Learn Live - Azure SQL Fundamentals: Introduction to Azure SQL (Ep. 1)

[Music] hi i'm anna hoffman and welcome to the first ever episode of learn live we're super excited to be joining you here today uh live from wherever you are from wherever you are streaming today uh my name is anna hoffman i'm a data applied scientist on the azure sql team and today i'm joined by bob ward bob why don't you introduce yourself thanks adam i am bob ward i'm a principal architect in azure data team famously now known as the person that presents with anna hoffman that's what i'm known as so i'm here today with my partner in crime anna to talk some azure sql pretty excited about it thanks anna for having me yeah i mean it's my pleasure and we're so excited this is gonna be such a fun series it's been a while since we've done this series so i'd love to like take a step back today of course is introduction to azure sequel but there's actually much more than this to come bob can you tell our viewers a little bit about like what we're doing here how it happened and how it's kind of evolved yeah sure i mean in a january of 2020 when was that that seems like an age ago you know we were joining forces together i'd spent time launching sql 19 you spent a lot of time working with azure sql i started on this journey uh you know here's a quick tip don't ask your boss what can i do next not a good idea because your boss will tell you here's what you can do next and our boss assad khan said bob i think you should spend some time in the clouds even though i had spent some time in azure take your experience with sql server and start assessing azure sql compared to sql server because a lot of people know who you are they know you know sql server they trust your judgment they trust your knowledge so that's that's what it was about i was like hey i'll go assess sql server and you kind of said hey i'm kind of working on this azure sql project it'd be kind of fun if you joined forces so that's kind of what led us to this whole concept of azure sql fundamentals yeah definitely and you know when we look at the workshop that we've built you know you'll hear us call it different things the azure sql workshop azure sql fundamentals azure sql bootcamp um but really i think a big part of it was being on top of the microsoft learn infrastructure and you know bob do you want to share what's what's so special about microsoft learn and that's that bottom left link here on the page azure sql fundamentals hey number one thing out of the gate if you think about sql server we always were able to do things for training wise to say hey download developer edition on your laptop and just do our training right but as we started building azure sql we thought to ourself huh how are they going to do azure sql without paying for azure sql to do our training we'd like to offer a free thing for you and having this sandbox environment with the learn team together is huge imagine now anna going to these different modules that we've talked about and trying out what we're going to show you over the next six weeks and doing with not getting charged to your azure subscription so having that sandbox is so big and then microsoft learn just has this amazing reputation for training at microsoft so being able to partner with them together to be on their platform and have that sandbox i think it's just been critical and it's been immensely a part of the success we've had of all this training yeah definitely agree and you kind of alluded to something that i'm not sure we've told everyone yet and that's this is running for six weeks so the next six weeks we're going to be going model through modules so what you see here and we're going to take you through all of azure sql fundamentals and we kind of think at the end of this you might be able to pass dp300 which is our relational databases in azure certification so pretty exciting and you know when we look at these modules bob uh can you tell us like how we came up with with this order or these modules or why these items are what we consider the most important sure i mean when we were doing this assessment i was doing this assessment you're working on azure sql together we're thinking to ourselves how do we organize this and ironically we had both together started working on a what to use when type session at events which led to this idea of like we got to start off in the first module and talk about what is azure sql microsoft had pivoted to this branding of this term azure sql so we need to kind of set the landscape and build the foundation for what are we talking about what's the history of this what is it all about why would you choose this right and then for me uh being a sql server person i thought to myself okay well what is one of the first things you do you have to install sql server and so we together thought well let's talk about the deployment experience how do you deploy azure sql how do you configure it and compare those experiences with a sql server perspective and then i remember throwing this tournament to you we were just talking online uh we used one note remember all the one notes we built to go uh design all this i said anna we gotta have meat and potatoes in there and you're like um meat and potatoes like what the heck what is this texan term you come up with and it all deals with these topics in the middle the heart of this security performance and availability are three critical parts of the sql server database engine so why not take the concepts in sql server that apply to that and tell you an azure sql how does that work and so here we have we have our five modules and we started down the process of you know going through all of this and then i remember one specific meeting you saying well there's got to be a way to put this all together right and i'm like yeah what do you have in mind and so you were the one that came with the idea of module six putting it all together so you should tell everybody you know when when we get to the end of six weeks what are you going gonna have yeah it's really exciting at the end of the six weeks you're gonna be able to like it says put it all together we've got some really challenging exciting case studies that are gonna bring it all together and we're going to use like a social media campaign type of style to use hashtags and try to get you all to submit your solutions to the challenges that we pose at the end of all this so i think it's going to be really fun you know i want to bring us back to today where we're going to be focusing on the first module which is introduction to azure sql so if you go to that aka.mslink you're going to be able to find the actual module we're going through today so we encourage you after this session or next week or whatever to go through this actual module so you can get a little bit into the details and and but understand we're going to be covering some of the same content and finally um this is live and interactive so want to encourage you to participate ask questions on whatever streaming platform you're coming in through uh we will be taking a look at what's coming up and what's being brought up to us uh on youtube we're getting a lot of comments already some people are saying you know hey this is exciting some people are saying thanks thanks for doing uh this so we've got a lot of engagement already so i'm just super excited for all of you to join us today and let's go ahead and talk about azure sequel how does that sound bob sounds like great idea to me awesome so today we're going to be talking about azure sql we're going to give you some background and history then we're going to talk about each of the deployment options and we're even going to dive into the service tiers compute tiers and all that other good stuff you have to talk about so you have to think about so like bob mentioned this is really your what to use win type of session before we dive deeper into the meat and potatoes of azure sequel so like to start off with um with this slide here on the left you have ray ozzy was the chief architect chief software architect at microsoft at the time he came to the professional developers conference in 2008 so this was like microsoft build back then and he got on stage and announced windows azure for the first time and he did so with this slide they're seeing here now this is a really interesting slide to me because what is front and center microsoft sql services so microsoft sql services is how we kind of first made our our our step into to the cloud and the the great thing about this and why i think it's so cool is from the beginning of windows azure when you think about it sql services or sql server has literally been at the center of our cloud so it's evolved a lot since then um but it really started out at the center of what we were trying to do and you know bob over the years i'm sure you've seen a lot of different things when it comes to azure sql and evolution yeah thanks again first of all i want to give a huge kudos to my colleagues anna if you look at the bottom of these slides right here these little dots are called breadcrumbs so you see here we're in the overview section you're going to see where we're headed right as we go along these slides we'll talk about azure vm mi database and comparison and summary so what i love about this series is as we're presenting these in slide decks in slide format every module has that look and feel so you kind of know where we are in this discussion and where we're headed in the rest of the modules so anna you're right i have a little bit of history buff and i actually admittedly have been at microsoft much longer than this timeline shows you but i do remember being at microsoft when all this started and you think back to 2006 we embark on these interesting projects at the same time red dog being the windows azure team and cloud db being the sql team saying how do we do services for things like compute and databases and yes you just mentioned in 2008 at this professional developers conference we announced windows azure and sql data services well sometimes it's about branding and so in 2010 we decided to come together when we actually launched the service this was preview in 2008 to to call it sql azure in fact many people today in the engineering team if you talk to them still use that term sql azure windows azure being a term back then because windows being the top brand at microsoft but notice as you look a little this timeline like even by the time 2012 came along we were really what we were doing is servicing a database and you're going to see the fruits of that in today's presentation of what is azure sql database so sql azure became azure sql database now in 2013 there's a very important thing we did and this interesting the timeline is because it kind of aligns when satya nadella came along and was in charge of our server and tools business and that is prior to this time we did not have a virtual machine offering that was stateful you couldn't run sql server in the azure virtual machine environment so the launch of this was very important for the progress of azure sql because now sql became this primary workload because you had a disk drive all of a sudden to store sql server data and then you see in 2014 we decided to rebrand windows azure to microsoft azure now why would we do that because all of a sudden now azure became this platform not just for running like windows and sql but we're running all sorts of different types of compute and applications hundreds of types of apps we're now showing up in the azure platform so why not call it microsoft azure because it wasn't just about windows anymore including things like linux vms right in 2015 we had a very very important evolution to the platform of azure sql database we call that version 12. and that involved a new paradigm and an architecture that allowed us to get a lot more predictable performance and isolation to make sure customers that want to deploy azure sql database i could do that in a very very performant way and then kind of look as we we marched further here at 16 we introduced something called elastic pools which we'll talk about we introduced something called a managed instance project cloud lifter so instead of just a database you're going to get now a pass service that's a sql instance and we're going to talk about what is the difference today between an azure sql database and managed instance and it says in 19 and 20 are all about us kind of just moving the game forward with things like hyperscale or serverless compute or even things like instance pools and finally as we lost all of this last year this branding that azure sql encompasses this family of infrastructures a service and platform as a service so so anna it's a it's a really long history it's important to kind of see that evolution of what we've learned along the way but perhaps the best way to really give everybody a feeling of what are we talking about is for you to talk about what is azure sql yeah definitely and i i think it's actually a good segue into the azure ecosystem we did get one question um and and this person says how is being an azure sql dva different from being a sql server dba and i think this might be a good lead into the azure ecosystem yeah thank you for saying that because you're right before we talk about azure sql what the heck is azure what's that experience like so if you're a sql server dba and you've never used azure before at all you kind of have to have an account in this service right you need to have an account a subscription to do things like deploy databases or instances and often what are you going to use you're going can use the azure portal which comes in the forms of browsers or even applications on phone devices to go in and navigate things in a graphical way you're going to see examples throughout all of our modules about the use of the azure portal to do various things now the azure marketplace is a series of products and services that you can subscribe to and purchase for azure well guess what one of the marketplace items is azure sql and we'll actually talk i think today at the end of this kind of this unified experience we've built on how you can go find out azure sql services through the azure portal through the marketplace now if you're kind of a programming person like me you'd like to know apis and command line interfaces so azure comes with a consistent set of command line interfaces and apis that are consistent across many different platforms and available to you things like azure cli the as cli which is available on linux windows and macbooks or powershell or even just using apis from a programming point of view and so a lot of things are always based on apis and cli and you're going to find out that azure sql integrates and snaps into that ecosystem where we have specialized apis and cli specific to azure sql you're going to see this term called resource groups a lot in our presentations and so behind the scenes the azure resource manager or arm provides an infrastructure for services like azure sql to deploy and manage different various pieces and so having that resource manager is critical and usually what you see is something called a resource group and then resources within that group as they interface to what arm is all about now azure monitor is a platform for monitoring and azure sql snaps into that you will see demonstrations and examples throughout our modules about how we take azure sql and integrate with the overall azure monitor system to look at things like metrics and logs going on with your deployment one of the decisions you're going to be making is where to deploy azure sql in your data center today you have known areas where you go put a sql instance in a virtual machine azure is all over the world there's so many regions data centers i've lost count so you're going to make some decisions where am i going to deploy my azure sql resources in given data centers and regions on the world some of those decisions could be based on compliance regulations that you must adhere to in your country and we'll talk some about that and then finally you would like to use this platform for a reason and that is you would like compliance you'd like trust and most of all you'd like a service level agreement wouldn't it be nice to go deploy an azure sql database and have guarantees of uptime based on agreement and a contract between what you have between us between yourself and microsoft so you know if you've never used azure sql before and your dba you need to understand these different parts of the ecosystem because they're going to come important to understand how you are going to deploy manage and configure a sequel outside of the sql environment itself and we call these things you know the azure ecosystem yeah definitely this is this is definitely like things that become you know common language as you start to work with azure now when we think about azure sql and we really want to get into this now i know you guys have been watching for a while you want to know what is azure sql what are the options and what scenarios do i use one versus the other so there are three main deployments in azure sql when we think about it from deploying things that run and are hosted and live in azure now the first one is sql server on an azure virtual machine and as we go from left to right through the slide what you're going to see is yes you give up some control and you give some more responsibility to azure but what you get in return is also pretty cool so let's start with sql server and azure virtual machine so in this case it's just sql server running on a virtual machine so you pick the os you pick the version you pick everything that this is running on and microsoft takes care of the infrastructure and the hardware so that's sql server and azure virtual machine a lot of times we see people using this uh when they need access to the os but we'll dive deeper into that after this now the next option is what moves us into our platform as a service or pas type of offering so sql server on azure virtual machine is an is or infrastructure as a service whereas paths or platform as a service is slightly different and let's talk about why so the first option is azure sql managed instance and you can think of this as hey let's abstract away the os from you and let's give you a managed sql server that is versionless that's what you can think about with azure sql manage instance because this is going to give you all the instant scoped capabilities that you're used to having in sql server now if you say you know i don't need these instance code capabilities and i want to try some of the newer things that have come to azure sql database like hyperscale and serverless that's when you know we go one level further and say okay let's abstract away the server and say i just want a managed database with the the latest innovations and the greatest tech and the most opportunities for cost savings and that's where azure sql database comes into play and today those are the old that's the only deployment option where you can get hyperscale and serverless so if we take a step back and kind of look at the spectrum there are different ways that might land you into each of these different deployment options and bob you know do you think it'd be useful if we kind of went through some scenarios and talked about some customers and dove a little bit deeper into this or you think this is good well we presented this before it's a great overview to kind of think of left to right on my furthest control my most control my less control my more capabilities but i think what was really interesting is take this from perspective like a challenge solution like what are the challenges that one of these is trying to solve what's the solution and what are the features of differentiators that's probably the right way to look at it awesome yeah so let's do just that and we're going to dive first into sql server on azure virtual machines um so if we think about a customer challenge or a potential customer challenge or scenario i want to migrate to the cloud as fast as possible but maintain my os and sql server functionality so maybe you're really just trying to do this lift and shift you want to you don't want to upgrade you don't want to move to version list yet uh one great example of this is if you're running on sql server 2008 which reached end of support a few years ago now i think um but if you move to a virtual machine you can get extended security updates so this can be one one reason you might decide to move to sql server in an azure virtual machine but then let's kind of drill more to some actual functionality here like what does it really come down to so if you need complete sql server and operating system access that could be a third-party app or something of that nature this is going to be your best bet right you need expansive different versions of sql and operating systems for whatever reason we've actually seen customers that have third party applications that require a specific sql version or operating system version or maybe you need to go to linux specifically or use containers so using vms sql and vms are going to be your best option there could be specific functionality really you know specific to your application like you need file stream or you need a simple recovery model database which you're going to find out later in our modules we don't offer that for past services but there's reasons behind that but for whatever reason you need that and then you might use some of our bi services maybe use analysis services or reporting services and so we actually have specific vms built for that type of functionality but if you did move sql server to azure vm like why would you do this versus just running vm on your own well first of all we talked about the fact it's infrastructure as a service so we're managing all the hardware and the host and so forth but consider like the example and i just talked about free extended security updates for non-supported versions of sql server or how would you like to have some automated backups and security update uh functionality like like it'd be great if we had some of that to combine with this complete access in other words can azure kind of help me along the way whether it's azure backup or excelling my performance with like blob caching for storing azure storage or wouldn't i like to look at a study that actually said there's a return on investment here for allowing azure to host the infrastructure but you control most of what happens with sql server in the operating system yeah definitely and one customer story that we really like to talk about when we think about sql server in a virtual machine is all scripts and the reason i like this customer is because they were able to take about 600 on premises virtual machines and failover using a technology called azure site recovery in a matter of just a few weeks so this included failing over their sql server vms to sql server on azure virtual machine and this was able to be done very quickly so you talk about moving quickly to azure this can be a great way to do it and one of the interesting things is after they moved to sql server and azure virtual machine they started to evaluate some of the capabilities related to managed instance and they started converting or migrating some of their instances to manage instance to take it full advantage of that platform as a service capabilities now you know we don't go into a ton of detail on sql server and azure virtual machine for the rest of this series um so you know bob i know our viewers would love it if you could give a little bit of a deep dive onto sql server and azure virtual machines yeah now we had a tough choice when building this workshop out right if you think about sql server itself and those engine capabilities you know those are going to be common inside the virtual machine of sql server itself but i couldn't help myself i was like you know we have to give the audience some taste of what it's like to do sql server and vm so let's let's spend some time doing that so first of all what about just deploying this thing like what are your choices well we talked about that idea of a marketplace so this is really interesting we will provide pre-installed versions of sql server and windows and linux for various versions out there that you can just choose from the marketplace make a couple of choices hit a button and about 10 minutes or so later a fully installed version of sql and the os will be available for you use that's kind of nice to have right and then you have this ability to say oh i'd like to pay as i go you know a subscription type model and we'll talk about in a second or you can actually bring your own license to the table from an existing sql server license so that's one way of doing it the other way of doing it is to say say i'd like to install an operating system like a windows version i do this all the time from the marketplace and then i will actually bring the sql bits myself i have the bits i can copy up the setup files and just install sql within the deployed virtual machine with the operating system this is an interesting mode that a lot of people realize let's say you got sql 2012 running an environment with windows server 2016. i just picked that combination azure migrate allows you to take the vhd the actual vm image from vmware or hyper-v and literally lift and shift the vm image the snapshot into azure just kick up the virtual machine and configure from there so that's an interesting option if you'd like to deploy sql that way now this is something a lot of people i don't think know enough of it's important so you'd like to take us some advantages in azure like you'd like to be able to use licensing that you already have like azure hybrid benefit as a possibility or you'd like to do these automated backups and security updates or even see your virtual machines with sql in the portal along with other azure sql assets so there's something called the sql is agent extension okay so if you use the marketplace we just put this in there for you and you can start using it if you install your own sql server you go through a process of actually installing and registering with this look at the advantages of that especially from it comes from this licensing point of view or managing your vm so sql it's not a big brother thing it's actually just trying to help you manage the vms you have sql uh already already out there so a lot of people don't realize that that's something certainly to take a look at to use now when it comes to sql a lot of it's about storage and sizes so when your environment you're used to picking virtual machines with certain configurations of memory uh cpus number of cores disks we have all those we have virtual machine sizing that we have we have guidance for you about picking the right vm sizes in azure that match your sql workload and then you have choices to make we're going to help guide you about putting your database and log on premium managed disk on separate disks taking advantage of things like recaching we're going to help you put tempdb on the local ssd drive for maximum performance and then we have something called ultra for super low latency needs for performance we have ultra disks things like your transaction log for say like in memory otp might use something like ultra so a lot of choices there for sizing and your storage needs you're going to have to kind of typically network this together so we offer virtual networks on azure you're going to learn more about virtual networking with azure paths but you should the same concept can apply though for your vm where you can integrate on premises assets and your networks with virtual networks and azure and then azure defender which is a new service that's kind of been out there a little bit for past we offer this to integrate in and apply this to your sql server azure virtual machine environments we'll talk more about azure defender and the security module those concepts like vulnerability assessments and events threat to protection can now be used for vms now high availability still matters when you're in the cloud and we have a lot of options for you but one thing people don't realize when you're in our infrastructure you know have built-in h8 at the vm at the node level you know for failovers that occur we take care of all those things for you because the vm is in azure but you might want to use some of our technologies like failover cluster instance or availability groups and we have that functionality in fact we have specifically built functionality so those can actually work really well in azure and we even have some interesting scenarios we announced last year where hybrid availability secondary replicas can be done like a dr only can be done for free with your licensing so check out those kind of type options as well even things like file stop file snapshot backups or integration with azure backup itself for a complete vm sql combination so if you look down this list you can see there's a lot of choices but a lot of functionality to make your sql and os investment integrate well with that azure ecosystem yeah and i think a lot of people are pretty excited about all these different options because what we're doing is we're really giving you a lot of control and a lot of options to kind of pick how you want your backups to be done how you want your high availability to be managed and that sort of thing and when we kind of look at comparing ias which is what we're just talking about infrastructure as a service or uh sql server and azure virtual machine uh with your more platform as a service pass type of offerings that you get with azure sql database and azure sql managed instance that uh choice of control and some of the configuration capabilities are a big benefit of virtual machines and one way i like to think about it especially when you look at things like business continuity availability your backups and geo-replication with ias all of these things are going to be possible for you to configure bob mentioned uh several of these on the previous slide however with pass or platform as a service or outdoor sql database manage instance um either of those options this is more of a built-in capability for example and we'll talk much more about it in coming modules but some of our architectures just actually you deploy a specific database and we deploy an always-on availability group behind the scenes for you so yes in is it's possible and configurable but a lot of times in past it's just kind of built in and the same goes for things like geo replication but there's even more think about scale so when you pick those vm sizes for is you can change those but in paths we have an advantage we could do things like auto scaling with serverless right things behind the scenes where you just make some declarative choices about ranges of cores and we'll just scale for you what about versionless anna you touched on that concept earlier about versionless this important concept even though we can offer automated updates which are security updates for vms the sql server engine in pass is versionless behind the scenes we are doing everything for you to update that version of sql server this has huge advantages because not only are taking advantage of security updates that we just do for you but if we find an important fix we need to make to the engine we can roll this out proactively to prevent you ever seeing a problem or we can light up new functionality that are versionless through a private or public preview program without installing a new major version of sql server and then all sorts of built-in monitoring and intelligent capabilities can now be delivered because in paths we have the ability to build and install and deploy software services that interact with your paths installation and help really you know provide some intelligent performance and monitoring capabilities you're going to see examples of that throughout this six week series definitely and so now that we've kind of compared is and passed let's go into some of the platform as a service offerings starting with azure sql managed instance now remember this is kind of abstracting away that os and giving you a managed sql server so some customer challenges we see is a customer wants to remove some of the management overhead so they don't want to deal with backups or availability or some of those things they want to take advantage of a lot of the built-in capabilities that we were just mentioning about however they do leverage or some of their systems leverage uh sql server instance scoped capabilities so what do i mean by that i mean things like clr machine learning services service broker uh database mail um dtc i think those are those are some of the big ones but those instant scope features that you think about when you think about scoping to a whole instance or features that are only available at an instance level uh this is where we find customers really landing in azure sql managed instance because it does have support for all of those you know most of the the features in sql server today yeah and that's a great point out and if you think about the actual key features first of all it comes down to the fact that this is an instance we'll talk about something called an instance pool shortly but i want you to think in terms when you connect with say management studio of you looking at an instance of sql server that just looks like your sql server that's when we say the surface area is really really important and you know actually the history behind this project is called cloud lifter i remember in those days when we did that the reason we did that is that customers were struggling a little bit to lift to the cloud and azure sql database deployment because of the features and things you just mentioned but it is a fully managed service things like platform service capabilities that you've mentioned already are going to be part of this thing but another thing that we uh we did cloudline for was native virtual network support so at the time we didn't have some of those cables for azure database and so you want to be able to take your managed instance and integrate it into your virtual network especially for an on-premises environment and manage this just comes with that type of capabilities and then you also want to be able to do things like azure active directory with on-premise identities that's going to be part of the solution as well but there's got to be again what are the differentiators to make this interesting for you besides having that instance notice this one this is interesting uh when you want to do a migration using things like data database migration service we can do almost zero downtime migration using log shipping type technology and we just announced even the ability to have apis for you to control that but yeah we can behind the scenes use log shipping to do a almost zero downtime migration from say a sql server on premises or even an ios type environment and then you're going to see later in the availability section where we're going to introduce something called failover groups which are a big part of doing managed distances across regions and then just like we talked about in vm independent studies have looked at what is going to be that return on investment for you where we're gonna do more of the managing of some of the things behind the scenes you do less managing of it right you don't have to worry as much about it are you gonna get that returning investment enforcer's done a pretty good job of that and then i remember putting the slide together and i thought to myself where really is this it's the best of sql server with the benefits of a managed service that's kind of the summary i think of the way you should think about it yeah i i think it's a great summary and i i've started saying it as well so i i'll give you kudos but i'm also going to say that and a customer we really like to talk about when we talk about manage instance is komatsu because they were able to migrate a pretty sizable database with almost no code changes in just a matter of days a terabyte and a half and now these this aren't just you know marketing but these are real numbers they had 49 cost reduction and 25 to 30 percent performance gain so that's pretty significant that's not just like five to ten percent like 49 cost reduction pretty significant uh we're seeing a lot of adoption of azure sql managed instance uh like bob said best of sql server the benefits of a managed service now the last deployment offering we have to talk about today is azure sql database and there are lots of customer challenges that kind of lands a customer in azure sql database so let's talk about a few of them um so a lot of times we have customers that want to build modern applications so they want to take advantage of things like cloud elasticity and also this idea of serverless and that's something you're only going to get an azure sql database today with hyperscale and serverless which we're going to talk about actually later in this session today as well as upcoming sessions now we also have customers who want to do things like manage multiple databases so perhaps they're a software as a service provider and they have many databases and they want to do things like resource share and cost optimize and that's also something that azure sql database has a pretty good solution that's been around for years to do but there are some other capabilities and features as well and let's dive into them yeah and again think about this your goal is to have a database that's what you want right even though behind the scenes we may dedicate an instance of sql server for you uh and you may have some feeling of that by looking at say a dmv or something the concept is you want a database and that's the abstraction point you'd like to work with and so the idea of a database is important or an elastic pool which is of course a series of databases and then just alluded to but because of the fact that you are going to work just at a database level behind the scenes we can offer some interesting technologies that we can't anywhere else things like you've already alluded to hyperscale and serverless compute those are very important aspects to the promise of azure sql database and you're going to hear us talk about that in today's session and other sessions as well in this in this series it is a fully managed service in fact it's the fullest of managed services again because you're doing a database and you're seeing the security module we're going to help try to solve some of your your concerns about uh you know exposing things on endpoints on the internet by something called private link supporter we're going to offer that in an azure sql database in addition we're going to go to the highest level availability here in this case by being able to offer something called availability zones and so if you look at differentiators that kind of technology allows us to offer the highest availability of any of our azure sql choices here through our sla and quite frankly the highest in the industry that's been published for a service level agreement of that kind of type number that's pretty significant and we'll talk more in availability about the entire sla system and you know what does that really really mean to you but this was interesting uh many of you out there using sql server understand the terms like rpo or recovery point objective or recovery time objective we're the only database out there offering a service level agreement for those kind of type numbers not just for sla and so just like some other studies you know some people have asked me before like this database is interesting but i don't really know are you really performant you know can you really meet the needs of my application so giggom did a study and looked at the performance using benchmarking to compare us to in other industry databases and so that's something you sort of can take a look up if you're kind of wondering like is it possible to perform with azure sql database and what kind of cost am i going to have compared to the competition yeah thanks bob and you know when we think about customers here because we always like to mention a real world customer scenario i like to talk about accuweather now there are a lot of customers we could have talked about but i like accuweather because they came to azure they came to azure sql database and said you know i don't want to worry about managing a sql server managing versions i just want a database and i want to build modern applications so they were able to very quickly get up and running and start building these new applications uh leveraging things like azure machine learning services azure to factory uh and then scale as their service either needed more or needed less compute or other services so it's been really powerful for them to use something like azure sql database and now what we're going to do bob and i know you love this section we're going to do some knowledge check questions i mean there are people right now that are just tuning in going i thought i was just going to sit back and kind of learn from anna and bob you're going to test them oh yeah i mean if you think yeah we're going to do lots of quizzes we got quizzes we got assignments we'll do all sorts of stuff bob yeah that's using that's what you really meant i really want to make bob look bad in front of a lot of people yeah okay uh-huh yeah so i i will say you know if you want to use the official poll then please head over to learn tv that's where the poll is going to be hosted uh otherwise if you're not on learn tv feel free to put in your answer in the chat whatever it is uh wherever you're streaming from we can also see those as well but if you want to be part of the overall uh ticker then you have to go into learn tv uh so that being said let's take a look at this question so the question is you want to migrate to the cloud but you leverage a third-party application which requires access to the os which is going to be the easiest azure sql deployment option to move to okay i'm gonna think about this while you're you're uh getting your answers in um you know anna how i roll process of elimination that's the best way to take multiple choice tests definitely we do have some answers coming in and learn as well as on youtube and twitch so keep it keep it coming in we also have someone on twitch they said bob you can never be bad i don't know what that was in reference to i'm not sure but i'll take it the first one yeah um i think someone said ranger shirt today i'm rocking my baseball team in honor of the baseball season coming up so you are yeah you're just ready for the next four because the past yeah i love the way you said that yes okay what do we got anna uh well we have someone who commented on twitch and i liked this comment they said it depends my favorite answer yeah that's what i should do let's take a look it looks like the most common answer is between a and b a has got about 80 of our votes on learn and for um for learn for youtube and twitch it looks like a is the the most common one as well you know before we answer this question you know it might be helpful i'm not sure if we exactly talked about elastic pools bob do you want to give a brief overview of where elastic pools falls into all this well darth we actually are going to talk a little bit more about pools on the next slide so anna i'm wondering you snuck it in when the next slide is where i get to talk about pools but about it's about it's about a collection of databases it's about a pool of databases that allows you to pay for performance across these and manage these as a pool of them and it all came out of the need for software as a service developers that said look i got like a thousand of these things and what am i gonna do pay for the maximum performance of all a thousand i'd like to share the cost and share the management across these which is why we call elastic pool that's really kind of what the concept is awesome thanks bobbin thanks darth vader um okay let's let's take a look at the answer what do you think the answer is bob well again we talked about and you know me i like to use process elimination so i'm going to tell you what it not is to get an answer of what it is so you want to migrate to the cloud that's a good one these all involve the cloud that's easy but you leverage a third-party app which requires access to the operating system now you're a little tricky because you say easiest to move to here right but it does say access to the operating system if you remember we talked about the fact that infrastructure as a service gives you operating system access platform or service does not give you operating system access because we manage even the vm itself not just the os in fact you don't even know what os sql's running on in a pass environment so the problem is here c and d are gone because databases you know those are paths environments and even though azure sql managed instance may feel like that to you like an instance again you don't have access to the operating system and that's a key tip in it because later i'm gonna i'm gonna give a tip for the audience you may you may get questions where we ask you about using the file system for managed instance you don't have access to the file system okay because you don't have access to the operating system so b can't be therefore the answer's got to be a yeah that's exactly right and i'm going to go ahead and put up the next question um but but we can go ahead and talk about it so this next question is another scenario so you should be able to see this on learn tv again or you can put your answer in the chat but the scenario is you want to migrate to the cloud and remove some of the management associated with sql server but your application leverages clr and service broker capabilities from sql server which azure sql deployment option will be the easiest to move to so this one's kind of a tricky one bob do you have any tips for folks as i think about this one well it's not just about easiest i'll that's not that was my tip it's more than just about easiest but i can see why somebody is going to look at that and may choose an answer that is not the exact right answer i would say there's a best answer here how about that the best answer uh yeah so go ahead and put put your answers in let's see i'm not seeing any just yet okay um [Music] i like how you said migrate to the cloud though that applies you threw that in there that was good it's good that you did that yeah i mean migration is a big piece of all this i know we don't dive a ton into it in this series but yeah looks like some answers are coming in i see quite a few of one answer okay um seeing right now about 90 folks saying manage instance and about 10 saying sql db okay and while we're waiting for your answers uh this is an interesting question that just came in uh malusi says is it possible to use ms sql server and link it to azure and how so yeah all sorts of ways to do that absolutely um you know for example uh you could even do a link server that's one certain way right uh and we even just mentioned the fact that for azure virtual machine you could have a azure vm be a secondary replica an async secondary replica as like a dr to your on-premises sql server that's an interesting scenario because we announced a licensing change that if that azure vm up there with is a secondary replica is just for dr purposes it's free how crazy is that free core licensing to do that with right so that is certainly one way to do it awesome thanks bob uh hopefully that helps with uh your question lucy uh okay so it looks like votes are kind of starting to slow down looks like we got 92 percent of people saying manage instance okay eight percent saying sequel db and a few people in the comments saying they were between b and c but couldn't decide yeah come on now you got to make a choice and a decision right so let's walk through it um we mentioned that uh surface level instance level features which is what clr and broker are think about those sql clr and broker are instance level functionalities you know that is something that is not an azure sql database which is why we did cloud lifter right so our already we got to eliminate you know c and d from our answers so now it comes to a and b and you know it is true that sql clr and service broker could run in a sql server in sql server and azure virtual machine but notice anna's first part of her question i wanted to remove some of the management associated with sql server and we talked about how platforms of service gives you that managed capabilities so i would eliminate a and see the best answer would be b yep that is correct okay it seems like most people were on board with that so great job to everyone participating [Music] so now that you've tested everybody anna maybe we should talk about pools right uh you know i think this is interesting because i remember you had this idea it was really good idea because because what did we do we did you know vm mi and db but then a lot of people are still torn like what is between mi and db again in fact i remember we were first doing this people were still asking us that so maybe it's kind of important for me to go through this a little bit um you know managed instance is that single instance with a surface area and native network support and fully managed but what we found was is we wanted a way to speed things up from a deployment point of view and actually provide a way to condense the number of instances that could show up in a virtual machine because with a single instance it's kind of dedicated so an instance pool came along which is in preview right now and so it can be a more cost effective way to do an instance in fact it's the only way to get down to two v cores because the minimum number v course for an instance is four so if you just want two v course for managed instance it's in public preview now you're gonna use this concept called instance pool and what you do is you go out there and build your instance pool first and then you deploy instances within it so the initial instance pool can take a little bit of time just like a single instance can but then the instances within the pool are pretty fast you know go through and deploy those things and then for azure sql database again it's a single database this is where your maximum storage with things like hyperscale or get that great capability with something called serverless which we're going to talk about and private link support but then you also have that pool concept for databases again this is a little different where remember you're kind of abstracted from the instance here right so you don't know where these exist but a pool concept is again i like a lot of these i'd like to share costs and management of these across a large number of these so i'm going to build an elastic pool and when you go in and do that deployment and i know anna in the module deployment you're going to see some of those options you get to choose hey i'm doing a database i'd like it to be part of a pool that i've already established so that's kind of really when you think about these two you've got managed instance on the side again it's a fully managed service but it is a full sql instance it's going to feel like that to you and you've got an azure database which you can have multiple databases for something called a logical server or a large number of databases is something called a pool yeah and i think this comparison is super helpful um you know figuring out which deployment option is going to be best for you and if you are migrating we do have tools that help you with that for example we just launched a new experience in azure migrate which allows you to actually assess your on-prem instances and it will recommend not only the deployment option but also the different service level objective and the service tier and the compute tier and that's what we're going to talk about next so we've talked about the deployment option so are you going to pick sql server and azure virtual machine azure sql managed instance or azure sql database and then within those are you gonna choose this pool's type of option or not uh so that's what we've talked about now once you've picked that there's still a few other choices that you have to make and really these choices are designed to give you flexibility but it's important to understand what the differences between them are so let's start by looking at the next decision that you have to make which is around purchasing models yeah and i want to emphasize to the audience we're just getting you introduced to this right so you're going to like in the next module we're going to talk about this again in fact we're going to repeat ourselves and talk about what these are because we're going to go through a sequence of your choices right but we wanted to introduce you these concepts because these are decisions you'll be making as you go through this process in fact if you're pre-planning about deciding mi or db this is going to be helpful to you so for a database an azure database you do have something called what we call a purchasing model to make a decision for the longest time we had this concept called a database transaction unit and that was your only choice and it's really kind of think of it as a pre-packaged way of looking at resources it's the simplified way for you to choose your compute like cpus memory io etc you just pick a dtu number and we have these levels called like basic standard and premium you pick levels within that and then we give you some sort of predictable performance number that your app would adhere to now we had a lot of feedback that people would come along and say i want some more flexibility i want flexibility for licensing and flexibility for choices to separate this choice of virtual cores and storage so therefore we came along with this vcore model now we say microsoft's recommendation because what we found is people wanted to apply that azure hybrid benefit they want to take their sql licenses and apply it to the cloud and this is the choice you need to make to get that plus people want things like reserve capacity or they'd like to pick the number of cores they'd like and storage the max storage separately and that's where this v core model you know comes into play if you want to make this choice you kind of have to decide for yourself but here's the good news you can go back and forth so if you if you go to the vcore model and say look dt is better for me or you go to dtu and i'd like the vcore model it is something you can use in the portal through apis to change that decision definitely and so that's the purchasing model and again like bob mentioned this is only available for azure sql database for azure sql managed instance you don't have this choice you just get the v core model so that's the purchasing models let's now shift and talk about the different service tiers that are in both sql database and manage instance but we're focusing again on the vcore model now the first service here that you have and is pretty self-explanatory by the name is a general purpose service tier now we're going to talk about this later in the availability module which is going to be week 5 of our series uh but just to give you a little glimpse sneak peek under the hood of what happens in general purpose is underneath we're deploying a failover cluster instance or something very similar to that if you're familiar with sql server and availability now this is like like the name suggests good for most this workload so this is your general purpose database uh it works really well it's cost efficient and there are some other options that can help with cost savings now the other main option that we have in both sql database and sql managed instance is business critical and again like the name suggests this is really meant for your mission critical type of workloads that need that low latency fast recovery and even some read scale right off the bat and again just to give you a little peek under the hood what's happening in the business critical service tier is when you deploy we're actually going to deploy an always-on availability group underneath the hood for you and that's how you're going to get this really fast recovery and readable secondary now we dive a little bit deeper into what that entails uh for general purpose you can think about this um as having remote storage so again we'll dive much deeper into this in the availability module and i've used some you know yelp style ways to compare the two so you're going to get more iops or io per second with the business business critical service here but that's also going to come at a cost because again we're maintaining this always-on availability group underneath the scenes if you need in-memory otp or using in-memory type of workloads you're going to want to use business critical as well because it's not supported in general purpose today and again like i like i mentioned about the always-on availability group that kind of explains the replicas you're seeing down at the bottom for business critical you have a primary and you have three secondaries uh but one of those secondaries can be used for read workloads and that's what that r stands for now on the other side you have uh you know less iops less cost no in-memory otp but you have an option another option between uh general between provision compute and serverless compute and i think bob you're gonna you're gonna dive deeper into that in just a moment is that right absolutely awesome so these are the main service tiers i'll hold off and let you guys have the suspense build on serverless uh these are the main service tiers available for sql database and sql managed instance but there is one more service tier for sql db yeah first of all i love the way he roll yelp style comparisons yeah only anna would say yelp style you know we've been alluding to hyperscale um hyperscale is something only available in azure sql database but has some very unique characteristics now notice here we say most business workloads but here's the key highly scalable storage and read scale requirements and so if you compare it to the other things we've shown here as far as storage goes and compute and replicas consider this hyperscale has a unique architecture which we'll show you in the availability module that uses a combination of local and remote storage okay and because of the way we actually do that the iops that you're going to get out of this are going to be slightly between what you see with the general purpose and brittle business critical type environment and the costs for this you're going to see are will again can be a little bit in between what you might pay for from general purpose and business critical but notice these last two points these are very important your largest database size capabilities come from hyperscale we publish 100 terabytes but we really think of it as unlimited storage and then notice the diagram compared what anna showed you in the left and right you get a primary compute you're just like you're going to work with anything else you're always going to have a primary replica you're going to be connecting to but notice here you get up to four four reed replicas that you can scale to as opposed to the fact that general purpose has no re-replicas and a business critical only has one re-replica and so you're gonna love the fact when we talk more about hyperscale during availability about this architecture and about the fact that for large scalable read scale operations you're gonna get the most availability here with hyperscale and because of the fact we're using local remote storage you're going to see some amazing functionality where we're going to talk about how point in time restore becomes blazingly fast because of the way we actually store the database files in azure storage but surface the data through to you through a tiered set of caching layers [Music] and so we'll let the suspensive side i mentioned earlier in azure sql database you have one other choice in the general purpose service tier and that's between provision compute and serverless compute so let's take a look at an example let's say i have a line of business application which most people use only during business hours so maybe between eight to five or nine to five in my time zone now after work or after office work hours people go home and they don't use the database anymore now with the provision compute model you're going to be paying for the compute you have whether or not you're using it or not and this has its benefits because it's guaranteed it's available it's there for you you've already budgeted for it but what serverless can do is in these periods for example you know the night time when no one's using your line of business application or the weekends where you know developers aren't working on an app or something like that uh you can actually enable something called an auto pause delay so what you can see here is in this example around 22 or 23 o'clock what happens is the database becomes inactive people aren't using it and no one's opening connections to the database and after that auto pos delay time which again is something that you can set here it's set to about two hours uh the database actually becomes pause so what does that mean that means that we actually have separated the compute from the storage and during this time we're essentially going to shut down sql server and so you're only going to have to pay uh for the storage at that point so this is obviously drastically going to reduce your cost uh if you have this kind of intermittent type of of workflow and in addition to that what uh serverless allows you to do is set a minimum and a maximum number of v cores so in this example you can see we set the min as one v core and the max as for v core and what it's going to do is it's going to auto scale you on a per second basis to the max of cpu or memory of what you actually need so this auto scaling capability is pretty powerful and they implemented a really interesting new cash reclamation system to kind of bring you back down after this period of high intensity cpu and this is great for you know line of business applications like i mentioned this scenario where people aren't really using it after hours on the weekend but it can also be great for things like if you have a new application you're not sure how much cpu it's going to need at a given time or if you're hand dealing with dev tests we've seen a lot of adoption in the serverless space for many many different use cases you know bob i'm sure you've also seen that as well with customers absolutely so now bob i thought it might be interesting you know if you have something you could show us we haven't seen any demos yet yeah what's a session with that demos i'd love to show you a demo anna cool let's take a look yeah so what i have here is a recording and i know sometimes people don't like recordings of demos but i recorded this for a very good reason because we're going to show you several aspects of serverless compute uh as a composed as compared to general purpose provision and some of this needs to be time lapse so it'd be better to show you kind of in a video so let me just kind of kick this off and let's talk through it so i've got an azure sql database and you're going to see here i'm going to zoom in i've got a general purpose to vcore general purpose we'll talk about gen 5 we're going to mention those things later during configuration so i'm going to run a performance test with my favorite tool called ostress okay anna you've seen me do stress before right gonna run a typical query against a sample database gonna kick off like eight users 25 iterations and i'm time lapsing the performance of running this workload on my two v chords so i'm going to kick this thing off for 20 o stress and it's going to come back and run in a certain time frame that for me takes a little too long so i time lapses a little bit it's going to get finished it's going to take about 14 seconds i want this to run faster i think i've got a cpu bottleneck right so what i'll do is i'll go back here in fact this is really cool the azure portal you can see the utilization it's almost 100 consumed for this general purpose to v core environment so i'm like hey i need more cpus no problem go into the portal use a little slider and say you know what give me eight cpus a couple minutes later the scaling operation finishes pretty low down time i run it it's going to come back boom five minutes that's pretty cool but you're probably asking yourself like what does that do with serverless so instead of having to do that operation i'm gonna take another database that's the exact same copy of that database and i'm gonna deploy it as a serverless database okay instead of it being this 2v core now what you're going to see in the portal something different remember anna mentioned about ranges you can you see right here in the portal now since i pick serverless i can say oh i'd like a max of eight cores but a min of two v cores also notice this auto pause that anna mentioned i'm gonna turn that off for now so i'm gonna go back i'm gonna run the same workload against serverless notice it took even faster than general purpose now that's not always guaranteed but i'd expect to have similar performance from my 8 v core deployment notice the portal looks different now with serverless i not just get utilization here on the left i get billing billing for cpu because anna mentioned building slightly different on the graph here on the left hand side of this peak is kind of like me getting billed for min v core usage okay so when i was not using it i got built for men when i was not you know running the app on the left and right side that's min but i'm getting billed in the middle there at the spike for the actual usage now here's what i'm gonna do i'm gonna turn on that auto pause feature anna talked about the minimum is one hour and of course this is where time lapse comes into play i'm to turn on auto pause and i'm going to run this thing but then i'm going to wait an hour and not use this now go look at the mesh metrics for billing okay shows here the last four hours i see a similar graph as i saw before you see how this works so i see my non-usage in orange shows up as min core usage you can see here the blue is min core uh billing and on the right side and then the maximum usage is there in the middle the orange is not using any at all the key is when i'm not using this but i've got min core not auto pause i still get build some but here the key is to the right auto pause is kicked in i'm not using it and i'm not billed for anything that is one of those interesting scenarios so i get ranges of cpus anna but i also get the ability to pause my database and not get billed for any compute usage just build for storage yeah it's it's a pretty powerful demo um i i really think serverless is just really cool personally especially for all my demo databases i'm always using serverless because you know it's all set up and ready to go but when i'm not using it it's paused and it keeps it really really cost efficient for me to use now i put up another poll for you guys to go ahead and take a look at um what i'll do is i'll just kind of explain the question and then we have a few questions in the chat so we can take some of them um and then we have one more question after this as well so thanks for all your participation uh so this question has another scenario and it says you're moving an application and a database to azure but your database is currently 62 terabytes with plans to grow you don't leverage any instant scope features today which azure sql deployment option will be the easiest to move to so go ahead and use either the uh learn tv poll or the chat to put in your answer to that question and while we're waiting for answers there uh we did have one question bob i think this is a great question about serverless um nick says would you be able to set specific times for example i always work between 9 and 17 hours and then stop yeah i mean you you have the ability to set auto pausing whenever you need to use it right you can enable auto pausing so you have the ability to go in since this is all based on clis and apis you could go in and set your auto pausing to be a certain time frame that i needed for and then you're not going to get charged in that time that's exactly like uh what you said earlier in that chart so if you say auto pausing i want to do auto pause and you say auto pause for an hour if you don't use the database between nine as you say here nine and 1700 then you wouldn't get charged so you're not setting a schedule up you're just saying the fact that when the database is idle during those times pause me so during that time frame if you didn't use it for a couple hours you wouldn't get charged but all of a sudden if you start using it again we'll ramp up and then we'll charge you only that time but then let's say you paused again for an hour so you did it again then you would become in the auto pause mode so you don't set a schedule up you just set for how long do i want to wait until i'm idle until i go to that pause state yeah exactly and one other thing that that you know you might want to consider when you think about what your auto pause delay is going to be is when this pauses remember sql server essentially shuts down uh so when it starts up again you are starting with a cold cache so if you're just hitting this intermittently but don't need you know need all of that stuff to be there for most of the day then maybe you want to set a little bit longer of an auto pause type of delay and i do want to call out though one thing anything you can figure with azure is programmable so if you wanted to change the configuration of auto paws or your ranges and so forth during certain times you could do that but it doesn't mean you set a schedule for pausing if that makes sense yeah definitely and um looks like we've got a lot of answers in for this uh question again this was about a 62 terabyte database and it seems like we had a few people say sql server and azure virtual machine uh but the vast majority of people are saying azure sql database single database um but however 20 percent of you are saying elastic pool so maybe that's something we need to talk about as well so bob do you want to take us through this one let's talk about elastic pool first lasting pool is a set of databases so you said here you wanted to actually have a database of 62 terabytes so elastic pool is not taking 62 terabytes and sharing it across a bunch of databases okay so that's eliminated here that's not the answer but it's good to know that elastic pool is not sharing a bunch you know data across databases but uh for 62 terabytes it's a set of a bunch of databases right so that wouldn't be the case now we didn't really specifically say azure sql managed instance doesn't support 62 terabytes but i called out that your maximum storage is not from managed instance so we're going to eliminate that one and sql server azure vm won't support that size today so the only option would be single database and it would be that hyperscale option that we mentioned yeah definitely so it seems like most of you all got that but maybe a few got confused on the elastic pools bit now i have one more question and this one should be an easy one because we've talked a lot about serverless today this scenario is you have an azure sql database with the serverless compute tier deployed with an auto pause delay of two hours after two hours of no activity what happens to your database and the incurred charges you know i always didn't like professors that said this should be easy i really did like that so it puts the pressure on yeah exactly [Music] yeah so go ahead and put in your answers for this um let me take a look because we have been getting a good number of questions in here um okay maybe this one i'm not sure if this is a question but miles says you don't pay for what you use you pay for what you provision so i'm guessing they're uh comparing uh provision compute for serverless compute yeah i think miles is a good point i mean let's just summarize it real quick so if you do provision compute you are paying for the number of cpus you choose whether you use them or not you pick eight you pay for eight and storage is independent of that right so you provision date you pay for that with serverless you're going to give us a range you're going to say hey i want a min of 2 potentially and a max of 8. if you're not turning on autopos that's key you don't turn on autopilos and you don't use any cpu you pay for men okay or you pay for what you use like you use four or six or even the max of eight okay if you turn on auto pause and you got two to eight and we pause you because you didn't use for use it for an hour even if you have minute two you don't pay for two you pay for none right that's kind of the key to it but one but be careful too it's very very important go look at the documentation about using things right in the database once you connect and start using it we're gonna start ramping you up and you're gonna start pay for that that min cpu so that's kind of how i would look at it thanks miles i think that's a great call out and again thanks miles for calling that out uh we have a few other questions coming in uh not all of them exactly relevant to this stream but um okay let's take a look at the answers so it looks like we've gotten a good number of responses in and they are mostly on a so azure will pause your database to stop compute costs and only charge you for storage so bob can you talk a little bit about how you would think about this question and answer yeah i think the key again is elimination um pausing is not about re changing a service tier so it's important about that concept service tier and don't be you know it's okay if you're like hey man i don't know what a service tier is we're going to talk about that term so many times right remember the service tiers anna talked about general purpose business critical hyperscale so you don't reduce a service tier you actually change how compute costs will happen for your database in fact as it turns out today serverless is only available for the general purpose service tier so you got service tiers and then under general purpose you can have provisioned or serverless okay that's not available in business critical or hyperscale so in that serverless compute tier which is part of general purpose you wouldn't change the tier you would only actually stop compute cost and only charge for storage during the time you're paused that's kind of the key yeah definitely so the answer is a which it seems like most of you were able to get which is awesome you've really been killing these uh knowledge checks so that's good we'll have to stay tuned for next week and hopefully stump more of you because that's my goal you will because you stumped me so yeah um so the kind of the the one of the last things you have to think about i mean if you can even think about making another choice is hardware and we don't have to spend a ton of time on this uh because when it comes to hardware typically there's one main option for you so in this case that's gen 5 or generation 5 and you can see some information about it on the left and you can you can kind of guess that you can kind of assume that as time goes on and hardware advances these options are also going to change for example there are some some future generations that became available recently uh some that are memory optimized versus compute optimized and there's also a newer version of hardware bob i love if you could share a little bit about it uh related to the dc series and this is for confidential computing yeah one just comment here you know it's not our intention to make this complicated for you to make these hardware choices right but we want to give you the option so most people are going to pick gen 5 but if you want to get the maximum number of cores for azure sql database you're going to pick m series for example but dc series interesting it's a new vm choice from the azure team which includes an intel chip specifically designed for enclave security so we recently announced the preview for azure sql database with secure enclaves and behind the scenes to make that work we're going to use intel sgx as the enclave the protected memory environment in the actual compute itself and so the dc series is exactly that it's built specifically to support our always encrypted secure enclave technology oh um but i think it's really interesting is what i was saying and it's something we'll dive into a lot in our security module which will be the third week so two weeks from now um now the last thing we have to think about or kind of you might be interested in is these interfaces for azure sql and i'll show you in a moment how we've kind of done this azure sql unified experience because again like bob mentioned we are trying to make this easier for you but there are some other tools and technologies that you're already familiar with using or you might be interested in using that you want to consider yeah anna you know um along the way when i started this assessment experience i thought to myself what am i going to use as a tool to kind of look at these things and i thought what does the majority of the world use and it's management studio so you need to have assurances that if you're a management studio user not only is it compatible to work with the azure sql choices we've talked about today but it knows how to snap to that so if you go into management studio and you pick what's called a logical server which we'll show you during deployment for your azure sql databases we will kind of change the menu options or the object explorer options that snap to only azure sql database or if use managed instance it'll look like a complete sql server instance in fact you could actually take azure sql data i think i've done this before you got to actually sql database azure vm and managed instance connect to all of them in a single pane of glass with management studio and kind of see those differences but azure data studio and that's one of your favorite tools to use that's a new and upcoming tool that we've built and just the notebook experience alone is worth using this tool you're going to see us do demonstrations throughout this series a lot with something called a notebook with azure data studio but there's more for example we announced private preview of a new migration experience to do in-line migrations to azure sql from sql server and it's going to be done through azure data studio so azure data studio one of the cool things about it people love it it works across operating systems so you're a macbook lover today you want a native experience use azure data studio linux windows and mac all works with that tool but i mentioned before earlier to the azure ecosystem that people like to use apis and command line interfaces and that's me so out of the gate notice here it says the word t-sql one of the things i emphasized to anna during our beginning journey and i think ann has a phrase it's just sequel was t-sql language is so known by so many people around the world that as they make this journey they're afraid do those t-sql skills go away hey here's the good news they don't just go away they get better we take the t-sql language and we enhance it to work with azure sql so all those t-sql skills you have all apply to everything you're seeing here in the cloud but maybe you're a powershell user that's going to work or new powershell commands we've built specifically for managing experience across these assets and then if you've never seen a cli i love a cli i use it all the time in the cloud shell works like on windows linux and macs and it's a great interface as well to do command line interfaces to anything you can do with azure sql but you really want to get gnarly i call it gnarly go to the rest ap i know my colleague dimitri is like bob have you tried a cli with rest to solve this problem like no i'm not a rest programmer i'm sorry but if you want to go to the lowest levels you can even do rest apis because all of our interfaces for azure sql paths actually support rest and then what's tried and true sql command and bcp they all work great all those tools work fine in fact again i use azure cloud show all the time to use those tools but yeah those tried true tools work just fine applying those to any azure sql assets we have today so look at our list here a unified experience in the portal but tried and true tools and new tools you know to help use your existing skills you know today but also i should grow you new skills in the experience with azure sql yeah and you know one thing i'll add is throughout this series if you stick with us which we really hope you will um throughout this series through the exercises you're gonna get to play with almost all of these different interfaces and tools uh and kind of see what makes sense to you you're gonna see like bob mentioned the azure cloud shell which is built into their browser into the azure portal and it allows you to do things like actual access azure cli powershell sql cmd all from within the portal uh using command line so there's a lot of really cool tools that we're going to take a look at uh but kind of to round out this session we thought it'd be interesting to take a look at this azure sql unified experience so at the very beginning uh bob mentioned how in 2020 we announced azure sql and we didn't just announce it as a thing we tried to follow through in some of the tools that we have so that you can see this azure sql experience as well so i'm going to hop over and show you an example of this so this is my azure portal view which you can see here um and what you can see is a few things you're probably familiar with seeing uh create a resource but what you also see is this azure sql button now this button is great because what it allows you to do is actually take a look at all of your databases uh all of your azure sql resources in a given subscription or you can set multiple you can set certain resource groups you can add other filters what's really cool about this is it's going to group them all together regardless of what azure sql deployment option they are so you can see i have my sql databases i can see if they're hyperscale general purpose business critical serverless i can see my manage instances my azure sql database logical servers which you'll learn a lot more about in the next session i can even see my sql virtual machine so this one's running sql server enterprise uh 2019 on windows server 2019. so there's lots of information here even elastic pools are here so i kind of get this single pane of glass to see everything that exists now similarly if i go to create a new resource um what's going to happen is i'm going to have more options but this should look familiar this is very similar to the way we design the slides where you have your three main deployment options and under each one you can dive into what those options are so for sql database you can dive in and see single database or elastic pool or database server for managed instance you can see options there and this is perhaps the coolest part is for the sql server in an azure virtual machine uh you can actually get every single image in one place now before this if any of you had been trying to deploy a sql server and azure virtual machine it was really hard to find the specific image you wanted and if i go all the way down what you're going to see is i can find a bring your own license sql server 2008 r2 sb3 on windows server and i can deploy that and make it easier for me to migrate and leverage right off the bat and you can see we also have images for other os's so not just windows but suse slash ubuntu red hat rel all those different versions are here for you to easily deploy a pre-installed image like bob talked about closer to the beginning so this is really awesome and one other thing you can see is we talked about the different service tiers so we'll go more into deployment in the next episode but just wanted to show you just like bob did earlier these different options you have under general purpose or business critical and then we also talked a little bit how hyperscale is available only in azure sql database today um so these are kind of your options this is kind of the unified experience that we're going for and yeah so if we kind of take a step back and think about all the things that we've done uh bob what do you think have we covered just so much do you think people are following what what's next what's going on i think some people are watching and they really love the overview we provided but they're also wondering like you know are you guys going to get into some gnarly deep stuff i can assure you harley in the next modules uh anna as we built this we're like bob i think we're going too deep so we are going to get deep as we move into the meat and potatoes of sql server but but think about what we talked about i mean i'm biased i think azure sql is the world's database i mean think of the options you have to put things in a virtual machine in a managed instance and an azure sql database and what we're going to show you in the next few weeks how easy it is to deploy this how to configure it how to secure it make it fast make it available the built-in capabilities we're going to show you i think makes it that but we have these options for you you want a vm we'll put it in a vm we'll give you some extra on top of it you need to run an azure right remember you're in a vm in azure so you just built in availability the vm because you're in azure but you want to manage the instance to get that built in availability we're going to give it to you in a database and go to hyperscale you have that as well you want to just lift and shift again you can take azure migrate and take a vhd image from vmware or hyper-v and just put it into the cloud you want an instance to get that power of paths remember the best of sql server diminished manage service that's managed instances and then if you want if you have a modern developer right now and you have a modern app and you just want to deploy a database and you want this thing to auto scale or pause when you need it or scale up to 100 terabytes you need azure database it's definitely the platform for you so i think about all these choices we have and out today i i really the intention we both had was to make sure people walked away from today's session and know what's possible with azure sql and then join us in this journey over the next five weeks to find out what do i do next and how do i use it yeah and we do have a whole journey ahead of us i think bob you you mentioned it quite well saying uh you know that this is really just the beginning we're going to get really deep so if this was an overview something you were already familiar with you know you're definitely going to want to join because we're going to go into some stuff uh you probably didn't even know existed in azure sql uh today now and a lot of demos coming up in the next ones right lots of demos there's so many demos uh so we're really excited to share them with you all and if you want to check out some resources in the meantime of course that forest resource is going to be for this learning path so if you want to start getting hands-on reviewing the modules then you can check it out at that first link uh if you want the slides that we're using a lot of people ask for our slides we actually host those on github uh so you're able to take them you can take that at that second link sql workshops um what else uh shameless plugs for other series that bob and i do one is azure sql for beginners so if there's something small you want to revisit this was a 61 video series that bob and i recorded that takes you bit by bit through the azure sql fundamentals content um and then another shameless plug for data exposed if you like these live streams sort of things uh we're actually streaming on wednesday bob you're coming on the show on wednesday well thanks for thanks for the notice so on wednesdays at 9 a.m pacific uh please join us on data exposed live and we also release uh recorded episodes on thursday you can find those on our youtube channel we have some books too because doing live videos and stuff isn't enough for us we have to go write some books uh bob wrote a book last year called azure sql revealed which follows pretty closely with this journey but it dives deeper into some of the extra stories and hidden gems that bob has found in his research and his just career and then we also have another book uh written by myself and several other azure sql program managers about practical azure sql database for modern developers so these are more this book is more geared towards folks who might be new to azure sql database or our developer heavy uh so definitely check out both those books they're really really great resources i read them both to help myself even though i was involved in some of them uh and uh yeah again if you want to access this module from learn today you can head to this aka link learn intro azure sql again it's part of this whole azure sql fundamentals learning path and bob that's that's not all is it yeah i just want to add one other thing that it you know honestly it wasn't even intentional for us i can't tell how many people that have gone through azure sql fundamentals or are going to go through the series that you're going to go through the next six weeks and literally just pass the dp300 exam i'm gonna have like i'm talking like a lot of people have contacted me and said like hey i just passed the exam i just did your stuff i'm like that's actually pretty amazing so if that's your thing and you want it and you need to pass that exam as part of your journey please this content can help you do that awesome thanks bob and you know just stick stick around with us you see the dates there but really we're here at the same place same time every monday for the next five mondays uh so go ahead and mark your calendars uh be sure to join us i'm just taking a quick look at the question we have about two minutes left um let's see [Music] um maybe this is this is a good one to end on bob i'm gonna throw okay see what you think uh this person says which approach would you take in a scenario where you are running a kubernetes cluster hosting multiple microservices where some of them depend on sql database would you host the sql server in a container or use azure sql database so this is kind of a tricky what to use when well you know um you take something like azure kubernetes service that provides a lot of platforms of service-like capabilities for kubernetes and a sql server container runs a greater than environment but you're not going to get built-in availability groups okay you're not going to get serverless scaling so if you need a lot of these paths capabilities that we provide azure sql database is going to be a better offering for you like a versionless sql server where a sql server container doesn't do that but hey kubernetes has built-in basic availability like a failover cluster instance and a sql server container runs really well in there it's really great for devops scenarios so a lot of people that choose kubernetes versus using a pass service but again remember pass gives you that ultimate built-in platform capabilities so that is going to be what does that person say it depends yeah but think of the container more of like your infrastructure service type capabilities because again it is just sql server running a container which is going to be kind of like running in a virtual machine like is awesome thanks so much bob thank you everyone so much for joining us today be sure to join us next week same place same time like comment subscribe all that stuff and we hope to see you next week on learn live bye everybody

Channel: Microsoft Developer

Views: 11,559

Keywords: Microsoft, Developer

Id: wcRb5RHriUc

Length: 89min 24sec (5364 seconds)

Published: Mon Mar 15 2021