 So, a very special guest with us today, Mr Sanjay Manwani, he has been managing and working closely with open source database technologies for the past 12 years and is currently managing the MySQL India team, which is involved in all aspects of SQL development, not only that, but he's also a speaker at multiple events, so without further ado, let's welcome Mr Sanjay Manwani, Mr Sanjay, over to you. Mr Sanjay, I think your mic is muted. Mr Sanjay, your mic, your mic is still muted. I'm so sorry. I'm so sorry. Okay, let me start. I'd like to thank you for this opportunity, for Forsyatian, thank you for the introduction. Quickly, let me get started. Not much time and too many slides. So, this is kind of our, it's about six or seven, since we've kind of started our database service on the Oracle Cloud Infrastructure. There are a lot of clouds that have MySQL, but there is this service that we have begun now, and this is a native service, so let me just get started. This is a safe hours statement in the previous presentation I was looking that people were asking, what is this for? This basically is that whatever I'm presenting is just for information purpose. If you make any decision based on this presentation, based on any date, then basically you can't say, Sanjay said this and therefore we did it. And that's what it prevents us from getting sued basically, so that's it. Quick agenda, I'll talk about the market trends, I'll talk about MySQL and its customers, kind of introduction very quick. I will talk about the database service, and I will talk about HeatWave, which is kind of the latest thing that we have implemented, and this is specially required for analytics workloads. This is a very, very new thing, extremely hot and it's available only on the MySQL database service on Oracle Cloud, so this is very, very specific to that. And I will also talk about a little bit about what our customers are saying we have and we are doing and how much do they like our services. So, you know, standard things, data is growing exponentially, I think it's really, really growing much and you know, earlier the people used to have text pages and they started having photographs, now there are videos, everybody wants to search everything, access everything. So we can imagine what data is going to be like with AI and I don't know, we'll have maybe holographic images soon on the web, you can see people in 3D and everything is going to be on the web, imagine how data is going to be, right. And it is moving to the cloud, like a lot of people had their own data centers where they were saving everything or they wanted to have it on their laptops. So we see the trend that, you know, as far as consumers are concerned, people are not keeping the data at home, they're too scared because hard disk crash and things like that. There is a lot of data that is increasing and going into the public cloud and that is the trend is, you know, is only expected to grow, right. In fact, now even now today you can't imagine that, you know, you will keep your pictures in a public cloud, right. But everybody does that back up and you take a picture and then goes into your Google photos or something like that. So 49% of the world's data is going to the right in public clouds. That's, you know, some of the predictions which are being made by IDC. And as far as even databases are concerned, right. Out of the databases, people are going to start using on the cloud. As people start trusting data on the cloud, they're going to say, okay, you know, yes, I'm good. I have my data center, I have my cloud. I have the trust on the cloud that the cloud is going to keep my data safe because that's the biggest thing, right. Why do banks want to have or governments want to have that data in houses because they have a worry that people are going to steal that data. But this trust is so slowly going to increase. There is, I think currently it's a time of hybrid databases where people have some data inside the data center, some data on the cloud or exclusively, you know, one or the other way, depending on, you know, how much they worry or how much trust they have on the cloud. But this trend is going to accelerate. It is going to, you know, move more and more of the databases is going to be on the cloud. So people are going to start using databases on the cloud. And as I said, I think the biggest worry that people have is, okay, is it secure? Is my data secure, right? There are thousands of people that are insiders, former employees, nations, criminals, curiosity. Because everybody wants to look at your data. Everybody wants to basically, you know, take away your data in some way or the other. So whether it's personal data, financial data, you have to basically keep it safe and secure, right. And as I said, security, if you look at, you know, CXO survey, right, what is it that they're worried about? Why, you know, why do they want to kind of move to the cloud? What people have realized is that these large cloud companies, they have, you know, they have a better capability of managing security than anything else. So they want to move to, you know, to the cloud, right. And there is a lot more security there. Obvious, there is cost, scalabilities of management and speed. But I think security is one of the big factors that people are trusting clouds now. And, you know, people still are, I mean, open source is more valuable and important today than it was a long time back. Because, you know, as far as open source is concerned, you can be sure that whether the company that manages, you know, particular open source data is there or not, at least you have, you know, the code of that open source. And there will be other people if the, you know, if there is value in the code itself, it is not important if the company that has the code or has created the code, is it doing well or not doing well, that is not important. But there will be multiple people that can support, you know, this open source thing. So the value of enterprise open source is also accelerating. It allows for better collaboration. It allows for, you know, as I said, survivability. It allows for migration between clouds. We have these many things that, you know, open source brings you. And that's the reason that, you know, conferences like FOS Asia, they are very, very important, right? Because then we talk about, you know, how important open sources. So there is going to be cloud, there is going to be open source, and the databases are moving to this, you know, kind of open source databases are going to move to the cloud. That's kind of the, you know, the thing. And right now what is happening is that a lot of innovation work that is happening, the new things that are happening. It is happening on, you know, open source platform. So if I want to do something new, exciting, interesting, it is sometimes, or many times it started by university students. And they, you know, the cheapest and the easiest way to manage is just, you know, using open source. So there are companies that want to therefore also leverage it. And there are students that want to leverage it. And together, I think there is a huge amount of open source innovation that is happening, and I don't need to say this open source databases are mainstream. People were, this is probably a slide from when people were a little worried about open source databases, but Gartner is saying it today that open source databases are kind of mainstream. Let's just talk about my skill a little bit. So if you look at, you know, DV engines, this is a company that rates databases and it talks about how popular databases are. And there are marks that are given. And I'm sure, you know, if you have any idea about databases, you would have looked at the DV engines. And this is from kind of March 2021 that we've talked about. And if you look at it, my skill is number two. Oracle is not an open source database. It's a proprietary database. My skill is a number one open source database out there. And it is, you know, if you look at the numbers given to it, we are quite far away from, if you look at the top five or six, they're, you know, at 400 and then suddenly, you know, when you get to IBM, it drops to 156. So, you know, kind of one-tenth of the usage of these kinds of databases. As you go below, you know, you see a lot of also churn there, right? Redis is coming up, Elasticsearch is going down, SquareSkill Light is coming up, Microsoft Access is going down and up. So these are things that keep moving around. What kind of is constant is kind of the top four or five that are quite stable and they keep on growing. So, what do you need to think? No, just basically the focus is my skill is number one and we are number one by far. And it is popular. People want to use my skill, you know, they are either using it or they say, okay, they are planning to use it because there's a lot of innovation, there's a lot of change that, you know, is happening and this is, these are kind of surveys from Stack Overflow, JetBrain, both of them in the same thing. And, you know, whether it is social e-commerce, SaaS, finance, fintech, you know, people can say, okay, but we have this kind of company, that kind of company. It doesn't matter what kind of company you are. You know, there are use cases, you know, Facebook, Twitter have huge use cases. New e-commerce companies have kind of the newest thinking that has been implemented using my skill. So this is kind of, you know, this thing that I want to talk about. Generics, let's get to specifics. What is this whole MySQL database service? As I said, you know, if you're managing your own MySQL, you're always worried. Do I have the latest security fix? You have to have a DBA team in place that gets the latest fixes. You need to make sure that security regulatory compliances are there. You know, there's a new patch that has come, security patch. There's a new security library I have to implement it. You know, I have to make sure that, you know, my cloud database and my on-premise database are in sync. Let us suppose I have a, you know, you know, my replicas which are there on different clouds or the other way around. I have a master environment and I have replicas somewhere else. How do I replicate, right, between? Then if there are multiple technologies, let us suppose, you know, I'm interested in Oracle, 70% of Oracle people use MySQL. How do I integrate to that? How do I get technical support? So then you need to work with a cloud that understands Oracle, which understands MySQL and preferably that has developed MySQL. I know there are a lot of implementations of MySQL that are there which have been done by Google, Microsoft, all these kind of companies. But what they're doing is they're taking what we have created, you know, making some changes to it, adding some things to it, deleting some things to it, making sure that it works on the cloud and then giving you that service, right? While it's not true or, you know, this is our strength that it's the team that has built the MySQL server is now bringing you this, you know, service available to you. And it is based on the Enterprise Edition. This is the first time I'm talking about Enterprise Edition on FOSS Asia because now, as far as cloud is concerned, it brings everything together. So whether it is, you know, the latest audit, whether it is the latest security features, whether it is, you know, all kinds of security, security key infrastructures that we integrate with, everything is today available to you because the Enterprise Edition is available to you. Again, the database service is easy because it's fully managed by us. So we will have the screens and if there's anything that goes wrong, we will take care of it. The moment you click, it is automatically provisioned and you get the latest version, right? It is 100% secure. We are guaranteeing it. There is data protection. There is advanced security and the security updates or something that's our problem. Because it's implemented in our cloud, we take care of all that, right? And it is, as I said, built on Enterprise Edition and it is on-premise compatible. So it's the same code that works both and it's built on the second generation of Oracle Infrastructure. I will talk about that a little bit. So what is it? So it's not only the database that we are taking care of. The OS is something that we are taking care of. The server we are taking care of, storage, data center, you don't have to worry about, you know, my networking is going to go down my whatever, you know, air conditioners are going to go down. Everything is taken care by us. So what is this second generation cloud? The second generation cloud, so we tried, Oracle tried a cloud first and then it realized, okay, there are certain changes that are required. We need to focus on security. You know, way better performance. It has to be cheaper and there has to be a complete open ecosystem that we need to create. So this is a second generation cloud that has been created and we are not talking about it, but if you read anywhere, all the analysts are saying it is the most secure cloud that is out there. And it's not that it's just the expensive cloud. It is also the cheapest cloud. So if you look at just how much does it cost you for 100 OCPUs and one terabyte of storage for my scale if we have the cheapest out there. I know a lot of people are, these are numbers that you need to look about, but we are way, way cheaper and we are not only cheaper as far as the my scale service is concerned. We are cheaper as far as storage is concerned, the backup is concerned. When you bring in data, when you take out data, all these things are way, way cheaper on the Oracle cloud and therefore basically your TCO comes really down. Let's talk about heat wave a little bit because that's our latest, newest and most interesting thing. So typically if you wanted to do analytics what you would do is you will have your data in something like a my scale or a front-end database then you will clean it up because there will be customer specific data. There will be some credit card numbers, health information. You'll clean it all up. You'll put it into an OLAP database and from that OLAP database then you're going to basically create your data cubes and then using those data cubes you're going to take out whatever information that you wanted with these huge reports and things like that and a part of that information is going to go back into your database because there is a single view that people want to have at their data. So there's a whole cycle and some of these cycles will take days, some of these cycles will take months. Definitely you can't do analysis in a very short time. It takes a long, long time. So for example when companies want to bring out their own reports so it takes a long, long time for bringing out their quarterly reports with HeatWare which is our kind of OLAP accelerator or OLAP extension if you can say. It is the same database that you're working on. So now there's a single MySQL database you can do OLTP and you can do analytics on it because as far as HeatWare is concerned it's like another engine and it is an in-memory engine built in and the moment you make any changes and I have a nice diagram for it I think that's good I'll go back to the diagram I will basically go back to the diagram So you're all the data and everything that is working works exactly the same way and as far as your OLAP queries are concerned suddenly as far as the user is concerned it is suddenly very, very fast it becomes really like 400x faster and if you look at the scaling earlier you were scaling to maybe 100 cores now suddenly you are scaling to 1000s of course so you have 1000 parallel threads which are working this is an absolutely new technology so basically your Oracle Labs had come out with this they had their own specialized hardware on how do you parallelize OLAPs and things like that and they built in that technology and they said ok who do we integrate with and they came to MySQL and MySQL said yes we would love because again it's a very, very complimentary thing we are very, very good with OLTP but we are not good with OLAP so why don't you bring this engine in and they integrated this engine just to show you how this works yes so basically do an insert it goes to our optimizer and from our optimizer it knows that ok this table particularly is a table which is required for my OLAP and obviously everything is in anodb so that's the source of the truth but the whole thing goes into your in-memory database which is Hitware and this is happening immediately it's not there is no time gap here that ETL has to happen Keenup has to happen nothing has to happen if you do want to make sure that there are certain Keenups happen you write the filters out right there and this information is going to go into your in-memory presentation Heatwave database there which is kind of thing and from that when you ask for a result the result comes and it basically query is going to get answered at the moment so there is no wait time here there is no cleaning here there is nothing that is happening everything is happening extremely extremely fast and it is happening extremely seamlessly as far as you are concerned as far as user is concerned they are giving a MySQL query and suddenly this MySQL query is fast right ok so it is completely fast and because it's a part of Oracle infrastructure anything that you know is built in so for example we have this you know beautiful reporting tool in which you can create pie charts you can create you know all kinds of waves and graphs and you know all those things so all those things are available to you automatically because any tool that is built for analytics that is going to directly work on top of MySQL and Heatwave right and you can have your dual TP applications pumping in the data to MySQL you can just add the queries of your OLAP and those queries will get answered and beautiful reports can be run using our analytics tools again as I said if you are talking about 400 GB and 64 cores this is the speed up that we are talking about right if MySQL query was taking 1700 seconds it comes down to 4.21 seconds it's really really fast you know they will basically you know are very fast and the queries that we are talking about we are talking about the TPCH queries right so we have looked at the TPCH queries and you said okay you know in general what is the speed of execute we are talking 64 cores here please remember right and obviously you know people are going to say okay what does this mean if we compare with Amazon Aurora because lot of people are working currently with Amazon Aurora so how much faster is heat wave if we compare it to Amazon Aurora so if you look at it it is about 30 times faster in certain cases okay from 5 days and these are you know remember first you have to do is you have to in any scenario you have to like these are terabytes and terabytes so 4 terabyte of data how much do I have what time does it take to load in the first place right in case of Aurora 4 terabyte data might take about 5 days to load because it is slow it is cleaning up cleaning up and everything is going to happen so our data preparation time is 4 hours right how much is the you know is the run time the run time is 1100 times faster so a query that would take 2.5 hours in Aurora is going to take you 8 seconds in heat wave right and is it how much cheaper is it it is way way cheaper so it is 3.5 times cheaper so you are preparing it quicker you are running it faster and you are paying less money so very very you know makes a lot of sense for for this right and people are going to say but you know why are you comparing with Aurora Aurora is not for OLAP it is heat wave that we use for you know you know it is basically redshift we use for our OLAP so why don't we compare with you know with redshift so fine let's compare with redshift here if you have our analytics queries are about 2.7 5 times faster and it is 3 times cheaper so again you know from the cost benefit just makes sense to run it on heat wave and again remember it is seamless it is a single you know screen that you are running everything is the same but then people say okay you know why are you comparing with this 110 $1000 thing I can get way cheaper Amazon redshift so you know how about you know how do why don't you compare redshift fastest shape so if it's a fastest shape we are 5 times faster and we are 37% cheaper if you look at okay it is the cheapest shape of Amazon redshift then we are 17 times faster and still we are you know 3% cheaper so you know I don't want to you know I can I can keep talking about my service and I can keep saying it's the best service out there but let's talk about what our customers are talking about what do they think what is it that they've seen so we had you know we took it to health engineers even health engineers in health engineers and they tried you know and nice skill and they say we discovered astonishing performance improvements from hours to seconds exactly what I'm talking about and remember this is not something that has been built in a day our labs team has taken years and years to develop this right and this is our you know very very kind of it's like you know those diamonds that are very precious difficult to come by but when once implemented you know these are some things that you can really enjoy and it comes very very easy for them because single seamless database they have one database and from that one database as far as analytics is concerned it happens seamlessly so what is mercury saying and remember these are real time cases so as far as our test cases are concerned we might say 400 times 1100 but there will be some averages that will become common and as far as customers are concerned they are saying ok we are at least seen 20 times faster analytics right and our acceleration is 45,000 faster than on-premise so if we have MySQL which they were meant to mercury was maintaining and they tried to you know compare it with happening on heat wave it's 45,000 faster times faster so not something that I say again something that our customers are saying what is Enix saying Square Enix saying is they wanted to have an analytics workload of social game infrastructure and their focus was their problem as far as ETL is concerned so when you move your workload from your standard database which you use for day to day LTP and then you want to do analytics like you know moving mountains you take one data you clean it up and you know all that cleaning you used to take them many-many you know difficult times so they are saying our analytics engine is 500 faster and we require 0 ETL so the ETL that again you know what I just told you this is exactly what our customers are also saying again SCSK cooperation we attend analytics engine is 10 times faster and no need for ETL so just you know repeating things that people are talking about right now I think if you are wanting to test this out and you know kick the tires of it you get about $300 credit so what I would say is just go down search for MySQL search for the service of Oracle Cloud try it out try your hands try your fingers and this is the latest newest thing in the fastest thing and it's really really state of the art so if there is state of the art analytics that you want to think about if there is state of the art scaling that you want to think about because this thing scales from you know 1 core to thousands of cores and as far as the scaling is concerned it is at about 0.8 so it's not 1 is 1 so if you have 1 core it's going to give you 1 performance if it's 2 cores it's going to give you 2, 3, 4, 2,000 but at least we have 0.8 so 1,000 cores are going to give you 800 or 8,000 820 times performance so it's not really 1,000 it's kind of quite linear so try it out there is there are free coupons that are available I think for India there is a $500 coupon that is there if you are from India from I think other parts of the country there is 300 new try it out for free and we are in the process of developing it we are in the process of you know implementing it not about I think 2,000 customers we have tried it again in the last 7 months that we have come out with we are getting extremely good results and you know we are getting very very positive things so I think if you are serious about analytics and cloud and you are interested in my scale this is the thing that you need to get your hand at youth go down there use your $300 and get it out is what I would you know what I would suggest again DMM cooperation and again you know everybody is talking about no changes and I think that is all that I have if there are questions I will be kind of happy to answer those questions and I will stop scaredessness you know my screen share here Hello Mr Sanjay thank you so much for your insightful presentation now let's move on to our Q&A segment so does anyone have any questions for Mr Sanjay I think currently there is no questions in the shared notes but if you would like you can voice out your questions and or type in the public chat now we have we have a very shy and quiet guy in the Steinforces share I think okay there's a question coming from Gaurav Soni he is wondering how Oracle can improve the performance compared to other cloud services so one of two things one is remember if you are using any other cloud service from any of the provider it may be a little different and therefore you might not get latest service there might be certain things that they might do faster because they have re-implemented certain things for it so I know that AWS has you know some things that are faster and slower but what I recommend to people always is that when you are talking about any cloud implementation and specially as I said in open source the advantage of open source is that whether I want to use MySQL on cloud number one or cloud number two or cloud number three I should be able to take it and I should be able to move it around as far as the core MySQL is concerned so are you getting locked in into a cloud that is something that you need to worry about with MySQL you know that this is the official implementation of MySQL and this is the official implementation which is available to you on the Oracle Cloud if somebody breaks it up extended and this is the same for any fork once they change things then when we come out with a newer version they have to make their changes most compatible and it is exactly the same with other cloud renders so they will have to create a new extension they will have to change something and then that extension will have to go into their into their cloud service if you look at certain providers they have not even right now we are on 8023 so there are even in 8.0 there are 23 version that have come and whenever we come out with a new version of 80 we also come up with a new version of or 5.7 and I think 5.7 or 80 or some number is going on so are they able to keep up that is something that you need to understand and remember tomorrow if you want to move from Google cloud to Oracle cloud because Oracle cloud is cheaper and tomorrow AWS becomes cheaper than us are you able to move your changes from one cloud to the other or if you want to just bring it back on premise so are you going to be able to do that and let us suppose you say I don't want to take my services from Oracle there is another provider that I want to take it from which supports core MySQL is that cloud provider going to be or even a company going to be able to help you on premise because this is an open source product that is something that you need to look at not only just performance A versus B I am going to talk about performance I am going to say this is the best another person is going to say this is the best remember as far as Heathrow and analytics is concerned because specific hardware wise and implementation wise currently it is offered only on the so if you are comparing analytics we are way faster than anybody because we understand the code we understand so you can't compare nobody has it Redshift I already compare we are way faster than that but core MySQL again I think there are there are certain things that certain people have done that are major faster but if you look at the version that they are supporting they are supporting a very very low version so again your call and don't don't believe me try on benchmarks as you saw we talked about 400x speed but when we took it to customer it's 20x faster for me so every implementation every difference is going to be different or it depends on your queries that you are using it depends on what exactly you are trying to work on yes typically I would say we have the latest we have the fastest but don't take my word for it as I said there are for 30 days $300 why don't you and you know how to run MySQL just take it go there try it out and I promise you everything installs in one click you go there start MySQL database on Oracle Cloud you say create it if you want to get created in 20 minutes in 20 minutes it will get created create your tables run your queries there is a due to face the previous presentation talk everything is going to work and just try it out use your $300 and you tell me how fast I see thank you so much Mr Sanjay he would like to follow up with say sure looking forward to try out the services with free coupons so audiences I think Mr Ivan has send 3 links on Oracle and MySQL you may want to check those out in the meantime does anyone want to voice out their questions we still have 6 minutes till the end of Q&A or Mr Sanjay do you have anything else yang share with the audiences so if you look at yes there are a lot of conferences that are happening on heatwave specifically and these are hosted by senior vice president who was in charge of the labs team and his name is Nipun Agarwal so search for Nipun Agarwal and I think on 24th there is an open conference that you can attend and Nipun is actually going to because it was introduction I just talked about very generic stuff about money and benchmarks and things like that but if you want to look at the implementation it's a very cool piece of technology how the nodes are there how the strut is there so if you want to attend any of these conferences search for our open conference you'll look for Nipun Agarwal he's the guy who was heading this whole thing so look for these conferences and go ahead and basically I don't know if this is what I would suggest Thank you Mr Sanjay okay we're down to last 4 minutes it's the last chance for you all to voice out does anyone want to unmute yourself and talk to Mr Sanjay or we can open up a discussion then we can share more about the open source about my SKL or Oracle