Andy Mendelsohn, Oracle | CUBE Conversation, March 2021

Andy Mendelsohn, Oracle | CUBE Conversation, March 2021

Show Video

[Music] the cloud has dramatically changed the way providers think about delivering database technologies not only has cloud first become a mandate for many if not most but customers are demanding more capabilities from their technology vendors examples include a substantially similar experience for cloud and on-prem workloads increased automation and a never-ending quest for more secure platforms broadly there are two prevailing models that have emerged one is to provide highly specialized database products that focus on optimizing for a specific workload signature the other end of the spectrum combines technologies in a converge platform to satisfy satisfy the needs of a much broader set of use cases and with me to get a perspective on these and other issues is andy mendelson is the executive vice president of oracle the world's leading database company andy leads database server technologies hello andy thanks for coming on hey dave glad to be here okay so we saw the recent announcements this is kind of your baby around next generation autonomous data warehouse maybe you could take us through the path you took from the original cloud data warehouses to where we are today yeah when we uh we first brought autonomous database out uh we were basically a second generation technology at that point you know we decided that what customers wanted was to the other you know the push of a button provision the really powerful oracle database technology that they've been using for years and um we did that with autonomous database and beyond that we provided a very unique capability that around self-tuning self-driving of the database which is something the first generation vendors didn't provide and this this is really important because customers today are you know developers and data analysts you know you know at the push of a button build out their their data warehouses but you know they're not experts in tuning and so what we thought was really important is that customers get great performance out of the box and that's one of the really unique things about autonomous data warehouse autonomous database in particular and then this latest generation that we just came out with also answers the questions we got from you know the data analysts and developers they said you know it's really great that i can press a button and provision this very powerful data warehouse infrastructure or database infrastructure from oracle but you know if i'm an analyst i want data you know so it's still hard for me to go and you know get data from various data sources transform them clean them up and get them to a way a place where i can start querying the data now i still need data engineers to help me do that and so we've done in the new release we said okay we want to give data analysts and data engineer data scientists developers is a true self-service experience where they can do their job completely without bringing in any you know any any engineers from their i.t organization and so that's what this new version is all about yeah awesome i mean look years ago you guys identified the i.t labor problem and you've been focused on r d and putting it in your r d to solve that problem for customers so we're really starting to see that hit now now gartner recently did some analysis they ranked and rated them some of the more popular cloud databases and oracle did very well i mean particularly particularly in operational categories i mean an operational side and the mission critical stuff you smoked everybody we had mark stamer and david floyer on and our big takeaways were that you're you're again dominating in that mission critical workloads that that that dominance continues but your approach of converging functionality really differs from some others that we saw i mean obviously when you get high ratings from gartner you're pretty stoked about that but what do you think contributed to those rankings and what are you finding specifically in customer interactions yeah so gardner does a lot of its analysis based on talking to customers finding out how their product these products that sound great on paper actually work in practice and i think that's one of the places where oracle database technology really shines it's it's uh it solves real-world problems um it's been doing it for a long time and as we've moved that technology into the cloud you know that continues you know the differentiation we've built up over the years really stands out you know you look at like amazon's databases they generally take some open source technology that isn't that new it could be 30 years old 25 years old and they put it up on the cloud and they say oh it's cloud native it's great but but in fact it's the same old you know technology that that doesn't really compete you know decade behind oracle's database technology so i think the gartner analysis really showed that sort of thing quite clearly yeah so let's talk about that a little bit because obviously i've learned a lot you know one of the things i've learned over the last many years of following this business a lot of ways to skin a cat and cloud database vendors if you think about you mentioned aws you know look at snowflake kind of right tool for the right job approach they're going to say that their specialty databases they're focused uh are better than your converged approach which they make you know think of as a you know swiss army knife what's your take on that yeah well the converged approach is something of course we've been working on for a long time so the the idea is pretty simple you know think about your smartphone you know if you can think back you know over 10 years ago used to have you know a camcorder and a a camera and a messaging device and also a dump phone device that all those different devices got converged into what we now call the smartphone why did the smartphone win it's just simply much more productive for you to carry one device around that that is actually best to breed in all the different categories instead of lots of separate devices and that's what we're doing with converge database over the years you know we've been able to build out technologies that are really good at transaction breasts at analytics for data warehousing now we're working on you know json technologies graph technologies the other vendors basically can't do this i mean it's much easier to build a specialty database that does one thing to build out a converged database that does end things really well and that's what we've been doing for years and again it's it's based on technology that uh you've invested in for quite a long time um and it's something that i think uh customers and developers and analyze analysts find to be a much more productive way of doing their jobs it's very unique and not common at all to see a technology that's been around as long as oracle database to see that sort of morph into a more modern platform i mean you mentioned aws uses leverages open source a lot you know snowflake would say okay hey we are born in the cloud and they are i think google bigquery would be another good example but but but that notion of boy i want to get your take on this born in the cloud those folks would say well we're superior to oracle's because you know they started you know decades ago not necessarily you know native cloud services uh how have you been able to address that i know you know cloud first is kind of the buzzword but but how have you you made that sort of transparent to users or or irrelevant to users because you are cloud first maybe you could talk about how you've able to achieve that and convince us that you actually really are cloud native now you know one of the things we we sort of like pointing out is that um oracle very uniquely has had this scale out technology for running all kinds of workloads not just analytic workloads which is what you see out in the cloud there but we can also scale out transaction processing workloads now that was another one of the reasons we do so well in for example the gardner analysis for trans operational workloads and that technology is really valuable as we went to cloud it lets us do some really unique things and the most obvious unique thing we we have is something we like to call you know you know cloud native you know instant elasticity and so with our technology if you want to provision a share you know some number of amount of compute to run your workloads you can provision exactly what you need you know if you need 17 cpus to get your job done you do 17 cpus when you provision your autonomous database our competitors who claim to be born in the cloud like snowflake and amazon they still use this this archaic way of provisioning uh servers based on shapes you know snowflake you know says what which shape cluster do you want you want 16 you want 32 you want 64.

no it goes up by a power of 2 which means if you compare that to what oracle does you you have to provision up to like twice as much cpu than you really need so if you really need 17 they make you provision 32. if you really need 33 they make your provision 64. so this is not a cloud native experience at all it's an archaic way of doing things and and we like to point out with our instant elasticity you know we can go from 17 to 18 to 19 you know whatever you want plus we have something called auto scale so you can set your baseline to be 17 let's say but we will automatically based on your workload scale you up to three times that so in this case be 51 and because of that true elasticity we have we are really the only ones that can deliver true pay as you go kind of you know just pay for what you need kind of capability which is certainly what amazon was talking about when they first called their cloud elastic but it turns out for database services these guys still do this archaic thing with shapes so that's a really good example of where we're quite better than the other guys and it's much more cloud native than the other guys i want to follow up on that uh just stay here for a second because you're basically saying we have we have better granularity than the so-called cloud native guys now you mentioned snowflake right you got you got the shapes you got to you got to choose which shape you want and it sounds like it sounds like redshift the same and of course i know the way in which amazon separates compute from storage is largely a tiering exercise so it's not as as is as smooth as you might expect but nonetheless it's it's good how is it that you were you were able to achieve this with a database that was you know born you know many decades ago is it i mean what is it in from a technical standpoint an r d standpoint that you were able to do i mean did you design that in in the 1980s how did you how did you get here yeah well um it's a combination of interesting technologies so autonomous database you know it has the oracle database software that software is running on a very powerful optimized infrastructure for database based on the exadata technology that we've had on prem for many years we brought that to the cloud and that technology is a scale-out infrastructure that supports you know thousands of cpus and then we use our multi-tenant technology which is a way of sharing large infrastructures amongst amongst separate uh clients and we divide it up dynamically on the fly so if there's thousands of cpus you know this guy wants 20 and this one wants 30 we we divide it up and give them exactly what they need and if they want to grow we just take some extra cpus that are in reserve and we give it to them instantly and so that's a very different way of doing things and that's been a shape based approach where you know what what snowflake and amazon do under the covers they give you a real physical server you know or a cluster and that's how they provision if you want to grow they give you another big physical cluster which takes a long time to get the data populated to get it get it working we just have that one infrastructure that we're sharing among lots of users and we just give you a little extra capacity we don't it doesn't it's done instantly there's no need for data to be moved to populate the new clusters that you know snowflake or amazon are provisioning for you so it's a very different way of doing things and you're able to do that because of the tight integration between you mentioned exadata tight integration between the hardware and software we got david floyer calls it the iphone of enterprise sometimes sometimes you get some grief for that but it's it's not a bad metaphor but is that really the sort of secret well the big secret under the covers is this you know exudated technology our real application cluster scale out technologies our multi-tenant technologies so these are things we've been working on for a long time and they are very mature very powerful technologies and they really provide very unique benefits in a cloud world where people want things to happen instantly and they want to work well for any kind of workload um you know that's that's why we call we talk about being converged we can do mixed workloads you can do transactions and analytics all in the same data the other guys can't do that you know they're really good at like you said a narrow workload like i can do analytics or i can do graph you know i can do json but they can't really do the combination which is what real world applications are like they're not pure one thing versus enough right thank you for that so one of the questions people want to know is can oracle attract you know new customers that aren't existing oracle customers so maybe you could talk about that and you know why should uh somebody who's not an existing oracle customer think about using autonomous database yeah that's a that's a really good question you know oracle if you look at our customer base has a lot of really large enterprises you know the biggest banks and the biggest telcos you know they run oracle they run their businesses on oracle and these guys are sort of the most conservative of the bunch out there and they are moving to cloud at a somewhat slower rate than the than the smaller companies and so if you look at who's using autonomous database now it's actually the smaller companies you know the same type of people that first decided amazon was an interesting cloud 10 years ago they're also using our technologies and it's for the same reason they're finding you know they don't have large it organizations they don't have large numbers of engineers to engineer their infrastructure and that's why cloud is so attractive to them and autonomous database on top of cloud is really attractive as well because you know information is the lifeblood of every organization and if they can empower their analysts to get their job done without lots of help from it organizations they're going to do it and you know that's really what's made autonomous database really interesting you know the whole self-driving nature is very attractive to the smaller shops that don't have a lot of sophisticated um i.t expertise all right let's talk about developers you guys are the stewards of the java community so obviously you know big probably you know the biggest most popular programming language out there but when i think of developers i think of guys in hoodies pounding away but when i think of oracle developers i might think of maybe an app dev team inside of maybe some of those large customers that you talked about but why would developers and or analysts be interested in in using oracle as opposed to some some of those more focused narrow use databases that we were talking about earlier yeah so if you're a developer um you want to get your job done as fast as possible and so having a database that gives you the most productive application development experience is important to you and so you know i was talking we've been talking about converged database off and on so if i'm a developer i have a given job to do a converged database that lets me do a combination of analytics and and transactions and do a little json and little graph all in one is a much more productive place to go because if i if i i don't have something like that then i'm stuck taking my my application and breaking it up into pieces you know this piece i'm going to run on say aurora on amazon and this piece i have to run on the graph database and here's some json i got to run that on some document database and then i have to move the data around the data gets sort of fragmented between these databases and i have to do all this data you know integration and and whatever with a converged database i have a much simpler world where i can just use one technology stack i can get my job done and then i'm future proof against change you know requirements change all the time so you build the initial version of the application and your users say you know that this is not what i want i want some something else and it turns out that something else often is why i want analytics and you use something like a you know a document stored technology that has really poor analytic capabilities and then so you have to take that data and you have to move it to another database and so with with our converged approach you don't have to do that you know you're already in a place where everything works everything that you need you can possibly need in the future is going to be there as well and so for developers i i think you know converged is the right way to go plus for people who are what we call citizen developers you know like the data analysts that they cuddle they write a little code occasionally but they're really after getting value of the data we have this really fabulous no code loco tool called apex and apex is again a very mature technology it's been around for years and it lets somebody who's just a data analyst he knows a little sql but doesn't want to write code get their job done really fast and we've published some benchmark on our website showing you know basically you can get the job done 20 to 40 times faster using a no co loco tool like apex versus something like you know just writing cutting lots of traditional code i'm glad you brought up apex we recently interviewed one of your former colleagues amit xavery and all he would talk about is low code no code and then in the apex announcement you said something to the effect of coding should be the exception not the rule did you mean that what do you mean by that yeah so apex is a tool that people use with our our database technology for building what we call data driven applications so if you got a bunch of data and you want to get some value out of it you want to build maybe dashboards or more sophisticated reports apex is an incredible tool for doing that and it's it's modern you know it builds applications that look great on your smartphone and it automatically you know renders that same user interface on a bigger device like a laptop desktop device as well and uh it's very it's one of these things that uh the people that use it just go bonkers with it it's a viral technology they get really excited about how productive they they've been using it and they tell all their friends and i think we decided uh i guess about a year ago when we came up with this apex service that you know we really want to start going bigger on the marketing around it because it's very unique nobody else has anything quite like it and it's it again it just adds value to the whole developer productivity story around an oracle database so uh that's why we have the apex service now and we also have apex available with every oracle database on the cloud god i want to i want to ask you about some of the features around 21c there are a lot of them you announced earlier this year maybe you could tease out some of the top things that we should be paying attention to in 21c yeah sure um so one of the ways to look at 21c is we're we're continuing down this path of a converged database and so one of the the marquee features in 21c is something we call blockchain tables so what is blockchain well blockchain was this technology that's under the covers behind bitcoin you know it's a way of creating a tamper-proof data store um that was used by the original bitcoin algorithms well developers actually like having tamper proof data objects and databases too um you know and so what we decided to do was say well if i create a sql table in an oracle database what if there's a new option that just says i want that table implemented using blockchain technology to make the table tamper proof and fully audited etc and so we just did that and so in 21c you can now get a basically another feature of the converged database that says uh you know give me a sql table i can do everything i can query it i can insert rows into it but it's it's tamper proof i can't ever update it i can't delete rows from it amazon did the their usual thing they took again some open source technology and they said hey we got this great thing called quantum ledger database and it does blockchain tables but but if you want to do blockchain tables in any of their other databases you're out of luck they don't have it you have to go move the data into this new thing and it's again one of their it's again showing sort of the problem with their their proprietary this proprietary approach of having specialty databases versus just having one conversion that does it all so that's the blockchain cable feature uh we did a bunch of other things um the one i i think is worth mentioning the most is is support for persistent memory so a lot of people out there haven't noticed this this very interesting technology that intel shipped a couple years ago called optane data center memory and what it is it's basically a hybrid of flash memory which is persistent memory and standard dram which is not persistent means you can't store a database in dram um and so with this persistent memory you can basically have a database stored persistently in memory all the time and so it's a very innovative new technology from a database standpoint it's a very disruptive technology to the database market because now you can have an in-memory database basic period all the time 24 7. and so 21c is the first database out there that has native support for this new kind of persistent memory technology and we think it's it's really important so we're actually making it available as uh to our 19c customers as well and uh you know that's another technology i'd call out that we think is very unique we're way ahead of the game there and we're going to continue investing moving forward in that space as well yeah so that layer in between dram and and persistent flash that's that's a great innovation and good game changing from a from a performance and actually the way you write applications but i gotta i gotta ask you i and all the analysts were wrong with juan recently juan loyza and and to listen to that introduction of blockchain and everybody wants to know is safra going to start putting bitcoin on the oracle balance sheet i'm about to get that leap yeah that's a good question who knows yeah i can't comment on speculation ah that would be interesting okay last question then we got to go uh look oracle the narrative on oracle is you're expensive and you're mean you know it's hard to do business with do you care are you doing things to maybe change that perception in the cloud yeah i think we've made a very conscious decision that as we move to the cloud we're offering a totally new business model on the club that is a a cloud-native model you pay for what you use um you have everyday low prices you don't have to negotiate with some salesman for for months to get get a good price um so yeah we really like the message to get out there that those of you who think you know what oracle's all about um you know i and how it might be to work with oracle on in from your on premises days um you should really check out how oracle is now on the cloud we have this autonomous database technology really easy to use really simple any analysts can help get value out of the data without any help from any other engineers it's very unique it's it's uh it's the same technology you're used to but now it's delivered in a way that's much easier to consume and much lower cost and so yeah you should definitely take a look at what we've got out there on the cloud and it's all free to try out we got this free tier you can provision free vms free databases um free apex whatever you want and uh try it out and see what you think well thanks for that i was kidding about me and a lot of a lot of friends at oracle some relatives as well and thanks andy for coming on thecube today it's really great to talk to you yeah it's my pleasure and thanks for watching this is dave vellante we'll see you next time [Music] you

2021-04-03 19:29

Show Video

Other news