companies go public uh my bet is on colibra most likely and maybe alation we'll see go public this year we we i'm also predicting that the scope of data governance is going to expand beyond just data it's not just data and reports we are going to see more transformations like spark jaws python even airflow we're going to see more of streaming data so from kafka schema registry for example we will see ai models become part of this whole governance suite so the governance suite is going to be very comprehensive very detailed lineage impact analysis and then even expand into data quality we already seen that happen with some of the tools where they are buying these smaller companies and bringing in data quality monitoring and integrating it with metadata management data catalogs also data access governance so these so what we are going to see is that once the data governance platforms become the key entry point into these modern architectures i'm predicting that the usage the number of users of a data catalog is going to exceed that of a bi tool that will take time and we already seen that that trajectory right now if you look at bi tools i would say there are 100 users to a bi tool to one data catalog and i i see that evening out over a period of time and at some point data catalogs will really become you know the main way for us to access data data catalog will help us visualize data but if we want to do more in-depth analysis it'll be the jumping-off point into the bi tool the data science tool and and that is that is the journey i see for the data governance products excellent thank you some comments maybe maybe doug a lot a lot of things to weigh in on there maybe you could comment yeah sanjeev i think you're spot on a lot of the trends uh the one disagreement i think it's it's really still far from mainstream as you say we've been talking about this for years it's like god motherhood apple pie everyone agrees it's important but too few organizations are really practicing good governance because it's hard and because the incentives have been lacking i think one thing that deserves uh mention in this context is uh esg mandates and guidelines these are environmental social and governance regs and guidelines we've seen the environmental rags and guidelines imposed in industries particularly the carbon intensive industries we've seen the social mandates particularly diversity imposed on suppliers by companies that are leading on this topic we've seen governance guidelines now being imposed by banks and investors so these esgs are presenting new carrots and sticks and it's going to demand more solid data it's going to demand more detailed reporting and solid reporting tighter governance but we're still far from mainstream adoption we have a lot of uh you know best of breed niche players in the space i think the signs that it's going to be more mainstream are starting with things like azure purview google dataplex the big cloud platform uh players seem to be uh upping the ante and and addressing starting to address governance excellent thank you doug brad i wonder if you could chime in as well yeah i would love to be a believer in data catalogs um but uh to doug's point i think that it's going to take some more pressure for for that to happen i recall metadata being something every enterprise thought they were going to get under control when we were working on service oriented architecture back in the 90s and that didn't happen quite the way we we anticipated and and uh to sanjeev's point it's because it is really complex and really difficult to do my hope is that you know we won't sort of uh how do we put this fade out into this nebulous nebula of uh domain catalogs that are specific to individual use cases like purview for getting data quality right or like data governance and cyber security and instead we have some tooling that can actually be adaptive to gather metadata to create something i know is important to you sanjeev and that is this idea of observability if you can get enough metadata without moving your data around but understanding the entirety of a system that's running on this data you can do a lot to help with with the governance that doug is talking about so so i just want to add that you know data governance like many other initiatives did not succeed even ai went into an ai window but that's a different topic but a lot of these things did not succeed because to your point the incentives were not there i i remember when starbucks oxley had come into the scene if if a bank did not do service obviously they were very happy to a million dollar fine that was like you know pocket change for them instead of doing the right thing but i think the stakes are much higher now with gdpr uh the floodgates open now you know california you know has ccpa but even ccpa is being outdated with cpra which is much more gdpr like so we are very rapidly entering a space where every pretty much every major country in the world is coming up with its own uh compliance regulatory requirements data residence is becoming really important and and i i think we are going to reach a stage where uh it won't be optional anymore so whether we like it or not and i think the reason data catalogs were not successful in the past is because we did not have the right focus on adoption we were focused on features and these features were disconnected very hard for business to stop these are built by it people for it departments to to take a look at technical metadata not business metadata today the tables have turned cdo's are driving this uh initiative uh regulatory compliances are beating down hard so i think the time might be right yeah so guys we have to move on here and uh but there's some some real meat on the bone here sanjeev i like the fact that you late you called out calibra and alation so we can look back a year from now and say okay he made the call he stuck it and then the ratio of bi tools the data catalogs that's another sort of measurement that we can we can take even though some skepticism there that's something that we can watch and i wonder if someday if we'll have more metadata than data but i want to move to tony baer you want to talk about data mesh and speaking you know coming off of governance i mean wow you know the whole concept of data mesh is decentralized data and then governance becomes you know a nightmare there but take it away tony we'll put it this way um data mesh you know the the idea at least is proposed by thoughtworks um you know basically was unleashed a couple years ago and the press has been almost uniformly almost uncritical um a good reason for that is for all the problems that basically that sanjeev and doug and brad were just you know we're just speaking about which is that we have all this data out there and we don't know what to do about it um now that's not a new problem that was a problem we had enterprise data warehouses it was a problem when we had our hadoop data clusters it's even more of a problem now the data's out in the cloud where the data is not only your data like is not only s3 it's all over the place and it's also including streaming which i know we'll be talking about later so the data mesh was a response to that the idea of that we need to debate you know who are the folks that really know best about governance is the domain experts so it was basically data mesh was an architectural pattern and a process my prediction for this year is that data mesh is going to hit cold hard reality because if you if you do a google search um basically the the published work the articles and databases have been largely you know pretty uncritical um so far you know that you know basically learning is basically being a very revolutionary new idea i don't think it's that revolutionary because we've talked about ideas like this brad and i you and i met years ago when we were talking about so and decentralizing all of us was at the application level now we're talking about at the data level and now we have microservices so there's this thought of oh if we manage if we're apps in cloud native through microservices why don't we think of data in the same way um my sense this year is that you know this and this has been a very active search if you look at google search trends is that now companies are going to you know enterprises are going to look at this seriously and as they look at seriously it's going to attract its first real hard scrutiny it's going to attract its first backlash that's not necessarily a bad thing it means that it's being taken seriously um the reason why i think that that uh that it will you'll start to see basically the cold hard light of day shine on data mesh is that it's still a work in progress you know this idea is basically a couple years old and there's still some pretty major gaps um the biggest gap is in is in the area of federated governance now federated governance itself is not a new issue uh federated governance position we're trying to figure out like how can we basically strike the balance between getting let's say you know between basically consistent enterprise policy consistent enterprise governance but yet the groups that understand the data know how to basically you know that you know how do we basically sort of balance the two there's a huge there's a huge gap there in practice and knowledge um also to a lesser extent there's a technology gap which is basically in the self-service technologies that will help teams essentially govern data you know basically through the full life cycle from developed from selecting the data from you know building the other pipelines from determining your access control determining looking at quality looking at basically whether data is fresh or whether or not it's trending of course so my predictions is that it will really receive the first harsh scrutiny this year you are going to see some organization enterprises declare premature victory when they've uh when they build some federated query implementations you're going to see vendors start to data mesh wash their products anybody in the data management space they're going to say that whether it's basically a pipelining tool whether it's basically elt whether it's a catalog um or confederated query tool they're all going to be like you know basically promoting the fact of how they support this hopefully nobody is going to call themselves a data mesh tool because data mesh is not a technology we're going to see one other thing come out of this and this harks back to the metadata that sanji was talking about and the catalogs that he was talking about which is that there's going to be a new focus on every renewed focus on metadata and i think that's going to spur interest in data fabrics now data fabrics are pretty vaguely defined but if we just take the most elemental definition which is a common metadata back plane i think that if anybody is going to get serious about data mesh they need to look at a data fabric because we all at the end of the day need to speak you know need to read from the same sheet of music so thank you tony dave dave meninger i mean one of the things that people like about data mesh is it pretty crisply articulates some of the flaws in today's organizational approaches to data what are your thoughts on this well i think we have to start by defining data mesh right the the term is already getting corrupted right tony said it's going to see the cold hard uh light of day and there's a problem right now that there are a number of overlapping terms that are similar but not identical so we've got data virtualization data fabric excuse me for a second sorry about that data virtualization data fabric uh uh data federation right uh so i i think that it's not really clear what each vendor means by these terms i see data mesh and data fabric becoming quite popular i've i've interpreted data mesh as referring primarily to the governance aspects as originally you know intended and specified but that's not the way i see vendors using i see vendors using it much more to mean data fabric and data virtualization so i'm going to comment on the group of those things i think the group of those things is going to happen they're going to happen they're going to become more robust our research suggests that a quarter of organizations are already using virtualized access to their data lakes and another half so a total of three quarters will eventually be accessing their data lakes using some sort of virtualized access again whether you define it as mesh or fabric or virtualization isn't really the point here but this notion that there are different elements of data metadata and governance within an organization that all need to be managed collectively the interesting thing is when you look at the satisfaction rates of those organizations using virtualization versus those that are not it's almost double 68 of organizations i'm i'm sorry um 79 of organizations that were using virtualized access express satisfaction with their access to the data lake only 39 expressed satisfaction if they weren't using virtualized access so thank you uh dave uh sanjeev we just got about a couple minutes on this topic but i know you're speaking or maybe you've spoken already on a panel with jamal dagani who sort of invented the concept governance obviously is a big sticking point but what are your thoughts on this you are mute so my message to your mark and uh and to the community is uh as opposed to what dave said let's not define it we spent the whole year defining it there are four principles domain product data infrastructure and governance let's take it to the next level i get a lot of questions on what is the difference between data fabric and data mesh and i'm like i can compare the two because data mesh is a business concept data fabric is a data integration pattern how do you define how do you compare the two you have to bring data mesh level down so to tony's point i'm on a warp path in 2022 to take it down to what does a data product look like how do we handle shared data across domains and govern it and i think we are going to see more of that in 2022 is operationalization of data mesh i think we could have a whole hour on this topic couldn't we uh maybe we should do that uh but let's go to let's move to carl said carl your database guy you've been around that that block for a while now you want to talk about graph databases bring it on oh yeah okay thanks so i regard graph database as basically the next truly revolutionary database management technology i'm looking forward to for the graph database market which of course we haven't defined yet so obviously i have a little wiggle room in what i'm about to say but that this market will grow by about 600 percent over the next 10 years now 10 years is a long time but over the next five years we expect to see gradual growth as people start to learn how to use it problem isn't that it's used the problem is not that it's not useful is that people don't know how to use it so let me explain before i go any further what a graph database is because some of the folks on the call may not may not know what it is a graph database organizes data according to a mathematical structure called a graph a graph has elements called nodes and edges so a data element drops into a node the nodes are connected by edges the edges connect one node to another node combinations of edges create structures that you can analyze to determine how things are related in some cases the nodes and edges can have properties attached to them which add additional informative material that makes it richer that's called a property graph okay there are two principal use cases for graph databases there's there's semantic proper graphs which are used to break down human language text uh into the semantic structures then you can search it organize it and and and answer complicated questions a lot of ai is aimed at semantic graphs another kind is the property graph that i just mentioned which has a dazzling number of use cases i want to just point out is as i talk about this people are probably wondering well we have relational databases isn't that good enough okay so a relational database defines it uses um it supports what i call definitional relationships that means you define the relationships in a fixed structure the database drops into that structure there's a value foreign key value that relates one table to another and that value is fixed you don't change it if you change it the database becomes unstable it's not clear what you're looking at in a graph database the system is designed to handle change so that it can reflect the true state of the things that it's being used to track so um let me just give you some examples of use cases for this um they include uh entity resolution data lineage uh um social media analysis customer 360 fraud prevention there's cyber security there's strong supply chain is a big one actually there's explainable ai and this is going to become important too because a lot of people are adopting ai but they want a system after the fact to say how did the ai system come to that conclusion how did it make that recommendation right now we don't have really good ways of tracking that okay machine machine learning in general um social network i already mentioned that and then we've got oh gosh we've got data governance data compliance risk management we've got recommendation we've got personalization anti-money money laundering that's another big one identity and access management network and i.t operations is already becoming a key one where you actually have mapped out your operation your your you know whatever it is your data center and you you can track what's going on as things happen there root cause analysis fraud detection is a huge one a number of major credit card companies use graph databases for fraud detection risk analysis tracking and tracing churn analysis next best action what-if analysis impact analysis entity resolution and i would add one other thing or just a few other things to this list metadata management so sanjay here you go this is your engine okay because i was in metadata management for quite a while in my past life and one of the things i found was that none of the data management technologies that were available to us could efficiently handle metadata because of the kinds of structures that result from it but grass can okay grafts can do things like say this term in this context means this but in that context it means that okay things like that and in fact uh logistics management supply chain it also because it handles recursive relationships by recursive relationships i mean objects that own other objects that are of the same type you can do things like bill materials you know so like parts explosion you can do an hr analysis who reports to whom how many levels up the chain and that kind of thing you can do that with relational databases but yes it takes a lot of programming in fact you can do almost any of these things with relational databases but the problem is you have to program it it's not it's not supported in the database and whenever you have to program something that means you can't trace it you can't define it you can't publish it in terms of its functionality and it's really really hard to maintain over time so carl thank you i wonder if we could bring brad in i mean brad i'm sitting there wondering okay is this incremental to the market is it disruptive and replaceable what are your thoughts on this space it's already disrupted the market i mean like carl said go to any bank and ask them are you using graph databases to do to get fraud detection under control and they'll say absolutely that's the only way to solve this problem and it is frankly um and it's the only way to solve a lot of the problems that carl mentioned and that is i think it's it's achilles heel in some ways because you know it's like finding the best way to cross the seven bridges of konigsberg you know it's always going to kind of be tied to those use cases because it's really special and it's really unique and because it's special and it's unique uh it it still unfortunately kind of stands apart from the rest of the community that's building let's say ai outcomes as the great great example here the graph databases and ai as carl mentioned are like chocolate and peanut butter but technologically they don't know how to talk to one another they're completely different um and you know it's you can't just stand up sql and query them you've got to to learn um yeah what is that carlos specter or uh special uh uh yeah thank you uh to actually get to the data in there and if you're gonna scale that data that graph database especially a property graph if you're gonna do something really complex like try to understand uh you know all of the metadata in your organization you might just end up with you know a graph database winter like we had the ai winter simply because you run out of performance to make the thing happen so i i think it's already disrupted but we we need to like treat it like a first-class citizen in in the data analytics and ai community we need to bring it into the fold we need to equip it with the tools it needs to do that the magic it does and to do it not just for specialized use cases but for everything because i i'm with carl i i think it's absolutely revolutionary so i had also identified the principal achilles heel of the technology which is scaling now when these when these things get large and complex enough that they spill over what a single server can handle you start to have difficulties because the relationships span things that have to be resolved over a network and then you get network latency and that slows the system down so that's still a problem to be solved sanjeev any quick thoughts on this i mean i think metadata on the on the on the word cloud is going to be the the largest font uh but what are your thoughts here i want to like step away so people don't you know associate me with only meta data so i want to talk about something a little bit slightly different uh dbengines.com has done an amazing job i think almost everyone knows that they chronicle all the major databases that are in use today in january of 2022 there are 381 databases on its list of ranked list of databases the largest category is rdbms the second largest category is actually divided into two property graphs and rdf graphs these two together make up the second largest number of data databases so talking about accolades here this is a problem the problem is that there's so many graph databases to choose from they come in different shapes and forms uh to bright's point there's so many query languages in rdbms is sql end of the story here we've got sci-fi we've got gremlin we've got gql and then your proprietary languages so i think there's a lot of disparity in this space but excellent all excellent points sanji i must say and that is a problem the languages need to be sorted and standardized and it needs people need to have a road map as to what they can do with it because as you say you can do so many things and so many of those things are unrelated that you sort of say well what do we use this for i'm reminded of the saying i learned a bunch of years ago when somebody said that the digital computer is the only tool man has ever devised that has no particular purpose all right guys we gotta we gotta move on to dave uh meninger uh we've heard about streaming uh your prediction is in that realm so please take it away sure so i like to say that historical databases are to become a thing of the past but i don't mean that they're going to go away that's not my point i mean we need historical databases but streaming data is going to become the default way in which we operate with data so in the next say three to five years i would expect the data platforms and and we're using the term data platforms to represent the evolution of databases and data lakes that the data platforms will incorporate these streaming capabilities we're going to process data as it streams into an organization and then it's going to roll off into historical databases so historical databases don't go away but they become a thing of the past they store the data that occurred previously and as data is occurring we're going to be processing it we're going to be analyzing we're going to be acting on it i mean we we only ever ended up with historical databases because we were limited by the technology that was available to us data doesn't occur in batches but we processed it in batches because that was the best we could do and it wasn't bad and we've continued to improve and we've improved and we've improved but streaming data today is still the exception it's not the rule right there's there are projects within organizations that deal with streaming data but it's not the default way in which we deal with data yet and so that that's my prediction is that this is going to change we're going to have um streaming data be the default way in which we deal with data and and how you label it what you call it you know maybe these databases and data platforms just evolve to be able to handle it but we're going to deal with data in a different way and our research shows that already about half of the participants in our analytics and data benchmark research are using streaming data you know another third are planning to use streaming technologies so that gets us to about eight out of ten organizations need to use this technology that doesn't mean they have to use it throughout the whole organization but but it's pretty widespread in its use today and has continued to grow if you think about the consumerization of i.t we've all been conditioned to expect immediate access to information immediate responsiveness you know we want to know if an uh item is on the shelf at our local retail store and we can go in and pick it up right now you know that's the world we live in and that's spilling over into the enterprise i.t world where we have to provide those same types of capabilities um so that's my prediction historical database has become a thing of the past streaming data becomes the default way in which we we operate with data all right thank you david well so what what say you uh carl a guy who's followed historical databases for a long time well one thing actually every database is historical because as soon as you put data in it it's now history it's no longer it no longer reflects the present state of things but even if that history is only a millisecond old it's still history but um i would say i mean i know you're trying to be a little bit provocative in saying this dave because you know as well as i do that people still need to do their taxes they still need to do accounting they still need to run general ledger programs and things like that that all involves historical data that's not going to go away unless you want to go to jail so you're going to have to deal with that but as far as the leading edge functionality i'm totally with you on that and i'm just you know i'm just kind of wondering um if this chain if this requires a change in the way that we perceive applications in order to truly be manifested and rethinking the way m applications work um saying that uh an application should respond instantly as soon as the state of things changes what do you say about that i i think that's true i think we do have to think about things differently that's you know it's not the way we design systems in the past uh we're seeing more and more systems designed that way but again it's not the default and and agree 100 with you that we do need historical databases you know that that's clear and even some of those historical databases will be used in conjunction with the streaming data right so absolutely i mean you know let's take the data warehouse example where you're using the data warehouse as context and the streaming data as the present you're saying here's a sequence of things that's happening right now have we seen that sequence before and where what what does that pattern look like in past situations and can we learn from that so tony bear i wonder if you could comment i mean if you when you think about you know real-time inferencing at the edge for instance which is something that a lot of people talk about um a lot of what we're discussing here in this segment looks like it's got great potential what are your thoughts yeah well i mean i think you nailed it right you know you hit it right on the head there which is that i think a key what i'm seeing is that essentially and basically i'm going to split this one down the middle is i don't see that basically streaming is the default what i see is streaming and basically and transaction databases um and analytics data you know data warehouses data lakes whatever are converging and what allows us technically to converge is cloud native architecture where you can basically distribute things so you could have you can have a note here that's doing the real-time processing that's also doing it and this is what your leads in we're maybe doing some of that real-time predictive analytics to take a look at well look we're looking at this customer journey what's happening with you know you know with with what the customer is doing right now and this is correlated with what other customers are doing so what i so the thing is that in the cloud you can basically partition this and because of basically you know the speed of the infrastructure um that you can basically bring these together and or and so and kind of orchestrate them sort of loosely coupled manner the other part is that the use cases are demanding and this is part that goes back to what dave is saying is that you know when you look at customer 360 when you look at let's say smart you know smart utility grids when you look at any type of operational problem it has a real-time component and it has a historical component and having predictives and so like you know you know my sense here is that there that technically we can bring this together through the cloud and i think the use case is that is that we we can apply some some real-time sort of you know predictive analytics on these streams and feed this into the transactions so that when we make a decision in terms of what to do as a result of a transaction we have this real time you know input sanjeev did you have a comment yeah i was just going to say that to this point you know we have to think of streaming very different because in the historical databases we used to bring the data and store the data and then we used to run rules on top uh aggregations and all but in case of streaming the mindset changes because the rules normally the inference all of that is fixed but the data is constantly changing so it's a completely reverse way of thinking of uh and building applications on top of that so dave menninger there seemed to be some disagreement about the default or now what kind of time frame are you are you thinking about is this end of decade it becomes the default what would you pin i i think around you know between between five to ten years i think this becomes the reality um i think you know it'll be more and more common between now and then but it becomes the default and i also want sanjeev at some point maybe in one of our subsequent conversations we need to talk about governing streaming data because that's a whole other set of challenges we've also talked about it rather in a two dimensions historical and streaming and there's lots of low latency micro batch sub second that's not quite streaming but in many cases it's fast enough and we're seeing a lot of adoption of near real time not quite real time as uh good enough for most for many applications because nobody's really taking the hardware dimension of this information like how do we that'll just happen carl so near real time maybe before you lose the customer however you define that right okay um let's move on to brad brad you want to talk about automation ai uh the the the pipeline people feel like hey we can just automate everything what's your prediction yeah uh i'm i'm an ai fiction auto so apologies in advance for that but uh you know um i i think that um we've been seeing automation at play within ai for some time now and it's helped us do do a lot of things for especially for practitioners that are building ai outcomes in the enterprise uh it's it's helped them to fill skills gaps it's helped them to speed development and it's helped them to to actually make ai better uh because it you know in some ways provides some swim lanes and and for example with technologies like ottawa milk and can auto document and create that sort of transparency that that we talked about a little bit earlier um but i i think it's there's an interesting kind of conversion happening with this idea of automation um and and that is that uh we've had the automation that started happening for practitioners it's it's trying to move outside of the traditional bounds of things like i'm just trying to get my features i'm just trying to pick the right algorithm i'm just trying to build the right model uh and it's expanding across that full life cycle of building an ai outcome to start at the very beginning of data and to then continue on to the end which is this continuous delivery and continuous uh automation of of that outcome to make sure it's right and it hasn't drifted and stuff like that and because of that because it's become kind of powerful we're starting to to actually see this weird thing happen where the practitioners are starting to converge with the users and that is to say that okay if i'm in tableau right now i can stand up salesforce einstein discovery and it will automatically create a nice predictive algorithm for me um given the data that i that i pull in um but what's starting to happen and we're seeing this from the the the companies that create business software so salesforce oracle sap and others is that they're starting to actually use these same ideals and a lot of deep learning to to basically stand up these out of the box flip a switch and you've got an ai outcome at the ready for business users and um i i'm very much you know i think that that's that's the way that it's going to go and what it means is that ai is is slowly disappearing uh and i don't think that's a bad thing i think if anything what we're going to see in 2022 and maybe into 2023 is this sort of rush to to put this idea of disappearing ai into practice and have as many of these solutions in the enterprise as possible you can see like for example sap is going to roll out this quarter this thing called adaptive recommendation services which which basically is a cold start ai outcome that can work across a whole bunch of different vertical markets and use cases it's just a recommendation engine for whatever you need it to do in the line of business so basically you're you're an sap user you look up to turn on your software one day and you're a sales professional let's say and suddenly you have a recommendation for customer churn it's going that's great well i i don't know i i think that's terrifying in some ways i think it is the future that ai is going to disappear like that but i am absolutely terrified of it because um i i think that what it what it really does is it calls attention to a lot of the issues that we already see around ai um specific to this idea of what what we like to call it omdia responsible ai which is you know how do you build an ai outcome that is free of bias that is inclusive that is fair that is safe that is secure that it's audible etc etc etc etc that takes some a lot of work to do and so if you imagine a customer that that's just a sales force customer let's say and they're turning on einstein discovery within their sales software you need some guidance to make sure that when you flip that switch that the outcome you're going to get is correct and that's that's going to take some work and so i think we're going to see this let's roll this out and suddenly there's going to be a lot of a lot of problems a lot of pushback uh that we're going to see and some of that's going to come from gdpr and others that sam jeeve was mentioning earlier a lot of it's going to come from internal csr requirements within companies that are saying hey hey whoa hold up we can't do this all at once let's take the slow route let's make ai automated in a smart way and that's going to take time yeah so a couple predictions there that i heard i mean ai essentially you disappear it becomes invisible maybe if i can restate that and then if if i understand it correctly brad you're saying there's a backlash in the near term people can say oh slow down let's automate what we can those attributes that you talked about are non trivial to achieve is that why you're a bit of a skeptic yeah i think that we don't have any sort of standards that companies can look to and understand and we certainly within these companies especially those that haven't already stood up in internal data science team they don't have the knowledge to understand what that when they flip that switch for an automated ai outcome that it's it's gonna do what they think it's gonna do and so we need some sort of standard standard methodology and practice best practices that every company that's going to consume this invisible ai can make use of and one of the things that you know is sort of started that google kicked off a few years back that's picking up some momentum and the companies i just mentioned are starting to use it is this idea of model cards where at least you have some transparency about what these things are doing you know so like for the sap example we know for example that it's convolutional neural network with a long short-term memory model that it's using we know that it only works on roman english uh and therefore me as a consumer can say oh well i know that i need to do this internationally so i should not just turn this on today great thank you carl can you add anything any context here yeah we've talked about some of the things brad mentioned here at idc in the our future of intelligence group regarding in particular the moral and legal implications of having a fully automated you know ai uh driven system uh because we already know and we've seen that ai systems are biased by the data that they get right so if if they get data that pushes them in a certain direction i think there was a story last week about an hr system that was uh that was recommending promotions for white people over black people because in the past um you know white people were promoted and and more productive than black people but not it had no context as to why which is you know because they were being historically discriminated black people being historically discriminated against but the system doesn't know that so you know you have to be aware of that and i think that at the very least there should be controls when a decision has either a moral or a legal implication when when you want when you really need a human judgment it could lay out the options for you but a person actually needs to authorize that that action and i also think that we always will have to be vigilant regarding the kind of data we use to train our systems to make sure that it doesn't introduce unintended biases and to some extent they always will so we'll always be chasing after them that's that's absolutely carl yeah i think that what you have to bear in mind as a as a consumer of ai is that it is a reflection of us and we are a very flawed species uh and so if you look at all the really fantastic magical looking supermodels we see like gpt three and four that's coming out z they're xenophobic and hateful uh because the people the data that's built upon them and the algorithms and the people that build them are us so ai is a reflection of us we need to keep that in mind yeah we're the ai's by us because humans are biased all right great okay let's move on doug henson you know a lot of people that said that data lake that term's not not going to not going to live on but it appears to be have some legs here uh you want to talk about lake house bring it on yes i do my prediction is that lake house and this idea of a combined data warehouse and data lake platform is going to emerge as the dominant data management offering i say offering that doesn't mean it's going to be the dominant thing that organizations have out there but it's going to be the predominant vendor offering in 2022. now heading into 2021 we already had cloudera data bricks microsoft snowflake as proponents in 2021 sap oracle and several of these fabric virtualization mesh vendors join the bandwagon the promise is that you have one platform that manages your structured unstructured and semi-structured information and it addresses both the beyond analytics needs and the data science needs the real promise there is simplicity and lower cost but i think end users have to answer a few questions the first is does your organization really have a center of data gravity or is it is the data highly distributed multiple data warehouses multiple data lakes on-premises cloud if it if it's very distributed and you you know you have difficulty consolidating and that's not really a goal for you then maybe that single platform is unrealistic and not likely to add value to you um you know also the fabric and virtualization vendors the the mesh idea that's where if you have this highly distributed situation that might be a better path forward the second question if you are looking at one of these lake house offerings you are looking at consolidating simplifying bringing together to a single platform you have to make sure that it meets both the warehouse need and the data lake need so you have vendors like data bricks microsoft with azure synapse new really to the data warehouse space and they're having to prove that these data warehouse capabilities on their platforms can meet the scaling requirements can meet the user and query concurrency requirements meet those tight slas and then on the other hand you have the or the oracle sap snowflake the data warehouse uh folks coming into the data science world and they have to prove that they can manage the unstructured information and meet the needs of the data scientists i'm seeing a lot of the lake house offerings from the warehouse crowd managing that unstructured information in columns and rows and some of these vendors snowflake in particular is really relying on partners for the data science needs so you really got to look at a lake house offering and make sure that it meets both the warehouse and the data lake requirement well thank you doug well tony if those two worlds are going to come together as doug was saying the analytics and the data science world does it need to be some kind of semantic layer in between i don't know weigh in on this topic if you would oh didn't we talk about data fabrics before common metadata layer um actually i'm almost tempted to say let's declare victory and go home in that this is actually been going on for a while i actually agree with uh you know much what doug is saying there which is that i mean we i remembered as far back as i think it was like 2014 i was doing a a study you know it was still at ovum predecessor omnia um looking at all these specialized databases that were coming up and seeing that you know there's overlap with the edges but yet there was still going to be a reason at the time that you would have let's say a document database for json you'd have a relational database for tran you know for transactions and for data warehouse and you had you know and you had basically something at that time that that resembles to do for what we're considering a day of life fast fo and the thing is what i was saying at the time is that you're seeing basically blur you know sort of blending at the edges that i was saying like about five or six years ago um that's all and the the lake house is essentially you know the amount of the the current manifestation of that idea there is a dichotomy in terms of you know it's the old argument do we centralize this all you know you know in in in in in a single place or do we or do we virtualize and i think it's always going to be a yin and yang there's never going to be a single single silver silver bullet i do see um that they're also going to be questions and these are things that points that doug raised they're you know what your what do you need of of of your of you know for your performance there or for your you know pre-performance characteristics do you need for instance hiking currency you need the ability to do some very sophisticated joins or is your requirement more to be able to distribute and you know distribute our processing is you know as far as possible to get you know to essentially do a kind of brute force approach all these approaches are valid based on you know based on the used case um i just see that essentially that the lake house is the culmination of it's nothing it's just it's a relatively new term introduced by databricks a couple years ago this is the culmination of basically what's been a long time trend and what we see in the cloud is that as we start seeing data warehouses as a checkbox item say hey we can basically source data in cloud and cloud storage and s3 azure blob store you know whatever um as long as it's in certain formats like you know like you know parquet or csv or something like that you know i see that as becoming kind of you know a check box item so to that extent i think that the lake house depending on how you define it is already reality um and in some in some cases maybe new terminology but not a whole heck of a lot new under the sun yeah and dave menger i mean a lot of this thank you tony but a lot of this is going to come down to you know vendor marketing right some people try to co-opt the term we talked about data mesh washing what are your thoughts on this yeah so um i used the term data platform earlier and and part of the reason i use that term is that it's more vendor neutral uh we've we've tried to uh sort of stay out of the the vendor uh terminology patenting world right whether whether the term lake house is what sticks or not the concept is certainly going to stick and we have some data to back it up about a quarter of organizations that are using data lakes today already incorporate data warehouse functionality into it so they consider their data lake house and data warehouse one in the same about a quarter of organizations a little less but about a quarter of organizations feed the data lake from the data warehouse and about a quarter of organizations feed the data warehouse from the data lake so it's pretty obvious that three quarters of organizations need to bring this stuff together right the need is there the need is apparent the technology is going to continue to verge converge i i like to talk about you know you've got data lakes over here at one end and i'm not going to talk about why people thought data lakes were a bad idea because they thought you just throw stuff in a in a server and you ignore it right that's not what a data lake is so you've got data lake people over here and you've got database people over here data warehouse people over here database vendors are adding data lake capabilities and data lake vendors are adding data warehouse capabilities so it's obvious that they're going to meet in the middle i mean i think it's like tony says i think we should there declare victory and go home and so so i it's just a follow-up on that so are you saying these the specialized lake and the specialized warehouse do they go away i mean johnny tony data mesh practitioners would say or or advocates would say well they could all live as just a node on the on the mesh but based on what dave just said are we going to see those all morph together well number one as i was saying before there's always going to be this sort of you know kind of you know centrifugal force or this tug of war between do we centralize the data do we do it virtualize and the fact is i don't think that work there's ever going to be any single answer i think in terms of data mesh data mesh has nothing to do with how you physically implement the data you could have a data mesh on a basically uh on a data warehouse it's just that you know the difference being is that if we use the same you know physical data store but everybody's logically manual basically governing it differently you know um a data mission is basically it's not a technology it's a process it's a governance process um so essentially um you know you know i basically see that you know as as i was saying before that this is basically the culmination of a long time trend we're essentially seeing a lot of blurring but there are going to be cases where for instance if i need let's say like observe i need like high concurrency or something like that there are certain things that i'm not going to be able to get efficiently get out of a data lake um and you know we're basically i'm doing a system where i'm just doing really brute forcing very fast file scanning and that type of thing so i think there always will be some delineations but i would agree with dave and with doug that we are seeing basically a a confluence of requirements that we need to essentially have basically the element you know the ability of a data lake and a data laid out their warehouse we these need to come together so i think what we're likely to see is organizations look for a converged platform that can handle both sides for their center of data gravity the mesh and the fabric vendors the the fabric virtualization vendors they're all on board with the idea of this converged platform and they're saying hey we'll handle all the edge cases of the stuff that isn't in that center of data gradient that is off distributed in a cloud or at a remote location so you can have that single platform for the center of of your your data and then bring in virtualization mesh what have you for reaching out to the distributed data bingo as they basically said people are happy when they virtualize data i i think yes at this point but to this uh dave meningas point you know they have convert they are converging snowflake has introduced support for unstructured data so now we are literally splitting here now what uh databricks is saying is that aha but it's easy to go from data lake to data warehouse than it is from data warehouse to data lake so i think we're getting into semantics but we've already seen these two converge so is that so it takes something like aws who's got what 15 data stores are they're going to have 15 converged data stores that's going to be interesting to watch all right guys i'm going to go down the list and do like a one i'm going to one word each and you guys each of the analysts if you wouldn't just add a very brief sort of course correction for me so sanjeev i mean governance is going to be the maybe it's the dog that wags the tail now i mean it's coming to the fore all this ransomware stuff which really didn't talk much about security but but but what's the one word in your prediction that you would leave us with on governance it's uh it's going to be mainstream mainstream okay tony bear mesh washing is what i wrote down that's that's what we're going to see in uh in in 2022 a little reality check you you want to add to that reality check is i hope that no vendor you know jumps the shark and calls their offering a data mesh project yeah yeah let's hope that doesn't happen if they do we're going to call them out uh carl i mean graph databases thank you for sharing some some you know high growth metrics i know it's early days but magic is what i took away from that it's the magic database yeah i would actually i've said this to people too i i kind of look at it as a swiss army knife of data because you can pretty much do anything you want with it it doesn't mean you should i mean that's definitely the case that if you're you know managing things that are in a fixed schematic relationship probably a relational database is a better choice there are you know times when the document database is a better choice it can handle those things but maybe not it may not be the best choice for that use case but for a great many especially the new emerging use cases i listed it's the best choice thank you and dave meninger thank you by the way for bringing the data in i like how you supported all your comments with with some some data points but streaming data becomes the sort of default uh paradigm if you will what would you add yeah um i would say think fast right that's the world we live in you got to think fast fast love it uh and brad shimon uh i love it i mean on the one hand i was saying okay great i'm afraid i might get disrupted by one of these internet giants who are ai experts so i'm gonna be able to buy instead of build ai but then again you know i've got some real issues there's a potential backlash there so give us the there's your bumper sticker yeah i i would say um going with dave think fast and also think slow uh to to talk about the book that everyone talks about i would say really that this is all about trust trust in the idea of automation and of a transparent invisible ai across the enterprise but verify verify before you do anything and then doug henson i mean i i look i think the the trend is your friend here on this prediction with lake house is uh really becoming dominant i liked the way you set up that notion of you know the the the data warehouse folks coming at it from the analytics perspective but then you got the data science worlds coming together i still feel as though there's this piece in the middle that we're missing but your your final thoughts we'll give you the last well i think the idea of consolidation and simplification uh always prevails that's why the appeal of a single platform is going to be there um we've already seen that with uh you know hadoop platforms moving toward cloud moving toward object storage and object storage becoming really the common storage point for whether it's a lake or a warehouse uh and that second point uh i think esg mandates are uh are gonna come in alongside uh gdpr and things like that to uh up the ante for uh good governance yeah thank you for calling that out okay folks hey that's all the time that that we have here your your experience and depth of understanding on these key issues and in data and data management really on point and they were on display today i want to thank you for your your contributions really appreciate your time enjoyed it thank you now in addition to this video we're going to be making available transcripts of the discussion we're going to do clips of this as well we're going to put them out on social media i'll write this up and publish the discussion on wikibon.com and siliconangle.com no doubt several of the analysts on the panel will take the opportunity to publish written content social commentary or both i want to thank the power panelist and thanks for watching this special cube presentation this is dave vellante be well and we'll see you next time [Music] you

(bright music) >> In the 2010s, organizations became keenly aware that data would become the key ingredient to driving competitive advantage, differentiation, and growth. But to this day, putting data to work remains a difficult challenge for many, if not most organizations. Now, as the cloud matures, it has become a game changer for data practitioners by making cheap storage and massive processing power readily accessible. We've also seen better tooling in the form of data workflows, streaming, machine intelligence, AI, developer tools, security, observability, automation, new databases and the like. These innovations they accelerate data proficiency, but at the same time, they add complexity for practitioners. Data lakes, data hubs, data warehouses, data marts, data fabrics, data meshes, data catalogs, data oceans are forming, they're evolving and exploding onto the scene. So in an effort to bring perspective to the sea of optionality, we've brought together the brightest minds in the data analyst community to discuss how data management is morphing and what practitioners should expect in 2022 and beyond. Hello everyone, my name is Dave Velannte with theCUBE, and I'd like to welcome you to a special Cube presentation, analysts predictions 2022: the future of data management. We've gathered six of the best analysts in data and data management who are going to present and discuss their top predictions and trends for 2022 in the first half of this decade. Let me introduce our six power panelists. Sanjeev Mohan is former Gartner Analyst and Principal at SanjMo. Tony Baer, principal at dbInsight, Carl Olofson is well-known Research Vice President with IDC, Dave Menninger is Senior Vice President and Research Director at Ventana Research, Brad Shimmin, Chief Analyst, AI Platforms, Analytics and Data Management at Omdia and Doug Henschen, Vice President and Principal Analyst at Constellation Research. Gentlemen, welcome to the program and thanks for coming on theCUBE today. >> Great to be here. >> Thank you. >> All right, here's the format we're going to use. I as moderator, I'm going to call on each analyst separately who then will deliver their prediction or mega trend, and then in the interest of time management and pace, two analysts will have the opportunity to comment. If we have more time, we'll elongate it, but let's get started right away. Sanjeev Mohan, please kick it off. You want to talk about governance, go ahead sir. >> Thank you Dave. I believe that data governance which we've been talking about for many years is now not only going to be mainstream, it's going to be table stakes. And all the things that you mentioned, you know, the data, ocean data lake, lake houses, data fabric, meshes, the common glue is metadata. If we don't understand what data we have and we are governing it, there is no way we can manage it. So we saw Informatica went public last year after a hiatus of six. I'm predicting that this year we see some more companies go public. My bet is on Culebra, most likely and maybe Alation we'll see go public this year. I'm also predicting that the scope of data governance is going to expand beyond just data. It's not just data and reports. We are going to see more transformations like spark jawsxxxxx, Python even Air Flow. We're going to see more of a streaming data. So from Kafka Schema Registry, for example. We will see AI models become part of this whole governance suite. So the governance suite is going to be very comprehensive, very detailed lineage, impact analysis, and then even expand into data quality. We already seen that happen with some of the tools where they are buying these smaller companies and bringing in data quality monitoring and integrating it with metadata management, data catalogs, also data access governance. So what we are going to see is that once the data governance platforms become the key entry point into these modern architectures, I'm predicting that the usage, the number of users of a data catalog is going to exceed that of a BI tool. That will take time and we already seen that trajectory. Right now if you look at BI tools, I would say there a hundred users to BI tool to one data catalog. And I see that evening out over a period of time and at some point data catalogs will really become the main way for us to access data. Data catalog will help us visualize data, but if we want to do more in-depth analysis, it'll be the jumping off point into the BI tool, the data science tool and that is the journey I see for the data governance products. >> Excellent, thank you. Some comments. Maybe Doug, a lot of things to weigh in on there, maybe you can comment. >> Yeah, Sanjeev I think you're spot on, a lot of the trends the one disagreement, I think it's really still far from mainstream. As you say, we've been talking about this for years, it's like God, motherhood, apple pie, everyone agrees it's important, but too few organizations are really practicing good governance because it's hard and because the incentives have been lacking. I think one thing that deserves mention in this context is ESG mandates and guidelines, these are environmental, social and governance, regs and guidelines. We've seen the environmental regs and guidelines and posts in industries, particularly the carbon-intensive industries. We've seen the social mandates, particularly diversity imposed on suppliers by companies that are leading on this topic. We've seen governance guidelines now being imposed by banks on investors. So these ESGs are presenting new carrots and sticks, and it's going to demand more solid data. It's going to demand more detailed reporting and solid reporting, tighter governance. But we're still far from mainstream adoption. We have a lot of, you know, best of breed niche players in the space. I think the signs that it's going to be more mainstream are starting with things like Azure Purview, Google Dataplex, the big cloud platform players seem to be upping the ante and starting to address governance. >> Excellent, thank you Doug. Brad, I wonder if you could chime in as well. >> Yeah, I would love to be a believer in data catalogs. But to Doug's point, I think that it's going to take some more pressure for that to happen. I recall metadata being something every enterprise thought they were going to get under control when we were working on service oriented architecture back in the nineties and that didn't happen quite the way we anticipated. And so to Sanjeev's point it's because it is really complex and really difficult to do. My hope is that, you know, we won't sort of, how do I put this? Fade out into this nebula of domain catalogs that are specific to individual use cases like Purview for getting data quality right or like data governance and cybersecurity. And instead we have some tooling that can actually be adaptive to gather metadata to create something. And I know its important to you, Sanjeev and that is this idea of observability. If you can get enough metadata without moving your data around, but understanding the entirety of a system that's running on this data, you can do a lot. So to help with the governance that Doug is talking about. >> So I just want to add that, data governance, like any other initiatives did not succeed even AI went into an AI window, but that's a different topic. But a lot of these things did not succeed because to your point, the incentives were not there. I remember when Sarbanes Oxley had come into the scene, if a bank did not do Sarbanes Oxley, they were very happy to a million dollar fine. That was like, you know, pocket change for them instead of doing the right thing. But I think the stakes are much higher now. With GDPR, the flood gates opened. Now, you know, California, you know, has CCPA but even CCPA is being outdated with CPRA, which is much more GDPR like. So we are very rapidly entering a space where pretty much every major country in the world is coming up with its own compliance regulatory requirements, data residents is becoming really important. And I think we are going to reach a stage where it won't be optional anymore. So whether we like it or not, and I think the reason data catalogs were not successful in the past is because we did not have the right focus on adoption. We were focused on features and these features were disconnected, very hard for business to adopt. These are built by IT people for IT departments to take a look at technical metadata, not business metadata. Today the tables have turned. CDOs are driving this initiative, regulatory compliances are beating down hard, so I think the time might be right. >> Yeah so guys, we have to move on here. But there's some real meat on the bone here, Sanjeev. I like the fact that you called out Culebra and Alation, so we can look back a year from now and say, okay, he made the call, he stuck it. And then the ratio of BI tools to data catalogs that's another sort of measurement that we can take even though with some skepticism there, that's something that we can watch. And I wonder if someday, if we'll have more metadata than data. But I want to move to Tony Baer, you want to talk about data mesh and speaking, you know, coming off of governance. I mean, wow, you know the whole concept of data mesh is, decentralized data, and then governance becomes, you know, a nightmare there, but take it away, Tony. >> We'll put this way, data mesh, you know, the idea at least as proposed by ThoughtWorks. You know, basically it was at least a couple of years ago and the press has been almost uniformly almost uncritical. A good reason for that is for all the problems that basically Sanjeev and Doug and Brad we're just speaking about, which is that we have all this data out there and we don't know what to do about it. Now, that's not a new problem. That was a problem we had in enterprise data warehouses, it was a problem when we had over DoOP data clusters, it's even more of a problem now that data is out in the cloud where the data is not only your data lake, is not only us three, it's all over the place. And it's also including streaming, which I know we'll be talking about later. So the data mesh was a response to that, the idea of that we need to bait, you know, who are the folks that really know best about governance? It's the domain experts. So it was basically data mesh was an architectural pattern and a process. My prediction for this year is that data mesh is going to hit cold heart reality. Because if you do a Google search, basically the published work, the articles on data mesh have been largely, you know, pretty uncritical so far. Basically loading and is basically being a very revolutionary new idea. I don't think it's that revolutionary because we've talked about ideas like this. Brad now you and I met years ago when we were talking about so and decentralizing all of us, but it was at the application level. Now we're talking about it at the data level. And now we have microservices. So there's this thought of have we managed if we're deconstructing apps in cloud native to microservices, why don't we think of data in the same way? My sense this year is that, you know, this has been a very active search if you look at Google search trends, is that now companies, like enterprise are going to look at this seriously. And as they look at it seriously, it's going to attract its first real hard scrutiny, it's going to attract its first backlash. That's not necessarily a bad thing. It means that it's being taken seriously. The reason why I think that you'll start to see basically the cold hearted light of day shine on data mesh is that it's still a work in progress. You know, this idea is basically a couple of years old and there's still some pretty major gaps. The biggest gap is in the area of federated governance. Now federated governance itself is not a new issue. Federated governance decision, we started figuring out like, how can we basically strike the balance between getting let's say between basically consistent enterprise policy, consistent enterprise governance, but yet the groups that understand the data and know how to basically, you know, that, you know, how do we basically sort of balance the two? There's a huge gap there in practice and knowledge. Also to a lesser extent, there's a technology gap which is basically in the self-service technologies that will help teams essentially govern data. You know, basically through the full life cycle, from develop, from selecting the data from, you know, building the pipelines from, you know, determining your access control, looking at quality, looking at basically whether the data is fresh or whether it's trending off course. So my prediction is that it will receive the first harsh scrutiny this year. You are going to see some organization and enterprises declare premature victory when they build some federated query implementations. You going to see vendors start with data mesh wash their products anybody in the data management space that they are going to say that where this basically a pipelining tool, whether it's basically ELT, whether it's a catalog or federated query tool, they will all going to get like, you know, basically promoting the fact of how they support this. Hopefully nobody's going to call themselves a data mesh tool because data mesh is not a technology. We're going to see one other thing come out of this. And this harks back to the metadata that Sanjeev was talking about and of the catalog just as he was talking about. Which is that there's going to be a new focus, every renewed focus on metadata. And I think that's going to spur interest in data fabrics. Now data fabrics are pretty vaguely defined, but if we just take the most elemental definition, which is a common metadata back plane, I think that if anybody is going to get serious about data mesh, they need to look at the data fabric because we all at the end of the day, need to speak, you know, need to read from the same sheet of music. >> So thank you Tony. Dave Menninger, I mean, one of the things that people like about data mesh is it pretty crisply articulate some of the flaws in today's organizational approaches to data. What are your thoughts on this? >> Well, I think we have to start by defining data mesh, right? The term is already getting corrupted, right? Tony said it's going to see the cold hard light of day. And there's a problem right now that there are a number of overlapping terms that are similar but not identical. So we've got data virtualization, data fabric, excuse me for a second. (clears throat) Sorry about that. Data virtualization, data fabric, data federation, right? So I think that it's not really clear what each vendor means by these terms. I see data mesh and data fabric becoming quite popular. I've interpreted data mesh as referring primarily to the governance aspects as originally intended and specified. But that's not the way I see vendors using it. I see vendors using it much more to mean data fabric and data virtualization. So I'm going to comment on the group of those things. I think the group of those things is going to happen. They're going to happen, they're going to become more robust. Our research suggests that a quarter of organizations are already using virtualized access to their data lakes and another half, so a total of three quarters will eventually be accessing their data lakes using some sort of virtualized access. Again, whether you define it as mesh or fabric or virtualization isn't really the point here. But this notion that there are different elements of data, metadata and governance within an organization that all need to be managed collectively. The interesting thing is when you look at the satisfaction rates of those organizations using virtualization versus those that are not, it's almost double, 68% of organizations, I'm sorry, 79% of organizations that were using virtualized access express satisfaction with their access to the data lake. Only 39% express satisfaction if they weren't using virtualized access. >> Oh thank you Dave. Sanjeev we just got about a couple of minutes on this topic, but I know you're speaking or maybe you've always spoken already on a panel with (indistinct) who sort of invented the concept. Governance obviously is a big sticking point, but what are your thoughts on this? You're on mute. (panelist chuckling) >> So my message to (indistinct) and to the community is as opposed to what they said, let's not define it. We spent a whole year defining it, there are four principles, domain, product, data infrastructure, and governance. Let's take it to the next level. I get a lot of questions on what is the difference between data fabric and data mesh? And I'm like I can't compare the two because data mesh is a business concept, data fabric is a data integration pattern. How do you compare the two? You have to bring data mesh a level down. So to Tony's point, I'm on a warpath in 2022 to take it down to what does a data product look like? How do we handle shared data across domains and governance? And I think we are going to see more of that in 2022, or is "operationalization" of data mesh. >> I think we could have a whole hour on this topic, couldn't we? Maybe we should do that. But let's corner. Let's move to Carl. So Carl, you're a database guy, you've been around that block for a while now, you want to talk about graph databases, bring it on. >> Oh yeah. Okay thanks. So I regard graph database as basically the next truly revolutionary database management technology. I'm looking forward for the graph database market, which of course we haven't defined yet. So obviously I have a little wiggle room in what I'm about to say. But this market will grow by about 600% over the next 10 years. Now, 10 years is a long time. But over the next five years, we expect to see gradual growth as people start to learn how to use it. The problem is not that it's not useful, its that people don't know how to use it. So let me explain before I go any further what a graph database is because some of the folks on the call may not know what it is. A graph database organizes data according to a mathematical structure called a graph. The graph has elements called nodes and edges. So a data element drops into a node, the nodes are connected by edges, the edges connect one node to another node. Combinations of edges create structures that you can analyze to determine how things are related. In some cases, the nodes and edges can have properties attached to them which add additional informative material that makes it richer, that's called a property graph. There are two principle use cases for graph databases. There's semantic property graphs, which are use to break down human language texts into the semantic structures. Then you can search it, organize it and answer complicated questions. A lot of AI is aimed at semantic graphs. Another kind is the property graph that I just mentioned, which has a dazzling number of use cases. I want to just point out as I talk about this, people are probably wondering, well, we have relation databases, isn't that good enough? So a relational database defines... It supports what I call definitional relationships. That means you define the relationships in a fixed structure. The database drops into that structure, there's a value, foreign key value, that relates one table to another and that value is fixed. You don't change it. If you change it, the database becomes unstable, it's not clear what you're looking at. In a graph database, the system is designed to handle change so that it can reflect the true state of the things that it's being used to track. So let me just give you some examples of use cases for this. They include entity resolution, data lineage, social media analysis, Customer 360, fraud prevention. There's cybersecurity, there's strong supply chain is a big one actually. There is explainable AI and this is going to become important too because a lot of people are adopting AI. But they want a system after the fact to say, how do the AI system come to that conclusion? How did it make that recommendation? Right now we don't have really good ways of tracking that. Machine learning in general, social network, I already mentioned that. And then we've got, oh gosh, we've got data governance, data compliance, risk management. We've got recommendation, we've got personalization, anti money laundering, that's another big one, identity and access management, network and IT operations is already becoming a key one where you actually have mapped out your operation, you know, whatever it is, your data center and you can track what's going on as things happen there, root cause analysis, fraud detection is a huge one. A number of major credit card companies use graph databases for fraud detection, risk analysis, tracking and tracing turn analysis, next best action, what if analysis, impact analysis, entity resolution and I would add one other thing or just a few other things to this list, metadata management. So Sanjeev, here you go, this is your engine. Because I was in metadata management for quite a while in my past life. And one of the things I found was that none of the data management technologies that were available to us could efficiently handle metadata because of the kinds of structures that result from it, but graphs can, okay? Graphs can do things like say, this term in this context means this, but in that context, it means that, okay? Things like that. And in fact, logistics management, supply chain. And also because it handles recursive relationships, by recursive relationships I mean objects that own other objects that are of the same type. You can do things like build materials, you know, so like parts explosion. Or you can do an HR analysis, who reports to whom, how many levels up the chain and that kind of thing. You can do that with relational databases, but yet it takes a lot of programming. In fact, you can do almost any of these things with relational databases, but the problem is, you have to program it. It's not supported in the database. And whenever you have to program something, that means you can't trace it, you can't define it. You can't publish it in terms of its functionality and it's really, really hard to maintain over time. >> Carl, thank you. I wonder if we could bring Brad in, I mean. Brad, I'm sitting here wondering, okay, is this incremental to the market? Is it disruptive and replacement? What are your thoughts on this phase? >> It's already disrupted the market. I mean, like Carl said, go to any bank and ask them are you using graph databases to get fraud detection under control? And they'll say, absolutely, that's the only way to solve this problem. And it is frankly. And it's the only way to solve a lot of the problems that Carl mentioned. And that is, I think it's Achilles heel in some ways. Because, you know, it's like finding the best way to cross the seven bridges of Koenigsberg. You know, it's always going to kind of be tied to those use cases because it's really special and it's really unique and because it's special and it's unique, it's still unfortunately kind of stands apart from the rest of the community that's building, let's say AI outcomes, as a great example here. Graph databases and AI, as Carl mentioned, are like chocolate and peanut butter. But technologically, you think don't know how to talk to one another, they're completely different. And you know, you can't just stand up SQL and query them. You've got to learn, know what is the Carl? Specter special. Yeah, thank you to, to actually get to the data in there. And if you're going to scale that data, that graph database, especially a property graph, if you're going to do something really complex, like try to understand you know, all of the metadata in your organization, you might just end up with, you know, a graph database winter like we had the AI winter simply because you run out of performance to make the thing happen. So, I think it's already disrupted, but we need to like treat it like a first-class citizen in the data analytics and AI community. We need to bring it into the fold. We need to equip it with the tools it needs to do the magic it does and to do it not just for specialized use cases, but for everything. 'Cause I'm with Carl. I think it's absolutely revolutionary. >> Brad identified the principal, Achilles' heel of the technology which is scaling. When these things get large and complex enough that they spill over what a single server can handle, you start to have difficulties because the relationships span things that have to be resolved over a network and then you get network latency and that slows the system down. So that's still a problem to be solved. >> Sanjeev, any quick thoughts on this? I mean, I think metadata on the word cloud is going to be the largest font, but what are your thoughts here? >> I want to (indistinct) So people don't associate me with only metadata, so I want to talk about something slightly different. dbengines.com has done an amazing job. I think almost everyone knows that they chronicle all the major databases that are in use today. In January of 2022, there are 381 databases on a ranked list of databases. The largest category is RDBMS. The second largest category is actually divided into two property graphs and IDF graphs. These two together make up the second largest number databases. So talking about Achilles heel, this is a problem. The problem is that there's so many graph databases to choose from. They come in different shapes and forms. To Brad's point, there's so many query languages in RDBMS, in SQL. I know the story, but here We've got cipher, we've got gremlin, we've got GQL and then we're proprietary languages. So I think there's a lot of disparity in this space. >> Well, excellent. All excellent points, Sanjeev, if I must say. And that is a problem that the languages need to be sorted and standardized. People need to have a roadmap as to what they can do with it. Because as you say, you can do so many things. And so many of those things are unrelated that you sort of say, well, what do we use this for? And I'm reminded of the saying I learned a bunch of years ago. And somebody said that the digital computer is the only tool man has ever device that has no particular purpose. (panelists chuckle) >> All right guys, we got to move on to Dave Menninger. We've heard about streaming. Your prediction is in that realm, so please take it away. >> Sure. So I like to say that historical databases are going to become a thing of the past. By that I don't mean that they're going to go away, that's not my point. I mean, we need historical databases, but streaming data is going to become the default way in which we operate with data. So in the next say three to five years, I would expect that data platforms and we're using the term data platforms to represent the evolution of databases and data lakes, that the data platforms will incorporate these streaming capabilities. We're going to process data as it streams into an organization and then it's going to roll off into historical database. So historical databases don't go away, but they become a thing of the past. They store the data that occurred previously. And as data is occurring, we're going to be processing it, we're going to be analyzing it, we're going to be acting on it. I mean we only ever ended up with historical databases because we were limited by the technology that was available to us. Data doesn't occur in patches. But we processed it in patches because that was the best we could do. And it wasn't bad and we've continued to improve and we've improved and we've improved. But streaming data today is still the exception. It's not the rule, right? There are projects within organizations that deal with streaming data. But it's not the default way in which we deal with data yet. And so that's my prediction is that this is going to change, we're going to have streaming data be the default way in which we deal with data and how you label it and what you call it. You know, maybe these databases and data platforms just evolved to be able to handle it. But we're going to deal with data in a different way. And our research shows that already, about half of the participants in our analytics and data benchmark research, are using streaming data. You know, another third are planning to use streaming technologies. So that gets us to about eight out of 10 organizations need to use this technology. And that doesn't mean they have to use it throughout the whole organization, but it's pretty widespread in its use today and has continued to grow. If you think about the consumerization of IT, we've all been conditioned to expect immediate access to information, immediate responsiveness. You know, we want to know if an item is on the shelf at our local retail store and we can go in and pick it up right now. You know, that's the world we live in and that's spilling over into the enterprise IT world We have to provide those same types of capabilities. So that's my prediction, historical databases become a thing of the past, streaming data becomes the default way in which we operate with data. >> All right thank you David. Well, so what say you, Carl, the guy who has followed historical databases for a long time? >> Well, one thing actually, every database is historical because as soon as you put data in it, it's now history. They'll no longer reflect the present state of things. But even if that history is only a millisecond old, it's still history. But I would say, I mean, I know you're trying to be a little bit provocative in saying this Dave 'cause you know, as well as I do that people still need to do their taxes, they still need to do accounting, they still need to run general ledger programs and things like that. That all involves historical data. That's not going to go away unless you want to go to jail. So you're going to have to deal with that. But as far as the leading edge functionality, I'm totally with you on that. And I'm just, you know, I'm just kind of wondering if this requires a change in the way that we perceive applications in order to truly be manifested and rethinking the way applications work. Saying that an application should respond instantly, as soon as the state of things changes. What do you say about that? >> I think that's true. I think we do have to think about things differently. It's not the way we designed systems in the past. We're seeing more and more systems designed that way. But again, it's not the default. And I agree 100% with you that we do need historical databases you know, that's clear. And even some of those historical databases will be used in conjunction with the streaming data, right? >> Absolutely. I mean, you know, let's take the data warehouse example where you're using the data warehouse as its context and the streaming data as the present and you're saying, here's the sequence of things that's happening right now. Have we seen that sequence before? And where? What does that pattern look like in past situations? And can we learn from that? >> So Tony Baer, I wonder if you could comment? I mean, when you think about, you know, real time inferencing at the edge, for instance, which is something that a lot of people talk about, a lot of what we're discussing here in this segment, it looks like it's got a great potential. What are your thoughts? >> Yeah, I mean, I think you nailed it right. You know, you hit it right on the head there. Which is that, what I'm seeing is that essentially. Then based on I'm going to split this one down the middle is that I don't see that basically streaming is the default. What I see is streaming and basically and transaction databases and analytics data, you know, data warehouses, data lakes whatever are converging. And what allows us technically to converge is cloud native architecture, where you can basically distribute things. So you can have a node here that's doing the real-time processing, that's also doing... And this is where it leads in or maybe doing some of that real time predictive analytics to take a look at, well look, we're looking at this customer journey what's happening with what the customer is doing right now and this is correlated with what other customers are doing. So the thing is that in the cloud, you can basically partition this and because of basically the speed of the infrastructure then you can basically bring these together and kind of orchestrate them sort of a loosely coupled manner. The other parts that the use cases are demanding, and this is part of it goes back to what Dave is saying. Is that, you know, when you look at Customer 360, when you look at let's say Smart Utility products, when you look at any type of operational problem, it has a real time component and it has an historical component. And having predictive and so like, you know, my sense here is that technically we can bring this together through the cloud. And I think the use case is that we can apply some real time sort of predictive analytics on these streams and feed this into the transactions so that when we make a decision in terms of what to do as a result of a transaction, we have this real-time input. >> Sanjeev, did you have a comment? >> Yeah, I was just going to say that to Dave's point, you know, we have to think of streaming very different because in the historical databases, we used to bring the data and store the data and then we used to run rules on top, aggregations and all. But in case of streaming, the mindset changes because the rules are normally the inference, all of that is fixed, but the data is constantly changing. So it's a completely reversed way of thinking and building applications on top of that. >> So Dave Menninger, there seem to be some disagreement about the default. What kind of timeframe are you thinking about? Is this end of decade it becomes the default? What would you pin? >> I think around, you know, between five to 10 years, I think this becomes the reality. >> I think its... >> It'll be more and more common between now and then, but it becomes the default. And I also want Sanjeev at some point, maybe in one of our subsequent conversations, we need to talk about governing streaming data. 'Cause that's a whole nother set of challenges. >> We've also talked about it rather in two dimensions, historical and streaming, and there's lots of low latency, micro batch, sub-second, that's not quite streaming, but in many cases its fast enough and we're seeing a lot of adoption of near real time, not quite real-time as good enough for many applications. (indistinct cross talk from panelists) >> Because nobody's really taking the hardware dimension (mumbles). >> That'll just happened, Carl. (panelists laughing) >> So near real time. But maybe before you lose the customer, however we define that, right? Okay, let's move on to Brad. Brad, you want to talk about automation, AI, the pipeline people feel like, hey, we can just automate everything. What's your prediction? >> Yeah I'm an AI aficionados so apologies in advance for that. But, you know, I think that we've been seeing automation play within AI for some time now. And it's helped us do a lot of things especially for practitioners that are building AI outcomes in the enterprise. It's helped them to fill skills gaps, it's helped them to speed development and it's helped them to actually make AI better. 'Cause it, you know, in some ways provide some swim lanes and for example, with technologies like AutoML can auto document and create that sort of transparency that we talked about a little bit earlier. But I think there's an interesting kind of conversion happening with this idea of automation. And that is that we've had the automation that started happening for practitioners, it's trying to move out side of the traditional bounds of things like I'm just trying to get my features, I'm just trying to pick the right algorithm, I'm just trying to build the right model and it's expanding across that full life cycle, building an AI outcome, to start at the very beginning of data and to then continue on to the end, which is this continuous delivery and continuous automation of that outcome to make sure it's right and it hasn't drifted and stuff like that. And because of that, because it's become kind of powerful, we're starting to actually see this weird thing happen where the practitioners are starting to converge with the users. And that is to say that, okay, if I'm in Tableau right now, I can stand up Salesforce Einstein Discovery, and it will automatically create a nice predictive algorithm for me given the data that I pull in. But what's starting to happen and we're seeing this from the companies that create business software, so Salesforce, Oracle, SAP, and others is that they're starting to actually use these same ideals and a lot of deep learning (chuckles) to basically stand up these out of the box flip-a-switch, and you've got an AI outcome at the ready for business users. And I am very much, you know, I think that's the way that it's going to go and what it means is that AI is slowly disappearing. And I don't think that's a bad thing. I think if anything, what we're going to see in 2022 and maybe into 2023 is this sort of rush to put this idea of disappearing AI into practice and have as many of these solutions in the enterprise as possible. You can see, like for example, SAP is going to roll out this quarter, this thing called adaptive recommendation services, which basically is a cold start AI outcome that can work across a whole bunch of different vertical markets and use cases. It's just a recommendation engine for whatever you needed to do in the line of business. So basically, you're an SAP user, you look up to turn on your software one day, you're a sales professional let's say, and suddenly you have a recommendation for customer churn. Boom! It's going, that's great. Well, I don't know, I think that's terrifying. In some ways I think it is the future that AI is going to disappear like that, but I'm absolutely terrified of it because I think that what it really does is it calls attention to a lot of the issues that we already see around AI, specific to this idea of what we like to call at Omdia, responsible AI. Which is, you know, how do you build an AI outcome that is free of bias, that is inclusive, that is fair, that is safe, that is secure, that its audible, et cetera, et cetera, et cetera, et cetera. I'd take a lot of work to do. And so if you imagine a customer that's just a Salesforce customer let's say, and they're turning on Einstein Discovery within their sales software, you need some guidance to make sure that when you flip that switch, that the outcome you're going to get is correct. And that's going to take some work. And so, I think we're going to see this move, let's roll this out and suddenly there's going to be a lot of problems, a lot of pushback that we're going to see. And some of that's going to come from GDPR and others that Sanjeev was mentioning earlier. A lot of it is going to come from internal CSR requirements within companies that are saying, "Hey, hey, whoa, hold up, we can't do this all at once. "Let's take the slow route, "let's make AI automated in a smart way." And that's going to take time. >> Yeah, so a couple of predictions there that I heard. AI simply disappear, it becomes invisible. Maybe if I can restate that. And then if I understand it correctly, Brad you're saying there's a backlash in the near term. You'd be able to say, oh, slow down. Let's automate what we can. Those attributes that you talked about are non trivial to achieve, is that why you're a bit of a skeptic? >> Yeah. I think that we don't have any sort of standards that companies can look to and understand. And we certainly, within these companies, especially those that haven't already stood up an internal data science team, they don't have the knowledge to understand when they flip that switch for an automated AI outcome that it's going to do what they think it's going to do. And so we need some sort of standard methodology and practice, best practices that every company that's going to consume this invisible AI can make use of them. And one of the things that you know, is sort of started that Google kicked off a few years back that's picking up some momentum and the companies I just mentioned are starting to use it is this idea of model cards where at least you have some transparency about what these things are doing. You know, so like for the SAP example, we know, for example, if it's convolutional neural network with a long, short term memory model that it's using, we know that it only works on Roman English and therefore me as a consumer can say, "Oh, well I know that I need to do this internationally. "So I should not just turn this on today." >> Thank you. Carl could you add anything, any context here? >> Yeah, we've talked about some of the things Brad mentioned here at IDC and our future of intelligence group regarding in particular, the moral and legal implications of having a fully automated, you know, AI driven system. Because we already know, and we've seen that AI systems are biased by the data that they get, right? So if they get data that pushes them in a certain direction, I think there was a story last week about an HR system that was recommending promotions for White people over Black people, because in the past, you know, White people were promoted and more productive than Black people, but it had no context as to why which is, you know, because they were being historically discriminated, Black people were being historically discriminated against, but the system doesn't know that. So, you know, you have to be aware of that. And I think that at the very least, there should be controls when a decision has either a moral or legal implication. When you really need a human judgment, it could lay out the options for you. But a person actually needs to authorize that action. And I also think that we always will have to be vigilant regarding the kind of data we use to train our systems to make sure that it doesn't introduce unintended biases. In some extent, they always will. So we'll always be chasing after them. But that's (indistinct). >> Absolutely Carl, yeah. I think that what you have to bear in mind as a consumer of AI is that it is a reflection of us and we are a very flawed species. And so if you look at all of the really fantastic, magical looking supermodels we see like GPT-3 and four, that's coming out, they're xenophobic and hateful because the people that the data that's built upon them and the algorithms and the people that build them are us. So AI is a reflection of us. We need to keep that in mind. >> Yeah, where the AI is biased 'cause humans are biased. All right, great. All right let's move on. Doug you mentioned mentioned, you know, lot of people that said that data lake, that term is not going to live on but here's to be, have some lakes here. You want to talk about lake house, bring it on. >> Yes, I do. My prediction is that lake house and this idea of a combined data warehouse and data lake platform is going to emerge as the dominant data management offering. I say offering that doesn't mean it's going to be the dominant thing that organizations have out there, but it's going to be the pro dominant vendor offering in 2022. Now heading into 2021, we already had Cloudera, Databricks, Microsoft, Snowflake as proponents, in 2021, SAP, Oracle, and several of all of these fabric virtualization/mesh vendors joined the bandwagon. The promise is that you have one platform that manages your structured, unstructured and semi-structured information. And it addresses both the BI analytics needs and the data science needs. The real promise there is simplicity and lower cost. But I think end users have to answer a few questions. The first is, does your organization really have a center of data gravity or is the data highly distributed? Multiple data warehouses, multiple data lakes, on premises, cloud. If it's very distributed and you'd have difficulty consolidating and that's not really a goal for you, then maybe that single platform is unrealistic and not likely to add value to you. You know, also the fabric and virtualization vendors, the mesh idea, that's where if you have this highly distributed situation, that might be a better path forward. The second question, if you are looking at one of these lake house offerings, you are looking at consolidating, simplifying, bringing together to a single platform. You have to make sure that it meets both the warehouse need and the data lake need. So you have vendors like Databricks, Microsoft with Azure Synapse. New really to the data warehouse space and they're having to prove that these data warehouse capabilities on their platforms can meet the scaling requirements, can meet the user and query concurrency requirements. Meet those tight SLS. And then on the other hand, you have the Oracle, SAP, Snowflake, the data warehouse folks coming into the data science world, and they have to prove that they can manage the unstructured information and meet the needs of the data scientists. I'm seeing a lot of the lake house offerings from the warehouse crowd, managing that unstructured information in columns and rows. And some of these vendors, Snowflake a particular is really relying on partners for the data science needs. So you really got to look at a lake house offering and make sure that it meets both the warehouse and the data lake requirement. >> Thank you Doug. Well Tony, if those two worlds are going to come together, as Doug was saying, the analytics and the data science world, does it need to be some kind of semantic layer in between? I don't know. Where are you in on this topic? >> (chuckles) Oh, didn't we talk about data fabrics before? Common metadata layer (chuckles). Actually, I'm almost tempted to say let's declare victory and go home. And that this has actually been going on for a while. I actually agree with, you know, much of what Doug is saying there. Which is that, I mean I remember as far back as I think it was like 2014, I was doing a study. I was still at Ovum, (indistinct) Omdia, looking at all these specialized databases that were coming up and seeing that, you know, there's overlap at the edges. But yet, there was still going to be a reason at the time that you would have, let's say a document database for JSON, you'd have a relational database for transactions and for data warehouse and you had basically something at that time that resembles a dupe for what we consider your data life. Fast forward and the thing is what I was seeing at the time is that you were saying they sort of blending at the edges. That was saying like about five to six years ago. And the lake house is essentially on the current manifestation of that idea. There is a dichotomy in terms of, you know, it's the old argument, do we centralize this all you know in a single place or do we virtualize? And I think it's always going to be a union yeah and there's never going to be a single silver bullet. I do see that there are also going to be questions and these are points that Doug raised. That you know, what do you need for your performance there, or for your free performance characteristics? Do you need for instance high concurrency? You need the ability to do some very sophisticated joins, or is your requirement more to be able to distribute and distribute our processing is, you know, as far as possible to get, you know, to essentially do a kind of a brute force approach. All these approaches are valid based on the use case. I just see that essentially that the lake house is the culmination of it's nothing. It's a relatively new term introduced by Databricks a couple of years ago. This is the culmination of basically what's been a long time trend. And what we see in the cloud is that as we start seeing data warehouses as a check box items say, "Hey, we can basically source data in cloud storage, in S3, "Azure Blob Store, you know, whatever, "as long as it's in certain formats, "like, you know parquet or CSP or something like that." I see that as becoming kind of a checkbox item. So to that extent, I think that the lake house, depending on how you define is already reality. And in some cases, maybe new terminology, but not a whole heck of a lot new under the sun. >> Yeah. And Dave Menninger, I mean a lot of these, thank you Tony, but a lot of this is going to come down to, you know, vendor marketing, right? Some people just kind of co-op the term, we talked about you know, data mesh washing, what are your thoughts on this? (laughing) >> Yeah, so I used the term data platform earlier. And part of the reason I use that term is that it's more vendor neutral. We've tried to sort of stay out of the vendor terminology patenting world, right? Whether the term lake houses, what sticks or not, the concept is certainly going to stick. And we have some data to back it up. About a quarter of organizations that are using data lakes today, already incorporate data warehouse functionality into it. So they consider their data lake house and data warehouse one in the same, about a quarter of organizations, a little less, but about a quarter of organizations feed the data lake from the data warehouse and about a quarter of organizations feed the data warehouse from the data lake. So it's pretty obvious that three quarters of organizations need to bring this stuff together, right? The need is there, the need is apparent. The technology is going to continue to converge. I like to talk about it, you know, you've got data lakes over here at one end, and I'm not going to talk about why people thought data lakes were a bad idea because they thought you just throw stuff in a server and you ignore it, right? That's not what a data lake is. So you've got data lake people over here and you've got database people over here, data warehouse people over here, database vendors are adding data lake capabilities and data lake vendors are adding data warehouse capabilities. So it's obvious that they're going to meet in the middle. I mean, I think it's like Tony says, I think we should declare victory and go home. >> As hell. So just a follow-up on that, so are you saying the specialized lake and the specialized warehouse, do they go away? I mean, Tony data mesh practitioners would say or advocates would say, well, they could all live. It's just a node on the mesh. But based on what Dave just said, are we gona see those all morphed together? >> Well, number one, as I was saying before, there's always going to be this sort of, you know, centrifugal force or this tug of war between do we centralize the data, do we virtualize? And the fact is I don't think that there's ever going to be any single answer. I think in terms of data mesh, data mesh has nothing to do with how you're physically implement the data. You could have a data mesh basically on a data warehouse. It's just that, you know, the difference being is that if we use the same physical data store, but everybody's logically you know, basically governing it differently, you know? Data mesh in space, it's not a technology, it's processes, it's governance process. So essentially, you know, I basically see that, you know, as I was saying before that this is basically the culmination of a long time trend we're essentially seeing a lot of blurring, but there are going to be cases where, for instance, if I need, let's say like, Upserve, I need like high concurrency or something like that. There are certain things that I'm not going to be able to get efficiently get out of a data lake. And, you know, I'm doing a system where I'm just doing really brute forcing very fast file scanning and that type of thing. So I think there always will be some delineations, but I would agree with Dave and with Doug, that we are seeing basically a confluence of requirements that we need to essentially have basically either the element, you know, the ability of a data lake and the data warehouse, these need to come together, so I think. >> I think what we're likely to see is organizations look for a converge platform that can handle both sides for their center of data gravity, the mesh and the fabric virtualization vendors, they're all on board with the idea of this converged platform and they're saying, "Hey, we'll handle all the edge cases "of the stuff that isn't in that center of data gravity "but that is off distributed in a cloud "or at a remote location." So you can have that single platform for the center of your data and then bring in virtualization, mesh, what have you, for reaching out to the distributed data. >> As Dave basically said, people are happy when they virtualized data. >> I think we have at this point, but to Dave Menninger's point, they are converging, Snowflake has introduced support for unstructured data. So obviously literally splitting here. Now what Databricks is saying is that "aha, but it's easy to go from data lake to data warehouse "than it is from databases to data lake." So I think we're getting into semantics, but we're already seeing these two converge. >> So take somebody like AWS has got what? 15 data stores. Are they're going to 15 converge data stores? This is going to be interesting to watch. All right, guys, I'm going to go down and list do like a one, I'm going to one word each and you guys, each of the analyst, if you would just add a very brief sort of course correction for me. So Sanjeev, I mean, governance is going to to be... Maybe it's the dog that wags the tail now. I mean, it's coming to the fore, all this ransomware stuff, which you really didn't talk much about security, but what's the one word in your prediction that you would leave us with on governance? >> It's going to be mainstream. >> Mainstream. Okay. Tony Baer, mesh washing is what I wrote down. That's what we're going to see in 2022, a little reality check, you want to add to that? >> Reality check, 'cause I hope that no vendor jumps the shark and close they're offering a data niche product. >> Yeah, let's hope that doesn't happen. If they do, we're going to call them out. Carl, I mean, graph databases, thank you for sharing some high growth metrics. I know it's early days, but magic is what I took away from that, so magic database. >> Yeah, I would actually, I've said this to people too. I kind of look at it as a Swiss Army knife of data because you can pretty much do anything you want with it. That doesn't mean you should. I mean, there's definitely the case that if you're managing things that are in fixed schematic relationship, probably a relation database is a better choice. There are times when the document database is a better choice. It can handle those things, but maybe not. It may not be the best choice for that use case. But for a great many, especially with the new emerging use cases I listed, it's the best choice. >> Thank you. And Dave Menninger, thank you by the way, for bringing the data in, I like how you supported all your comments with some data points. But streaming data becomes the sort of default paradigm, if you will, what would you add? >> Yeah, I would say think fast, right? That's the world we live in, you got to think fast. >> Think fast, love it. And Brad Shimmin, love it. I mean, on the one hand I was saying, okay, great. I'm afraid I might get disrupted by one of these internet giants who are AI experts. I'm going to be able to buy instead of build AI. But then again, you know, I've got some real issues. There's a potential backlash there. So give us your bumper sticker. >> I'm would say, going with Dave, think fast and also think slow to talk about the book that everyone talks about. I would say really that this is all about trust, trust in the idea of automation and a transparent and visible AI across the enterprise. And verify, verify before you do anything. >> And then Doug Henschen, I mean, I think the trend is your friend here on this prediction with lake house is really becoming dominant. I liked the way you set up that notion of, you know, the data warehouse folks coming at it from the analytics perspective and then you get the data science worlds coming together. I still feel as though there's this piece in the middle that we're missing, but your, your final thoughts will give you the (indistinct). >> I think the idea of consolidation and simplification always prevails. That's why the appeal of a single platform is going to be there. We've already seen that with, you know, DoOP platforms and moving toward cloud, moving toward object storage and object storage, becoming really the common storage point for whether it's a lake or a warehouse. And that second point, I think ESG mandates are going to come in alongside GDPR and things like that to up the ante for good governance. >> Yeah, thank you for calling that out. Okay folks, hey that's all the time that we have here, your experience and depth of understanding on these key issues on data and data management really on point and they were on display today. I want to thank you for your contributions. Really appreciate your time. >> Enjoyed it. >> Thank you. >> Thanks for having me. >> In addition to this video, we're going to be making available transcripts of the discussion. We're going to do clips of this as well we're going to put them out on social media. I'll write this up and publish the discussion on wikibon.com and siliconangle.com. No doubt, several of the analysts on the panel will take the opportunity to publish written content, social commentary or both. I want to thank the power panelists and thanks for watching this special CUBE presentation. This is Dave Vellante, be well and we'll see you next time. (bright music)

>> Announcer: Live, from Las Vegas, it's the Cube, covering HPE Discover 2017. Brought to you by Hewlett Packard Enterprise. >> Hey, welcome back everyone. We are here live in Las Vegas for day two of three days of exclusive coverage from the Cube here at HPE Discover 2017. Our two next guests is Bill Mannel, VP and General Manager of HPC and AI for HPE. Bill, great to see you. And Dr. Nick Nystrom, senior of research at Pittsburgh's Supercomputer Center. Welcome to The Cube, thanks for coming on, appreciate it. >> My pleasure >> Thanks for having us. >> As we wrap up day two, first of all before we get started, love the AI, love the high performance computing. We're seeing great applications for compute. Everyone now sees that a lot of compute actually is good. That's awesome. What is the Pittsburgh Supercomputer Center? Give a quick update and describe what that is. >> Sure. The quick update is we're operating a system called Bridges. Bridges is operating for the National Science Foundation. It democratizes HPC. It brings people who have never used high performance computing before to be able to use HPC seamlessly, almost as a cloud. It unifies HPC big data and artificial intelligence. >> So who are some of the users that are getting access that they didn't have before? Could you just kind of talk about some of the use cases of the organizations or people that you guys are opening this up to? >> Sure. I think one of the newest communities that's very significant is deep learning. So we have collaborations between the University of Pittsburgh life sciences and the medical center with Carnegie Mellon, the machine learning researchers. We're looking to apply AI machine learning to problems in breast and lung cancer. >> Yeah, we're seeing the data. Talk about some of the innovations that HPE's bringing with you guys in the partnership, because we're seeing, people are seeing the results of using big data and deep learning and breakthroughs that weren't possible before. So not only do you have the democratization cool element happening, you have a tsunami of awesome open source code coming in from big places. You see Google donating a bunch of machine learning libraries. Everyone's donating code. It's like open bar and open source, as I say, and the young kids that are new are the innovators as well, so not just us systems guys, but a lot of young developers are coming in. What's the innovation? Why is this happening? What's the ah-ha moment? Is it just cloud, is it a combination of things, talk about it. >> It's a combination of all the big data coming in, and then new techniques that allow us to analyze and get value from it and from that standpoint. So the traditional HPC world, typically we built equations which then generated data. Now we're actually kind of doing the reverse, which is we take the data and then build equations to understand the data. So it's a different paradigm. And so there's more and more energy understanding those two different techniques of kind of getting two of the same answers, but in a different way. >> So Bill, you and I talked in London last year. >> Yes. With Dr. Gho. And we talked a lot about SGI and what that acquisition meant to you guys. So I wonder if you could give us a quick update on the business? I mean it's doing very well, Meg talked about it on the conference call this last quarter. Really high point and growing. What's driving the growth, and give us an update on the business. >> Sure. And I think the thing that's driving the growth is all this data and the fact that customers want to get value from it. So we're seeing a lot of growth in industries like financial services, like in manufacturing, where folks are moving to digitization, which means that in the past they might have done a lot of their work through experimentation. Now they're moving it to a digital format, and they're simulating everything. So that's driven a lot more HPC over time. As far as the SGI, integration is concern. We've integrated about halfway, so we're at about the halfway point. And now we've got the engineering teams together and we're driving a road map and a new set of products that are coming out. Our Gen 10-based products are on target, and they're going to be releasing here over the next few months. >> So Nick, from your standpoint, when you look at, there's been an ebb and flow in the supercomputer landscape for decades. All the way back to the 70s and the 80s. So from a customer perspective, what do you see now? Obviously China's much more prominent in the game. There's sort of an arms race, if you will, in computing power. From a customer's perspective, what are you seeing, what are you looking for in a supplier? >> Well, so I agree with you, there is this arms race for exaflops. Where we are really focused right now is enabling data-intensive applications, looking at big data service, HPC is a service, really making things available to users to be able to draw on the large data sets you mentioned, to be able to put the capability class computing, which will go to exascale, together with AI, and data and Linux under one platform, under one integrated fabric. That's what we did with HPE for Bridges. And looking to build on that in the future, to be able to do the exascale applications that you're referring to, but also to couple on data, and to be able to use AI with classic simulation to make those simulations better. >> So it's always good to have a true practitioner on The Cube. But when you talk about AI and machine learning and deep learning, John and I sometimes joke, is it same wine, new bottle, or is there really some fundamental shift going on that just sort of happened to emerge in the last six to nine months? >> I think there is a fundamental shift. And the shift is due to what Bill mentioned. It's the availability of data. So we have that. We have more and more communities who are building on that. You mentioned the open source frameworks. So yes, they're building on the TensorFlows, on the Cafes, and we have people who have not been programmers. They're using these frameworks though, and using that to drive insights from data they did not have access to. >> These are flipped upside down, I mean this is your point, I mean, Bill pointed it out, it's like the models are upside down. This is the new world. I mean, it's crazy, I don't believe it. >> So if that's the case, and I believe it, it feels like we're entering this new wave of innovation which for decades we talked about how we march to the cadence of Moore's Law. That's been the innovation. You think back, you know, your five megabyte disk drive, then it went to 10, then 20, 30, now it's four terabytes. Okay, wow. Compared to what we're about to see, I mean it pales in comparison. So help us envision what the world is going to look like in 10 or 20 years. And I know it's hard to do that, but can you help us get our minds around the potential that this industry is going to tap? >> So I think, first of all, I think the potential of AI is very hard to predict. We see that. What we demonstrated in Pittsburgh with the victory of Libratus, the poker-playing bot, over the world's best humans, is the ability of an AI to beat humans in a situation where they have incomplete information, where you have an antagonist, an adversary who is bluffing, who is reacting to you, and who you have to deal with. And I think that's a real breakthrough. We're going to see that move into other aspects of life. It will be buried in apps. It will be transparent to a lot of us, but those sorts of AI's are going to influence a lot. That's going to take a lot of IT on the back end for the infrastructure, because these will continue to be compute-hungry. >> So I always use the example of Kasperov and he got beaten by the machine, and then he started a competition to team up with a supercomputer and beat the machine. Yeah, humans and machines beat machines. Do you expect that's going to continue? Maybe both your opinions. I mean, we're just sort of spitballing here. But will that augmentation continue for an indefinite period of time, or are we going to see the day that it doesn't happen? >> I think over time you'll continue to see progress, and you'll continue to see more and more regular type of symmetric type workloads being done by machines, and that allows us to do the really complicated things that the human brain is able to better process than perhaps a machine brain, if you will. So I think it's exciting from the standpoint of being able to take some of those other roles and so forth, and be able to get those done in perhaps a more efficient manner than we're able to do. >> Bill, talk about, I want to get your reaction to the concept of data. As data evolves, you brought up the model, I like the way you're going with that, because things are being flipped around. In the old days, I want to monetize my data. I have data sets, people are looking at their data. I'm going to make money from my data. So people would talk about how we monetizing the data. >> Dave: Old days, like two years ago. >> Well and people actually try to solve and monetize their data, and this could be use case for one piece of it. Other people are saying no, I'm going to open, make people own their own data, make it shareable, make it more of an enabling opportunity, or creating opportunities to monetize differently. In a different shift. That really comes down to the insights question. What's your, what trends do you guys see emerging where data is much more of a fabric, it's less of a discreet, monetizable asset, but more of an enabling asset. What's your vision on the role of data? As developers start weaving in some of these insights. You mentioned the AI, I think that's right on. What's your reaction to the role of data, the value of the data? >> Well, I think one thing that we're seeing in some of our, especially our big industrial customers is the fact that they really want to be able to share that data together and collect it in one place, and then have that regularly updated. So if you look at a big aircraft manufacturer, for example, they actually are putting sensors all over their aircraft, and in realtime, bringing data down and putting it into a place where now as they're doing new designs, they can access that data, and use that data as a way of making design trade-offs and design decision. So a lot of customers that I talk to in the industrial area are really trying to capitalize on all the data possible to allow them to bring new insights in, to predict things like future failures, to figure out how they need to maintain whatever they have in the field and those sorts of things at all. So it's just kind of keeping it within the enterprise itself. I mean, that's a challenge, a really big challenge, just to get data collected in one place and be able to efficiently use it just within an enterprise. We're not even talking about sort of pan-enterprise, but just within the enterprise. That is a significant change that we're seeing. Actually an effort to do that and see the value in that. >> And the high performance computing really highlights some of these nuggets that are coming out. If you just throw compute at something, if you set it up and wrangle it, you're going to get these insights. I mean, new opportunities. >> Bill: Yeah, absolutely. >> What's your vision, Nick? How do you see the data, how do you talk to your peers and people who are generally curious on how to approach it? How to architect data modeling and how to think about it? >> I think one of the clearest examples on managing that sort of data comes from the life sciences. So we're working with researchers at University of Pittsburgh Medical Center, and the Institute for Precision Medicine at Pitt Cancer Center. And there it's bringing together the large data as Bill alluded to. But there it's very disparate data. It is genomic data. It is individual tumor data from individual patients across their lifetime. It is imaging data. It's the electronic health records. And trying to be able to do this sort of AI on that to be able to deliver true precision medicine, to be able to say that for a given tumor type, we can look into that and give you the right therapy, or even more interestingly, how can we prevent some of these issues proactively? >> Dr. Nystrom, it's expensive doing what you do. Is there a commercial opportunity at the end of the rainbow here for you or is that taboo, I mean, is that a good thing? >> No, thank you, it's both. So as a national supercomputing center, our resources are absolutely free for open research. That's a good use of our taxpayer dollars. They've funded these, we've worked with HP, we've designed the system that's great for everybody. We also can make this available to industry at an extremely low rate because it is a federal resource. We do not make a profit on that. But looking forward, we are working with local industry to let them test things, to try out ideas, especially in AI. A lot of people want to do AI, they don't know what to do. And so we can help them. We can help them architect solutions, put things on hardware, and when they determine what works, then they can scale that up, either locally on prem, or with us. >> This is a great digital resource. You talk about federally funded. I mean, you can look at Yosemite, it's a state park, you know, Yellowstone, these are natural resources, but now when you start thinking about the goodness that's being funded. You want to talk about democratization, medicine is just the tip of the iceberg. This is an interesting model as we move forward. We see what's going on in government, and see how things are instrumented, some things not, delivery of drugs and medical care, all these things are coalescing. How do you see this digital age extending? Because if this continues, we should be doing more of these, right? >> We should be. We need to be. >> It makes sense. So is there, I mean I just not up to speed on what's going on with federally funded-- >> Yeah, I think one thing that Pittsburgh has done with the Bridges machine, is really try to bring in data and compute and all the different types of disciplines in there, and provide a place where a lot of people can learn, they can build applications and things like that. That's really unusual in HPC. A lot of times HPC is around big iron. People want to have the biggest iron basically on the top 500 list. This is where the focus hasn't been on that. This is where the focus has been on really creating value through the data, and getting people to utilize it, and then build more applications. >> You know, I'll make an observation. When we first started doing The Cube, we observed that, we talked about big data, and we said that the practitioners of big data, are where the guys are going to make all the money. And so far that's proven true. You look at the public big data companies, none of them are making any money. And maybe this was sort of true with ERP, but not like it is with big data. It feels like AI is going to be similar, that the consumers of AI, those people that can find insights from that data are really where the big money is going to be made here. I don't know, it just feels like-- >> You mean a long tail of value creation? >> Yeah, in other words, you used to see in the computing industry, it was Microsoft and Intel became, you know, trillion dollar value companies, and maybe there's a couple of others. But it really seems to be the folks that are absorbing those technologies, applying them, solving problems, whether it's health care, or logistics, transportation, etc., looks to where the huge economic opportunities may be. I don't know if you guys have thought about that. >> Well I think that's happened a little bit in big data. So if you look at what the financial services market has done, they've probably benefited far more than the companies that make the solutions, because now they understand what their consumers want, they can better predict their life insurance, how they should-- >> Dave: You could make that argument for Facebook, for sure. >> Absolutely, from that perspective. So I expect it to get to your point around AI as well, so the folks that really use it, use it well, will probably be the ones that benefit it. >> Because the tooling is very important. You've got to make the application. That's the end state in all this That's the rubber meets the road. >> Bill: Exactly. >> Nick: Absolutely. >> All right, so final question. What're you guys showing here at Discover? What's the big HPC? What's the story for you guys? >> So we're actually showing our Gen 10 product. So this is with the latest microprocessors in all of our Apollo lines. So these are specifically optimized platforms for HPC and now also artificial intelligence. We have a platform called the Apollo 6500, which is used by a lot of companies to do AI work, so it's a very dense GPU platform, and does a lot of processing and things in terms of video, audio, these types of things that are used a lot in some of the workflows around AI. >> Nick, anything spectacular for you here that you're interested in? >> So we did show here. We had video in Meg's opening session. And that was showing the poker result, and I think that was really significant, because it was actually a great amount of computing. It was 19 million core hours. So was an HPC AI application, and I think that was a really interesting success. >> The unperfect information really, we picked up this earlier in our last segment with your colleagues. It really amplifies the unstructured data world, right? People trying to solve the streaming problem. With all this velocity, you can't get everything, so you need to use machines, too. Otherwise you have a haystack of needles. Instead of trying to find the needles in the haystack, as they was saying. Okay, final question, just curious on this natural, not natural, federal resource. Natural resource, feels like it. Is there like a line to get in? Like I go to the park, like this camp waiting list, I got to get in there early. How do you guys handle the flow for access to the supercomputer center? Is it, my uncle works there, I know a friend of a friend? Is it a reservation system? I mean, who gets access to this awesomeness? >> So there's a peer reviewed system, it's fair. People apply for large allocations four times a year. This goes to a national committee. They met this past Sunday and Monday for the most recent. They evaluate the proposals based on merit, and they make awards accordingly. We make 90% of the system available through that means. We have 10% discretionary that we can make available to the corporate sector and to others who are doing proprietary research in data-intensive computing. >> Is there a duration, when you go through the application process, minimums and kind of like commitments that they get involved, for the folks who might be interested in hitting you up? >> For academic research, the normal award is one year. These are renewable, people can extend these and they do. What we see now of course is for large data resources. People keep those going. The AI knowledge base is 2.6 petabytes. That's a lot. For industrial engagements, those could be any length. >> John: Any startup action coming in, or more bigger, more-- >> Absolutely. A coworker of mine has been very active in life sciences startups in Pittsburgh, and engaging many of these. We have meetings every week with them now, it seems. And with other sectors, because that is such a great opportunity. >> Well congratulations. It's fantastic work, and we're happy to promote it and get the word out. Good to see HP involved as well. Thanks for sharing and congratulations. >> Absolutely. >> Good to see your work, guys. Okay, great way to end the day here. Democratizing supercomputing, bringing high performance computing. That's what the cloud's all about. That's what great software's out there with AI. I'm John Furrier, Dave Vellante bringing you all the data here from HPE Discover 2017. Stay tuned for more live action after this short break.