Larry Ellison Introduces Oracle Autonomous Database Cloud

statements in this presentation relating to Oracle's future patient's beliefs intentions and prospects are forward-looking statements and are subject to material risks and uncertainties a detailed discussion of these factors and other risks that affect our business is contained in our SEC filings including our most recent report on Form 10-k and Form 10-q particularly under the heading risk factors this presentation is intended to outline our general product direction it is intended for information purposes only and may not be incorporated into any contract it is not a commitment to deliver any material code or functionality and should not be relied upon in making purchasing decisions the development release and timing of any features or functionality described for Oracle's products remains at the sole discretion of Oracle please welcome Larry Ellison [Applause] [Music] [Applause] hello everybody everybody welcome and recall that at Oracle OpenWorld last fall we announced the Oracle autonomous database and this is an update on the autonomous database now that it's generally available to all of our customers ok so right now our first autonomous service is available in the Oracle public cloud and it's the autonomous data warehouse and it became available just about fully available just about a week ago even though for months our customers have been using it and trying it and moving their data and moving their applications to the cloud and actually when I finish with this update we're gonna have a demonstration and then we're gonna have a customer panel and they can talk about their experiences with the Oracle autonomous database that's a good time to be talking to talk about our kind of our strategy versus Amazon the the Oracle autonomous database is the first of several autonomous pass services we will be delivering this year now everyone knows and everyone gives rightful credit to Amazon for kind of inventing the market for infrastructure as a service and they noticed that in a lot of ways computers it's more efficient to rent computers than to buy computers so Anna Amazon pioneered the notion of putting a virtual machine on a computer and renting you compute and renting you storage and letting you go ahead and build applications run applications on their cloud when using this thing called infrastructure as a service well I think we all think infrastructure of services is very valuable Oracle is going into that business but the way Oracle plans on and as is the process of differentiating itself from Amazon is to offer a complete suite of platform services that are at a higher level than low-level infrastructure stuff in other words if you built an application infrastructure as a service it's not very much different and that'sthat's the whole idea not very much different from you having bought a computer gotten a virtual gotten a VM gotten an operating system put up a network and you're using the could not those compute resources and those storage resources in your data center Amazon said no do that in our data center and the process is a little bit as a little bit as faster it's easier for you to acquire computers it's faster you come acquire storage and you're very familiar with this particular OLS and this VM you know how it works you know how to develop apps and you can do it over here in our data center in our cloud rather than your data center and you can save money we do that but that's not our strategy going forward our strategy going forward is to develop a high level layer of software called pass services that sit on top of infrastructure so rather than developing database applications like you used to develop you'll have a new a new set of tools for applications development for mobility for integration services for database services that are not business as usual you know writing programs on infrastructure need a cloud infrastructure or your data center infrastructure instead you'll be using these very high level services that'll allow you to build applications must much faster mobile applications much faster much more securely much more safely than you ever could so these autonomous path services basically allows your organization to become much more efficient in developing applications and running applications spend much much much less money developing applications running applications and the resultant application to themself are dramatically more secure and dramatically faster and therefore lower cost then had you developed them the old way and the first to these past services that we're going to talk about is the autonomous database now the autonomous database is and is my slide is based on a technology as revolutionary as the internet you know Oracle used to have a tagline the internet changes everything well this technology same deal changes everything machine learning for the first time the great promise of AI and AI was developed by a guy named Josh Lederberg at Stanford University many many years ago later on believe it or not he ran my medical foundation he's he's Nobel laureate ran Rock ran Rockefeller was president of Rockefeller they're great vial of Biology Research Institute in New York but believe it or not he was the father at Stanford University of artificial intelligence and for a long time people looked at the promise of AI but it never quite delivered to its promise until very recently and with the advent of the latest generation of AI neural networks combined with machine learning we are doing things that hitherto of considered unimaginable by computers you know the applications the applications are self-driving cars which even though we had an accident Ray's a tragic accident recently will be should be much safer much safer than human drivers who get in accidents all the time we lose about 30,000 people a year in the United States to traffic accidents all over the world obviously the number is much much much larger so self-driving cars will not only be much more much more more safe than human drivers they will allow transportation services that are extraordinarily low cost already people are finding it easier to ride around in ubers than own a car and drive himself and the car and do all of those and do all of those things imagine when the ubers don't have drivers or the drivers really are computers how much did how much less expensive that's going to be how much safer that's going to be and we have every important car company in the world is working on this right now everyone and they've made a trim and a lot of companies that aren't current companies are working on it they've made enormous progress computer vision facial facial recognition most people I think have smartphones or tablets that you no longer log on to and an editor enter passwords maybe maybe for a while used your fingerprint and now even that's too inconvenient as it does facial recognition and then everyone kind of understands how that works they under seen that application more remarkable our cities in China where the stoplight recognizes every single person who lives in that city and every single person who visits that city and catalogues that they just crossed at this intersection at that time basically it knows where you are and what you're doing pretty much all the time well that's an in fact China is the leader right now in facial recognition the you can you can like or dislike that particular application in terms of its implication on privacy but the technology the technology is extraordinary and we could have a very interesting debate about these technologies I mean is that technology dangerous could it be used for evil could it be used for keeping true you know keeping track of people yeah invading your privacy every technology every technology has the potential for misuse I can imagine I mean clearly every every Technol every technology has been at times misused I can imagine back the first cave person who discovered fire and a person that was sharing the cave with him or her which whoever discovered fire said I can't believe you discovered fire we can't let the secret get out people can mute misused the fire for all sorts of terrible things people can be burned but no but just cook food and protect us against wild animals heat up to cave in the winter no no that's true but it's terrible it's guy no you know two people who do horrible things as someone's be burned at the stake one day and there was this huge debate well it's I mean airplanes every every every technology can be used but by-and-large technologies have improved our lives and made lengthened our lives made us more healthy and you helped educate us and generally have proved improve general welfare and prosperity of course any technology can be misused we are using machine learning to make our software smarter to make our software self-driving okay a two-minute tutorial on machine learning what machine learning does is it ends what you it learns kind of like people learn it ins it learns through experience it learns by inspecting large quantities of data just like little kids look around look around the room constantly processing lots and lots and lots of data they listen to people - okay the people talking to them and eventually they they they learn they also learn how to talk by understanding and repeating patterns that they're hearing machine we're learning kind of works the same way you you know they look at large not large amounts of data and they detect patterns in that data so the very first thing machine learning ever detected was a cat he was looking a lot of visual information and eventually it said I've seen that before and it was a cat picture so gradually distinguishes one kind of object from another kind of object it distinguishes normal patterns in the data from a novel anomalous patterns in the data so if you teach a computer to look at cells has taught a lot of blood cells the computer could say wants a normal red cell that's a T cell blah blah blah it's going through all of this and then it will detect a cell that kind of hasn't seen before or it's very rare and that in fact they know it's an anomalous cell in some cases in certain kind of patterns they know it's a cancer cell similarly we if you process computer logs and you look at lots and lots of login data and it turns out that your CFO is always logging in from Palo Alto where they live or Redwood Shores where they work or New York where they're visiting and giving presentations or London but it was really strange when they locked it logged in from a village in the Ukraine that's an anomalous piece of data maybe we shouldn't allow that login we so we can detect so we can make our computer systems much more secure by having automated intelligent detection of anomalies anomalous logins or other anomalous things another example when you write an Oracle application especially an analytic application you end up sending the database queries to ask this question in that question and uh and if all of a sudden the system sees a query unlike you know the look doesn't look anything like a query they've ever seen before like you know even even it's not about social security numbers you know even even it's not kind of an obvious attempt to steal someone's identity it's just a sequel a sequel query they've never seen before you flag it you look at it it's an anomaly and it might be some attempt to break it to break into the system but this is all now that but now this is all automated this is all automated these introduced intrusion detection into your computer system must be automated because if cyber criminals and cyber terrorists are using networks of computers botnets networks of computers to attack our computers and we think are human beings can form a defensive wall against these computer attacks we are dreaming it has to be our computers versus their computers or we have no chance so we make this huge investment in automating our cyber security systems using machine learning detecting anomalous events so we then can respond this is by the way a very big deal I would say the Oracle autonomous database is probably the most important thing we have ever done and it's very different from what other people are doing in the data management industry and it's desperately needed because the word you know one of the very very worst things that goes on in our industry is data gets stolen personal data gets stolen there's massive fraud and okay your credit card gets stolen and someone charges a bunch of things to get your credit card that's bad you know that's bets very expensive much worse is when nation-states steal all of this information and use it to under undermine our essential infrastructure discover who our intelligence agents are around the world etc etc and that's happened and that has happened the Office of Personnel Management federal government a big federal office lost all of its applications all of its security applications were stolen by a state actor that allowed them to detect most of our intelligence agents and our embassies around the world so we need automated systems intelligent systems to protect our data first the first thing we have to do again is detect an anomalous event and do that in all I made it way that is necessary but not sufficient to protect ourselves to protect ourselves we have to then block any intrusion we have to detect the intrusion the attempted intrusion and then you have to block the intended intrusion and one of the unique things about the Oracle autonomous database is once an intrusion or a problem has been detected the system the system can automatically patch itself with no human intervention once once a flaw security flaw has been detected the system will automatically patch itself and that that vulnerability will be closed off without any scheduled downtime without any human intervention note what happened at a credit agency recently that lost a lot of data at law had a lot of data stolen they knew about their vulnerability for months and they actually attempted the vulnerability was in a data management system called Apache struts they detected there was a problem it was published that there was a problem they tried to find all instances of Apache struts that they had that that they had they found 80% of them they patched 80% of them six months later after the update the vulnerability was detected unfortunately the other 20 20 percent were not detected and they lost vast sums of of customer data and as a result the company the company management was discharged and company lost a tremendous amount of value and but the worst problem is people you know lost you know suddenly had their their identities were vulnerable now we know that that data was not stolen by fraudsters great thing as cyber criminals did not steal that data how do we know because the credit cards never showed up in the dark web that means nation nation states stupid it they're not trying to you know just steal your money okay the Oracle autonomous database is 100% self-driving it is fully autonomous it needs no human intervention it will provision itself it will tune itself it will patch itself it will upgrade itself it will add storage it will add compute it will delete compute it will do all of that without human intervention it is a hundred percent self-driving the automated the automated cyber defense is I wish we could say we had a hundred percent on that it's cyber defense we were constantly putting more and more information into our cyber into our cyber defense but it still is a combination it's Silva cut is a combination of human beings and and machine learning ok there are lots of benefits to autonomous systems the first one is the elimination of human labor I mean obviously there's an enormous cost savings to saying that a lot of things we normally thought of as database administration is now handled by the computer it's now fully automated there's again what I said provisions upgrades patches tooms does all of that expands contracts there are no fixed there are no fixed configurations you have to pay for this very different than Amazon you know it's interesting that Amazon calls itself the Elastic Compute cloud however if you look at Amazon's database to date two databases redshift and Aurora can they automatically expand as you need more capacity do they automatically get bigger when you need more capacity no not elastic at all you have to shut the system down deallocate the storage move the data from point A to point B you know get more compute capacity than bring it back up that's very different than if you need more compute capacity or more storage with the Oracle autonomous database you just say that storage at compute and it just does it automatically we're elastic they're not you pay for a fixed configuration and you have you have scheduled downtime if you want to change it but there's no human labor is a very big deal it's a very big deal because human beings make mistakes pilots make errors almost all airline crashes are pilot error not a problem with the airplane human beings make mistakes some human beings who you think work for you work for themselves or somebody else human beings can be malicious you have a large you know you're a large bank or whatever and you could have people working for you or who are not working in your best interest some people can be malicious and if you get people out of the database administration business that's a less opportunity for human error and eliminates the possibility of malice someone intentionally doing something to sabotage your systems so getting rid of and by the way the vast cause the vast majority of the cost of running an Oracle database or any kind of database for that matter is human labor it's not the cost of this of the disk storage it's not the cost of the compute it's not the cost of the flash storage it's not the cost of the network or the interface cards or anything else it's people people are the sources of the costs and people are the source of almost all the errors okay uh we have be cut because Oracle the article autonomous database is fully automated we can guarantee our SLA guarantees we will guarantee contractually guarantee 99.995% uptime and that includes everything that includes upgrades patching everything that's sick a couple minutes a month will be down there's nothing close there's nothing close to this as I said with Amazon you just want to add more storage to your database to your database yeah I don't want to add another processor database you have to shut the whole thing down you want to patch it you have to shut the whole thing down you want to do an upgrade to shut the whole thing down plus the system always it constant not only is this the tuning of the system fully automated it's continuous one of the great problems with databases although including the Oracle database historically has been when you built the application you loaded your data and then you tune your database you added indexes you created partitions you did all of these things to get it running just so and then you went off and we've worked and worked on your next project in the meantime the data distribution in your database changed the database got larger you got to let us to historical data in it the database could have changed its shape it's changed you know changed its size and all that all that clever tuning you did 18 months ago is no longer optimal the database performance isn't good the database hasn't been tuned for 18 months has been running for 18 months but it was tuned once in the beginning great thing about the Oracle anonymous database is conflict it's constantly collecting these logs that's constantly learning looking at queries it's constantly looking at results and it's constantly tuning itself creating indexes dropping indexes creating your purchases doing all of these things with continuous tuning this leads to a huge savings because we consume fewer resources because of this continuous tuning because this automated tuna what Oracle does is guarantee if you take a work load from Amazon that's running at Amazon either on an Oracle database at Amazon or on an Amazon redshift database and by June Aurora Amazon Aurora database and you move it to the Oracle cloud and we guarantee your bill from Oracle just your cloud bill I think do it later another way they know it's just a cloud bill will guarantee it's less than half of what Amazon's charging you just because our database is so the autonomous database is so much more efficient we compress the data we use less storage we have better algorithms we will use less compute there's no magic here you say well Amazon chargers and Oracle charges essentially the same thing for storage and the same thing for compute how could you be less than half the price I've look at your prices I've looked at your prices and they're exactly the same it can't be true well of Amazon charges by the minute and we charge by the minute and we do in one minute what it takes Amazon five minutes to do Amazon is five times more expensive than we are and that's pretty much as you'll see what the demonstrations about to come up that's pretty much what the ratio really is on an Oracle database we Amazon running an Oracle database to add Amazon will cost you five times five times what it cost you to run in the Oracle cloud because it will take you five times longer you use you'll use five times more resource five times more compute at Amazon then you will use it Oracle to do the exact same thing and we've tested it over and over again and we've invited other people to look at our numbers and look at our tests and if they can get better results for Amazon we're getting we'll publish your results but it's five times more expensive to run an Oracle database at Amazon than at Oracle and by the way running a redshift database at Amazon versus an Oracle database with Oracle almost 10 times more expensive at Amazon okay and but and of course and that's not even that's not even counting the labor savings by automating the DBA function that's not counting the cost of downtime that we have 99.995% uptime dramatically better than em do we have no cost for downtime because we have essentially no downtime we have no cost for database administration because it's completely automated we're not even counting that we're just taking the Amazon bill and saying that's by itself is five times higher all right dude the autonomous database is a combination of things it's we've taken Oracle 18c the latest version of our database and added a lot of features to the Oracle Oracle 18c we put it into the Oracle cloud the Oracle autonomous database it turns out runs entirely on Exadata infrastructure so it runs on this purpose-built infrastructure that's very very fast we compress all but we compress all the data storage we have a very fast interconnect we do all of these things but suffice to say the running on the exited infrastructure is way cheaper way cheaper than running on on the Amazon infrastructure and then we have we have automation driven by policy so you tell us when you setup a database whether that's a test Deb database it's tested test and development you don't need 99.995% availability so don't give me - don't give me disaster I don't want a pen I want to pay for disaster recovery I want to pay for this is that I want to optimize on cost not availability and reliability and recover ability in that case for test death a major major production database I want to guard against every failure I want a guard against you know even if your your data center goes down even there's a fire in your data center I wanted you know you've got to cover everything I did this application can never go down I want the 99.995 so you can you see you can optimize for lowest possible cost you can optimize for the highest possible availability you can optimize for the highest possible performance you can do combinations thereof but you just tell us what's important to you and we will based on those policies we will then do all the tuning create the indexes create via the the failsafe copies with active Data Guard in other data centers in other regions for that matter we'll do all of that in an automated fashion okay Oracle 18c has lots and lots of cool features in it it's a again it's now generally available in our cloud as part of the autonomous database we've put in lots of OLTP features lots of our all of our in memory stuff is getting just keeps getting faster and faster and faster that's not what I'm here to talk about or emphasize today again we put a lot of features into this but the big thing the big thing we've done here are the new machine learning the new machine learning algorithms the ability for this thing to run autonomously in the public cloud and by the way in what we call clouded customer so you can get this both in the public you know in the Oracle public cloud or you can get this in your data center where and something we call cloud and customer well we actually take an exudate a machine and what's called an oracle cloud machine and we actually build a portion of our of our public cloud if you will we put up a portion of our public cloud behind your firewall in your data center if you want that we gave you the option all right you know I've got very strict data privacy rules here in Switzerland I'm the bank I'm not allowed to move my data outside the country so I'm not moving it to you Oracle not to move it to your data center I really but I want to take advantage of all these features of the autonomous database so I really want I want this capability okay we can actually take and it really is our it's exact same hardware and software that runs in our cloud our public cloud we'll take that and build like a little mini part of our public cloud behind your firewall and then we can run our autonomous database behind your phone and by the way you pay for it exactly the same way as you would pay for it if it was in the public cloud so it's still subscription we still manage the database for you that's still autonomous well you know uh it's a joint responsibility if you will for managing the database we have some responsibilities and you have some you develop the applications you update you update the applications we backup the database we do you know we tune the database we meaning either Oracle people but more precisely Oracle software does all of that so there's a couple a couple of ways of the of the way we deliver this and again this is the the biggest the biggest issue but what autonomy means autonomy means no human labor driverless cars no human labor no human error no human cost okay I'm pressing the button there we go all right I've said this I've said this ready the continuous adaptive performance tuning this is actually a very very big deal anyone anyone who's a professional DBA knows that periodically you have to go back and tune these data warehouses people the shape of the data changes the size of the data you know the amount of data changes the kind of queries people are asking change the kind of analytics people are doing change these applications evolve over time and they put different strains on the system so how you tuned it you know three months after you built the applications very different than how it needs to be tuned three years after the application has been running we do it continuously it saves you money okay next slide please yeah only one of the two would work okay uh I mentioned I mentioned this earlier this is a dynamic elastic Hardware allocation this is a very big deal in Amazon if you want to make redshift bigger add another computer add a you know and you know another osa not another processor more compute more storage shut the system down move the data you can be down for days here throw a button now you know a full disclosure right now our elastic Hardware allocation we actually have left it in your hands if you want to expand the system you throw a switch it will immediately add more computer immediately add more storage we do will do that for you automatically I think starting in this summer but right now for the first version of the autonomous data warehouse which is out right now it has all this elastic capability you can add compute you can add storage but you have to throw switch we will do that we will do that by policy sometimes by the summer but again this is very different even though you throw the switch there's no downtime just we're just waiting for you to tell you want more performance you want more capacity you want more storage and we immediately in a add capacity without any downtime so the irony is the Amazon Elastic Compute cloud is not elastic but the Oracle cloud is and whether this is this and this is very very important because the whole idea of you know the great thing our Amazon says is you know you kind of get what you pay for you get what you use well that's just not true if you have fixed database configured configurations and you have to define your configuration once you know and then that's what you pay for you can't make the configuration larger for at the end of the month for peak loads and make it smaller in the middle of the month or less load you don't get what you use you get what you were allocate you get you have to you have to scale it for you know for maximum maximum usage because it's so expensive and time-consuming to take the system down and add capacity hours you actually can if there's a lot of demand not capacity there's less demand reduce capacity just throw a switch in our case it really it is true you pay for what you use that is not true at Amazon all right there is a lot of stuff built into the Oracle 18c database that allows us to do some things which much must seem like magic because no one else does them we have a fault-tolerant configuration that means if a computer fails the computer fails the database keeps running we have this thing called real application clusters and that's that's built into Exadata that is if you have a fault tolerant configuration if you buy policy have said I want to minimize downtime we give you a fault-tolerant configuration that means you have multiple processors accessing the same database and one of those processors should fail you might slow down a little bit but the system keeps running the system never comes to a stop that's simply not true of redshift or Aurora or even Oracle at Amazon you run Oracle at Amazon you don't get that feature you don't get the real application clustering feature they can't they don't support it you don't get that you don't get that fault-tolerant configuration they don't have fault tolerant configurations for database at Amazon we also have this thing called active Data Guard so if the regional data data center should fail we immediately failover to another data center automatically with no interruption they don't do that I've mentioned earlier that we have the ability to patch the database while it's running well how could that possibly be how can you change that how can you change tires at 140 miles an hour well I didn't again full disclosure if you're driving 140 miles an hour and you decide you need to change a tire you got to slow to 130 because let's say you have eight processors processing this database load and what we do is we take one of them offline so all you have is seven so you end up with 7/8 the performance you had before you slow down a little bit not that anyone would notice and you take this one computer down and you patch this off run that one computer and you bring it back up then you bring the neck yeah you do that over no yeah the computer one gets patched comes back online computer two goats that gets goes offline it gets patched comes back online it's all based on this idea that we have multiple non shared memory computers running the same database we can detach one of them upgrade this offer and bring it back online no one else has any of this stuff can't do it they don't have fault tolerance they don't have rolling upgrades they don't they can't patch without downtime they have don't have automated patching none of it with any of them not even not even when you're running Oracle at Amazon do you get those features you certainly don't get them with redshift and Aurora you can even go from Oracle 18 C to the sometime in the future delivered Oracle 19 or 20 and you can actually do those upgrades without downtime even in the case of a malicious user deleting lots of other users from this system not a problem we have these things hopefully this thing called flashback where we can go back in time to before that user deleted all of that information maliciously and we can recover it all without taking the system down okay guaranteed 99.995 reliability we will put that in our contract we will put it in our contract and we will pay you if we miss and there are no exceptions this isn't 99.5 but not counting when you have to take the system down for Pacha well you don't have to count everything very simple what are the exceptions what are the what are the down what's the down time that's not counted in that calculation all down time is counted in that calculations Amazon is very different Amazon has a long list of things they don't count Amazon claims have 99% availability but you can't count down time for patching that's not part of the 99% calculation what about 400 not progress well what if you need to add more store no no that's down time for storage that's don time for configuration change we don't count that as part of the 99% what what if Amazon is doing software is soft its own software upgrade no no no that's not counted as part of the 99% calculation what is counted nothing so I mean it's it's pretty funny when you when you read this list anyway uh the autonomous data warehouse is now available to all of our customers fully automated highly efficient we guarantee you no human labor for database administration elastic configurations and half the bring in your Amazon bill will cut it in half guaranteed ninety-nine point and dramatically more a lot reliable and secure than Amazon all right so that brings me to a serious it's actually this is not this is not the demo the demo someone else is going to do this this is kind of a bunch of demos I did at Oracle OpenWorld but I just want to recap to show you how we compare with Amazon when you're when you're kind of running Oracle at Amazon versus Oracle I'll do this very fast versus Oracle in the Oracle public cloud the autonomous database in the Oracle public cloud all right let me just go to the next slide oh yeah the cost this is a very low cost service so the great thing about the Oracle autonomous database is you can use our customers can use their existing licenses you don't have to buy new Oracle licenses you can use your existing Enterprise Edition licenses and bring them to our cloud and we use those licenses are the same licenses you bought for on-premise use you get to reuse those licenses in our cloud and you can reuse your Enterprise Edition license you can reuse your RAC license so you can reuse your active Data Guard license you can reuse all of that stuff no additional cost the only thing you pay for you if you bring your own license the only thing you pay for is the infrastructure the compute capacity and the storage and that's it and that's the cost of compute and the cost of storage the incremental cost when you bring your own license and fort you can get started for 272 dollars a month so it's we think it's a very attractively priced you know service considering it's fully autonomous very reliable very secure okay I mentioned earlier we have a whole bunch of autonomous services you know scheduled for this year the autonomous data warehouse is the first one it's out now we'll have the autonomous version of our OLTP database kind of the other half of oracle data warehousing an OLTP we'll have that out in June so we'll have the two main Oracle services out by June on top of that we have other data management services the Express database and the no sequel and the no sequel database but it's not we're not stopping there we're building a complete layer of services our autonomous analytics service is coming out this summer so when you pair you have an autonomous database you then pair that up with autonomous analytics and visualization we have an autonomous mobility service for mobile applications we have an autonomous developer service we have an autonomous integration service if you will and so Marc Benioff just spent a lot of money from you'll soft to get mule soft integration services good company nothing wrong with it but it's kind of a last generation company these are not autonomous service services that mule soft offers it's kind of conventional old-fashioned way of plugging things together things in the public cloud two things on-premise they can do all that we can do all that but our new version of integration services are autonomous there are autonomous past services this is a whole new layer to the cloud of highly intelligent services for data management application developers including mobile application developers including development of analytics including the integration between your modern cloud systems and your existing on-premise systems all of that all of that is now going to be available in the summer as autonomous has services this is how we plan to compete with Amazon not yes where we have infrastructure as a service we think we have better infrastructure pasture infrastructure more robust infrastructure blah blah blah but what's highly differentiated which is fundamentally different from what they do are these these autonomous PA's services it's not they don't have anything like that mule soft doesn't have anything like that this is a unique no one has any anything like the autonomous data warehouse or the autonomous olt oracle OLTP oracle database and remember we've been working on this problem this database problem for 30 years we're gonna mess it up for a very very long time so this so our database it's an in-memory database it's an object database it's a relational database it's an IOT database it does all of these things and now it does them it does them reliably it's a fault tolerant database and now it's a highly secure database and now it's an autonomous database so this is highly differentiated technology all right I think I've said all of this automated upgrade patch tune minimize the computer resources we use therefore it's very very low cost very reliable very secure all right I'm gonna just skip through this or we'll be here for dinner ok this this is oracle at oracle vs oracle at amazon that's my first set of slides and my second set of set of slides will be oracle at oracle vs redshift analyst amazon's data warehouse at amazon all right this is just saying how bad the Amazon offering is it's not automated you know it's not efficient you know as you say it's not elastic you know you can you know you you you you have to have a fixed configuration you max out your fixed configuration you pay for it when you don't use it they don't compress the data blah blah blah it's not reliable it's not fault tolerant it's labor-intensive there are lots of human errors very expensive all right this is the thing this is a partial list of Amazon claims 99% reliability except for all this stuff in doesn't include downtime for computer capacity doesn't include downtime for maintenance doesn't do downtime for upgrades and patching blah blah blah okay all right so we matched up these configurations we get we came up with the best Oracle configuration we could come up with and we also came up with the most cost of cost performant configuration of Amazon that we could come up with we published all this all of these benchmarks so and we've invited people to try to do better at Amazon if anyone who does better in Amazon we said we will just publish your numbers were will reproduce your numbers and we'll publish your numbers rather than ours so these are the this is the current state of the numbers and these numbers have been out there since the fall of last year and they've moved a little bit but not much all right we looked at a variety of different applications financial market marketing insurance and retail all analytics okay the first demo it was was the Oracle database again in Oracle versus the Oracle database in Amazon all right financial workload so here's the slide we have it now we're gonna start up this was a little meteor I had on this tried couple computers on screen on the stage with me and we're running these benchmarks and you see after 20 seconds Oracle's you know more than halfway done same exact workload same exact database same everything we're halfway done you know they're not not even ten percent done I'll keep going fast and by the time we finished by the time we finished we did in 34 seconds Amazon did it in two hundred and fifty five seconds same exact same exact thing took us four cents cost four cents at Oracle cost twenty three cents at Amazon same exact workload and remember we're ninety nine point and reliable we're more secure we don't have any labor for for database administration they have all of that but forget that this is just your Amazon compute and storage bill just the infrastructure costs there are almost six times more expensive just for the infrastructure another one this is a retail workload and to go through this quickly after 20 seconds we're 70% done still not 10% done by the time we finish in 35 seconds again but you know almost six times cheaper than Amazon same cost per minute same cost per minute but six to one so if you don't want the 99% 99.995% reliability if you don't want the automation of the DBA functions go to Amazon but you have to pay extra if you really if you really don't want the reliability it's gonna cost you more all right Insurance workload surprise-surprise same thing after 20 seconds we're well ahead of them for going about four times faster by the time we're done Amazon is eight times more than eight times more expensive in this particular application but it was it was test after test after test we invite you we invite all of our customers take an Oracle to a database to Amazon try it please try it you can do it it's gonna cost you a lot more all right okay so or so running Oracle on Amazon is much more expensive right than running Oracle in the public cloud but that's not taking into account the other benefits of labor savings error reduction reliability improvements and all these other things security improvements the next the next one is looking at Amazon this is not the Oracle database at Amazon but Amazon's premier data warehouse called redshift this is Amazon's best of the best this is their own stuff well it's not really their own stuff it's that open source database that they close sourced they took it from open source and put it on Amazon and then they said it's not open source anymore it's now redshift all right so here we go same financial analysis workload we did before this is on redshift we see after you know 15 seconds were more than half done Amazon's not 10% done we go on we go on nine times more expensive to run redshift on this able this is the same exact data the same exact queries same workload then I go on here's the next one okay more than 15 times slower to do the same thing well we said one you know Oracle is always bragging about performance always Oracle's always bragging about performance great you have a fast database nice but it's really expensive we know that you can't fool us we know Oracle charges a fortune for its database you know you can't fool us yeah you're fast but you're expensive actually there are 15 times more expensive than we are because there are 15 times slower and they charge by the minute and we charge about a minute so this fast thing translates one-to-one to cost savings it's a big deal all right here's another one market you know marketing marketing data same thing after 23 seconds you know we're you know we're done they're still working finally in the end this time they're over over 13 times more expensive again we did so many of these tests these are a handful of the tests I'm gonna stop I'm gonna stop right here and just say that that we invite you all of our customers who have workloads at Amazon and familiar with em is going to go ahead and try try a database workloads where I want at Amazon try one in the Oracle public cloud measure to things which is easier to use and which costs more which is more reliable just try it and it's not close it's not close we are you know there are 10 times mark there be there between 5 and 10 times sometimes 15 times more expensive than we are we fully automate the DBA function we give you 99.5% reliability and but if you want all of that cat we give you that automated automated intrusion and detection automated patching we're never again will you have to take downtime or wait for a downtime window to click to close a security vulnerability like you do have to you have to wait at Amazon never again that will be automated away for you you get all of that but you have to be willing to pay less with that I'd like to turn it over to the guys who are going to be doing the demonstration and and we're gonna have a customer panel after that so please stay with us thank you very much you you

Channel: Oracle

Views: 33,315

Keywords: announcement, Autonomous, Database, oracle autonomous database cloud, launch, self-driving, Larry Ellison, autonomous data warehouse, Autonomous Database, database, AI, larry, oracle autonomous database, oracle AI, yt, oracle autonomous data warehouse, 30530190, artificial intelligence, retire-2018, self-securing, larry ellison, autonomous database, EXT-18-83, self-repairing, fy17, oracle autonomous, data warehouse, fy18, English, database cloud, autonomous database cloud

Id: EgHkLQsg6sg

Length: 60min 29sec (3629 seconds)

Published: Wed Mar 28 2018