State of Play with Dataverse with James Oleinik

[mark]: 4:19
in this episode we will be discussing the data today's guest is from washington in the US he works as a director of product management at microsoft you can find links to his linked in any kind of blogs content that he held out there in the show notes welcome to the show james

[james_oleinik]: 4:38
thank you for having me

[mark]: 4:40
good to have you on good to be discussing data verse i think the ryan jones was the last person many moons ago maybe three years ago we had a discussion on data verse

[james_oleinik]: 4:52
that's awesome

[mark]: 4:52
and of course it's it's a different beast

[james_oleinik]: 4:56
it's grown and the i think the first question i answer after you know what is data verse is is what happened to ryan jones and i'm the new ryan jones oh

[mark]: 5:08
is that is that right it was so funny because

[james_oleinik]: 5:10
that's right

[mark]: 5:10
it did come it did come up about i don't know maybe two weeks ago on a v p side channel chat was we'sroan jones gone and i was like well last i saw he was in customer data platform

[james_oleinik]: 5:22
that's right

[mark]: 5:22
sorry customer insight is where he was and that's where understanding moved to and that was from n v p summit conversations with him then

[james_oleinik]: 5:30
yeah and ironically my previous role before joining data verse was in our insights area supporting some of the teams like customer insights and some of the industry club investments there so he and i past like ships in the night

[mark]: 5:47
well

[james_oleinik]: 5:47
me data verse and he to the inside world so go figure right

[mark]: 5:52
so so some some things just popped to mind when you know mentioning his name now and i remember when he first joined the team man he told me some stuff probably more of record stuff around what he landed in as an you know the product had evolved in a very different way pre james phillips as an i'm talking about the dynamics swede products and there was there was kind of like five i think there were five core messages and i know one was around stabilizing the platform high availability and um and there were some other things including the word can of worms that that needed work on and so was three or four years ago right and we've seen i suppose a lot of a lot of people a lot of companies and our come in and join the platform and be part of it and i'm always interested understanding or getting people that are working with it to how do you articulate the data verse how do you talk about it in a non to start with let's go non technical term so from the point of view people go oh is it another data base how do you how do you tell the story of what is the date of us and why you know dynamics power apse et cetera the whole gammit why arn't they just but on um u azure sequel and you know or any other high available storage direct on data lake or something like that

[james_oleinik]: 7:28
that's right well i think that that's that's a great question so maybe the explain like i'm five answer is that i think of more i think of data verse as yes it is much more than just another sequel data base the way that we think about data verse is it is effectively azure in a box right and it's aspiring to bring low code data platform to bear in an approachable way such a it's easy as building an excell table building kind of related tables together and building kind of your data model just like you would do and say building a list and share point online right and the key the key point is that we're doing so um by actually leveraging so much more than just a sequel and sure that's where our roots are but we have for example the ability to have essentially a fully scalable data platform where we support relational and non relational data images files relevant search or yes more analytical data on the day like that's all behind the scenes and so we're not just as our sequel we're actually think of all of the azure data services brought together first and foremost and that is kind of part of but being a data platform but i think the other part of it is the fact that our aspirations are not to build posit past services that developers can go build their own solution that's on top of m and if you go try to build an application web application or mobile application on top of a past you need more than just storage right um and and the thing is is that we also bring those other types of it's events business logic insights security extensibility and a lm and of course it's an a p i ter on top of it like we bring those to bear as well as a part of the platform because first and foremost data verses built on a sure it is built on all these composite technologies so we aspire to be not just another sequel data based but essentially that full platform for bringing your data and then be able to build kind of interconnected applications and processes a top that be through the power platform for dynamics or you know through other means in the in the next three years to come it so that was a really long explain like five but effectively data verse is not an asracequl data base it's azure in a box in other words

[mark]: 10:02
yeah that's brilliant because you know i remember back in the day when we had things on prim and so pre anything online and back then you know you would have to set up a uh your front answer is your back end service your sequel for service your exchange service there was a lot more than just storage right there was a lot more than just the data base that you had to line up there was a whole bunch of other services and of course i like the way you describe that is a really in a box and in a highly ap driven consumable way including

[james_oleinik]: 10:33
that's right

[mark]: 10:33
all compute memory et cetera and of course his multiple storage technol he's now used in the platform there was when when that shift happened when we saw the splitting away of data into things like blob storage and cosmos deb has there been any more thinking like has there been much of an evolution under the hood of what's going on like one of the things that i remember a couple of years ago the conference hustling ryan i was when when you know are you thinking about how are you thinking about a i at a platform level like is there learning that could happen ah you know whether it's design protocols or or i don't know like how do you think i suppose that my question was two parts to him how you thinking about security at something that would never potentially be surfaced and a i that would potentially be service but was creating a more robust intelligent system or ecosis that you have created with the data this

[james_oleinik]: 11:43
yeah so i think i think security and a i are both separate beasts and enough themselves but of course they're they're inter connected um so i think if you look at first of all where dat verse came from right this very operational world kind of and i was very transactional in that right a builder like kind of like when it was launched actually might have been like more like three years ago right it was still it was very much around bringing a i to this transactional business process right i want to recognize this form i want to process this image right and and that conti used to be the case today and right and the way we brought in a builder is it's effectively a first party component it's as it's kind of a composite part of that business process the same way you would string together you know a plugging call and extensibility call a trigger in flows was just kind of part of the platform um now we're where i think about security an a i going forward and you mentioned kind of these new storage paradems and i think that's where things do become interesting because part about data verse living up to its monitor is more than just a sequel data bat it has to live up to the fact that it's more about just highly normalized relational models of data based on sequel um and we leverage cosmos d b when we want to optimize for certain types of work loads and certain types of high performance looks right we want to leverage the appropriate search capability is at cognize search under behind the scenes when we want to index things based off of certain types of quare patterns right we want to use a d l s and dataliketechn logs for more analytical scenarios right and so i think the point i'm getting to is that data us is expanding to these other storage technologies right and the way we build on the current oxtensibility paradims right the types of capabilities take a as an example how does it plug into kind of a real time or even an event streaming type of work load that we pull together how do you bring a i to an analytical space where data is not as highly normalized it's in star schema it's at a different scale and this is the forefront of our thinking right now and i'll give you one example um that one of the sceneras where we're doing right now and although it's not it's i think it's on our release notes is the fact that will be supporting essentially seamless archival of data from our operational store to our cold storage so dat every date of our environment now comes with essentially data like tech ology right and what you'll be seeing kind of this fall is the fact that you can set up our civil policies for your environment to start moving that data it's still there right it's gonna but it's going to be moving from that kind of hot or of sequel to the colder storage of data lake underneath the covers and guess what we're going to bring compute like snaps to bear to your same app your same apis that that's still going to be honored that same security for your operational ceneros are going to be honored but re moving things to a different place right and number one that's powerful to live up to the data platform monitor versus just the sequel data base monitor but then the fact that that data can also be made available in that kind of more file based paratime now we're thinking about how do we bring new aiparadyms to bear because you can store more data and data like you can bring together more more interconnected data together to run kind of new types of ai models this is what you see for example a lot of customers doing even with sanaplinkof data verse but they dropped that data into senaps they have pipe plans that bring together the rest of their data together from their enterprise to orchestrate you know new insights new features now m l work loads snap spark type of work loads right and they're doing that today leveraging kind of like past technology but what if you could do that within data verse and that certainly on that's on our mind as we kind of expand the technology s how do we bring more value to bear over time if that makes sense

[mark]: 16:07
so it's obviously getting you know when you say a zero in a box that's definitely starting to consume a lot of the other services available in a zoo is you know correct me if i'm wrong but the bizaps team which a lot of its under penned with data verse is it the biggest consumer of azure nowadays i think there's

[james_oleinik]: 16:29
now

[mark]: 16:29
talk of it

[james_oleinik]: 16:30
it depends on which meter you're looking at i think we're certainly up there if not the highest in terms of as your sequel um and some of the compute pair times were certainly up there and growing in cosmo t b growing in a d s and growing in some of the if you look at like as cognitive search right i'm not sure if we're the number number one vendor and all of those but we are absolutely growing and i'd say probably holistically we're probably up there for some of the most as our consumed revenue of any kind of ice feed out there um and and yeah and like i think it's the best way of explaining kind of why data verse is because we aspire to bring the value of what you would have to compile on your own with those requite technologies and we're trying to bring those to bear in a low code way so that you can accelerate your own development and not have to manage all those pieces yourself

[mark]: 17:27
yeah you know if we look at the origins or where data verse came from there would have been choices made back then in architecture around field types and i was on another course somebody around the tool to migrate azor sorry migrate access access base data bases into the environment and there was there was some limitation around it might have been floating point is it that there's different size between the way access and and data verse treated when you think of are there any things that are that you think that at some point you're going to need to tackle around what's legacy that is not either well your telemetary will tell you if it's not been used i suppose it's an easier decision but around perhaps not what's legacy but what's what are you thinking about new what should what should be aditive because of the way the world is changed whether we're getting closer to things like quantum computing and things like that like we know so much more than we did eighteen years ago when i first start and with you now dynamics the world the technology s advanced so much so when you're looking at that data structures what are you thinking

[james_oleinik]: 18:52
m so i to me i think i don't actually think of it is in terms of kind of like a data type for a column um i actually think to me i think part of the revolution that's coming is that there's a couple of things that are going on and i think data versus kind of at the center of it i think number one there's just i think that people believe that oh because we have cloud that for example every every company has been fully has realized their digital transformation and far from it i mean we far from i think realized the amount of data that even had kind of moves from on prem to the cloud it um and i think there's a second thing which is that i think that there are other essentially parades of data that we have only yet to really tackle and bring in to kind of even just thinking around like more line of business type scenaras like we haven't yet realized what means to stream video right and have that have that be interconnected and trigger work flows right like

[mark]: 20:00
wow

[james_oleinik]: 20:01
we we can't even like to be honest with you we're having trouble just like streaming events from a website and to be able to act on them right because if you look at like there's so much structure and value to kind of a common com record you know and there's so much rich semantics around it and in the the owners you're taking on the person entering you can take real clear action on that but the same cannot be said for a stream of data coming in for a particular process less can be said around a stream of bites about a bunch of images coming in and so yer to me like that is one thing that not only like just the historical operational data that has yet to be digitized let alone these new streams of data which are much less rich it will be much more higher in volume and that to me is i think the kind of next horizon that business apps and i think a lot of cloud technologies cloud data have to figure out how we're going to tackle and to me a i to go back to your previous conversation is part of that because how else are you going to build a business process round a stream of event data or a stream of video if there's not something automatically ind of deriving some meaning or semantics around it that you can tie essentially a process to right it's not really that possible and so i think there's so much to me that's really really interesting and it's kind of like that next horizon of data scale um that we have to figure out right and and the good news is that there's a bunch of really smart people in azure building past technologies there's a bunch of people the industry pushing the needle on that and so i'm confident that maybe not quite as far out as quanta computing but just talking about like bringing more data in through different paradimes i am confident that like that is very much in the in our wheel house in the next next three years to go and start to realize some of those so snares within even within the realm of data verse i think

[mark]: 22:05
nice nice tell us about turnkey insights what are you thinking now

[james_oleinik]: 22:11
i mean when i think turnkey insights i asked the question it was the last time that you've shipped an n to n business process without a report in it when's the last time someone's looked at that report and not asked a question right and how how hard was it when you asked that question how much work was it to go like at that back right and then my last question would be once you have that insight didn't you ask like hey why isn't that back in my source system for me to act upon an ameninful way right and so to me it's just the more scenaros i see brought to the platform and i think a great example of them was at loman field which is one of the stadiums here in the puget sound area when they were doing vaccine management date they chose the power platform right and so they were literally doing you know registrations on a portal that data flowing into data verse people having running powerapps on tablets checking people in an and there was an operational report and power connected to data verse right there tracking everything tracking inventory driving processes and so like to me like in a in a nutshell turnkey insight about you can't have an enton solution that's really complete these days that doesn't have analytics and visualizations that are part of that and we can do so much more than just in my mind a bunch of reports and statements on top of our data like we can actually derive insights and add a lot on top of that and to me that's really what it's a part of because i just think it's just inherent part of any composite solution customers ends up piecing together and we should make much more turn key within the platform today and that is some area that we're trying to actually invest in

[mark]: 23:58
i like it you mentioned health care there as with the covid scenario you know building apse out and the data sitting within data verse tell me about some other just reads that you have seen without mentioning names like really

[james_oleinik]: 24:15
sure

[mark]: 24:15
embrace the technology like that of that you're like wow you're pushing the limits we love this

[james_oleinik]: 24:23
well you know i don't i don't even have to not name names because like the cool thing is crost has the industry clouds right and it's it's really part in personal with the way that microsoph sells one microsoph to customers right and there is a cloud cloud for health care there is a cloud for for retail there is a cloud for sustainability right there is a cloud for financial services and guess what they're all built on they're all built on dynamic staffs which are built on data verse right and the way that they go and go to market they go the market with these applications that are built on the platform and they are able to customize and tailor them even within those industries to the work loads of the customers right and that's the power the platform right and so like i don't have to go too far i can name names because it is a core part of the way that microsoph lands that on microsopvision with our cloud strategy and actually there's if you do get another person on your podcast you should try to get some other cory sanders or someone for that team

[mark]: 25:28
nice

[james_oleinik]: 25:29
because you know he's helping lead that strategy for within kind of scotus organization right now for how we're really going after those kind of industry plays

[mark]: 25:41
i like it i don't know in your role i know other people in your in your team does we should go out and meet with some clients from time to time right you might be involved in some executive briefings and such what do you do when c to says to you so tell us about the limitations like you know just like a job interview you so what re you what are your strength what are your weaknesses that thing what are the things that you how do you explain limitations with the technology

[james_oleinik]: 26:12
well so like maybe i'll give you a couple of examples right um so you know if you take kind of like the four tenants of like what we aspire to be like what do we think makes data verse unique right they are follows you know we we believe that we have the potential to be kind of as easy as excel but connected to your full enterprise data we believe that we can accelerate your development by bring the power of azure in like low code functional language like service side language we believe we bring all the n we bring enterprise grade security in compliance that you've already implemented and your trusted clowns we believe that where this planet scale data platform that doesn't has no limits right these are these four tenants but i would say that like look yeah we aspire to be as easy right but it's still to hard today right and customers the activation energy is too much people people are able to figure it out right but it's because and once they see that value they keep building on it but it is too hard right it is not approachable enough right like we do bring the power of azure to bear but there's still too much pro code there's too much in java script there's too much in plug ins that are folco right you know we do bring the power of those hundreds of hours that customers spent on figuring security and compliance for three sixty five and azure we bring that to bear but it's not turnkey right you do have to con figure more controls you do have to understand kind of our security concerns we don't do enough by default right and we aspire to be this planet gale data platform but there are still limits today right we're still proving that out right like many customers have to kind of like like even delete or build their own orcival solutions today right in order to had some of those limits and we know that so like what i would say is that look like we are heads and tails where we were three years ago when this face was that of ryan jones and there was red hair here okay and so we had advanced every single one of these kind of value props like we've shipped better make our experiences there is more that's easier that's brought to power effects and easier to build on there is so much more you get out of box in terms of security and our limits have gone up probably tenfold since then but we have a long way to go right is basically what i would say um and and that's why like were like we have a corpproduct team this is core to busnbstrthis is like in my mind if we achieve kind of like the goals that charles has set forward for kind of his organization right like danvers has to grow we have to scale to the next scale of orkloads and the next order of magnitudes of customers in order for us to be successful and we will i'm very confident in that

[mark]: 29:01
i like it it's been very interesting discussion with you james i've enjoyed it is there any final things at a top of mind that you'd like to say before i let you go

[james_oleinik]: 29:13
well i think number one i'd like to hear from the new zealand community around both why data verse and why not data verse right and so i think the feedback from the community right is golden and treasure right and we do believe we have something here and we do believe we have a long way to go and so please give us that feedback you know we were joking before the podcast started which is maybe i won't be able to get out to new zealand but if there are opportunities or using groups to come in and hear that feedback i am game and my team are game because we think we have something special here we think that microsoph has something special here with the power platform with dynamics with the platform that kind of spans it all and so you know if you're interested if you're building on top of if you couldn't build something let us know let me know i'm really interested to hear that for you because we have to

[mark]: 30:11
i like it

[james_oleinik]: 30:12
we have to figure it out in my mind

[mark]: 30:14
well the data centers open in new zealand about mid next year and of course you know once the m three six five work loads have moved in and and stabilized of course dynamics short step behind and data verse in everything power plate so

[james_oleinik]: 30:30
that's

[mark]: 30:30
maybe

[james_oleinik]: 30:30
right

[mark]: 30:30
she'd come for the big launch party that i'm sure is going to happen once we've got our own data centers in country

[james_oleinik]: 30:36
what do they do instead of breaking champagne do they like break a dev box or storage

[mark]: 30:42
and i would hope it's still i hope it's still champagne

[james_oleinik]: 30:46
fair enough, fair enough there you go

[mark]: 30:46
even though it's not a ship but it's going to be phenomenal anyhow thanks james

[james_oleinik]: 30:51
thank you so much