Azure Service Fabric overview and the road ahead – BRK2030

Azure Service Fabric overview and the road ahead – BRK2030


Welcome to this session regular about service fabric road map and our overview. I’m going to bit of a warm up session really to a series of talks at follow after this. Link to as I go through this talk and there, we go more in depth to how you service fabric today. We also look at up to have company coming coming releases of service, Irving, Texas Remesh and we also diving into. How we using build it from a developer perspective so but in the end of this talk? I want to give you an idea of what service is. How many people use of a topic today? Quite a few bill how many people have no idea what server gives. OK and I mostly I want to come out of this with the concept of the road map of where it’s going. Ultimately, what it’s all about most of what’s happening now? When we talk about journey to the cloud is that people are thinking about how they’re actually trying to build a global service. And effectively hear the conversations change pretty much over the last 2 years now where ever and realizes now that the sort of the defector approach for billing services is a microservers approach. Become the standard way where UD couple an application down into smaller pieces of code deploy them as independent pieces of code that then. Communicate between the model and that means effectively that you need a platform to be able to build those applications on. That’s exactly what service fabric keys, but of course, you know order to do that. You know there’s a lot of challenges. There’s a lot of difficulties with building. Microsoft application here. There’s independent deployments. This communication between the services. There’s rollouts and upgrades and diagnostics, even have a good Telemetry built into your platform. Do you understand how to diagnose individual services that are running you have security issues you have to figure out? How you can do orchestrate you want a platform that figured out the best place to run your services on? And effectively Microsoft went through this journey itself. You know it moved from box products to the cloud in a built a platform on the way to achieve this so. Yeah, I might have protein Stephanie how you build services how you build things that run inside the cloud. But of course, there’s complexity involved in edge and distributed systems approach of things and you need a platform to help you. I’m super Sonic is the platform that we built inside Microsoft allows you to. Target and build services that run scale from a micro services approach. We use it to host and run huge numbers of our own workloads and we gave it away for free. About 3 years ago when we first released it all. There’s been running in G 8 is a product for couple of years now. Under platform effectively spans everything from the developer technologies, where you can build an application. You can build a set of micro services. We provide you built in runtimes. Communication stacks all the way through to drop regional site how you do deployments and roll outs and upgrades how you do. Around those. Applications as an orchestrator not need makes decision by the way things run, you can scale the underlying physical architecture. Infrastructure itself to add new machines and scale down machines to create a cluster. As we talk about really don’t know 3 pillars, but how you build applications are you deploy applications and how you operate applications on this platform. Because in the end, you’re a customer who wants to figure out business problems and how you build an application that runs at scale. Build a platform itself. The platform actually can host in all sorts of different environments. You can run your local dev box. You can host in your own data center. We have a version called Service Arctic stand alone, that you can run inside your own. Hardware inside, where if you have your detail bedroom VMS or physical hardware. We version, which we call Azure Service fabric today. That allows you set up a cluster and then I’ll touch on another version that we’ve released for service very mesh. So when you look at those pillars bills deploy operate in to build killer we have tooling there, we integrate with. From Visual Studio and visual studio code when youbring.net applications with the clips in all service code. Look at the Java site recreate vested in making Java First Class Citizen service fabric runs on windows and runs on Lennox. Well, not that excited we tend to focus on having building. Java applications inside debit of course, increasingly we see a lotof.net core there. A lot of people focus on service orbing.net core apps running on Linux as wellas.net core ABS building on windows itself. And then the whole approach is really that you can bring any programming framework of your choice and run at. Of course, we provide you reliable act is reliable services. Add programs to help you builders Microsoft Access yourself as runtimes that we include as part of service topic which. Which is why we talk about as a Microsoft application platform rather than just an orchestrator of things? In the end, you need more than just an orchestrated in order to achieve fill in global service. We look at the deployment side of these things, increasing this is focused around our dev OPS. I’m going to show you some tools around here how. You can take your code and you deploy it through a pipeline. We deployed through divorce pipelines, Jenkins and Maven cradle is what we look at on the Linux side of things. We reached great with third parties such as octopus deploy and we have all the environment, such as Azure. Dev test labs allows you to send her. Temperate clustered all your testing inside of that and in his so shut down the cluster overnight to save money. So we had a great service fabric of the Azure offerings, and inside you operate increasingly focusing on the diagnostics in the Telemetry that comes from your actual deployment. It’s incredibly important to focus on your devops your monitoring and diagnostics in your security as your basics on as you build a service. Before, you can get going to building the application architecture and everything else like this. So there Logan Latics Azure app. Insights are the 2 that we have from clever to the built into Azure. And we’ve integrated with a lot of third parties running through the Internet rated dog new relic have dynamics. All of those are available in the Azure marketplace. You can download install as alternatives for looking at Telemetry and diagnostics. It come from the applications you build and deploy. So sick inside of what you’re what are the offerings we have I mentioned a server it runs anyway. And I could be done with a number of different choices of how you want to approach and build. Where you want to run your applications to stand alone version allows you to think about 3 things you think about standing up with physical hardware. Signing up the cluster and stunning applications. But you get full control and benefit River that runs. With this in VMS inside your deployments or whether it’s physical hardware. If you want to move to until you could take that next you can run it. Another Azure cloud. Providers and we have a lot of customers who run on premise products. Very been very successful in a number of deployments particular receive financial companies have a big investment already in the data center where they want to run their service. Inside Asia, Divisor service fabric that allows you to build a cluster, we’ve had this out for couple of years now. You build a cluster environment deploy a service fabric application Internet cluster, you survive. We still entirely free here. You just get paid for the resource is it, you’re given and you could scale out an underlying VMS that you see. But the platform just deployed inside those set of VMS. We take care of the patching of the VM selves the patching of the service fabric runtime inside there. Scale out that cluster scaling back in. We hooking up the diagnostics few we integrate with underlying service service. Infrastructure so we control reboots of the machines, he don’t affect your applications and to whole bunch of intelligence. Clusters run scalable and flexible and can be able to take advantage of Azure platform, but of course, you actually get to see the machines, you actually get see the infrastructure. Who is your responsibility to think about the scale of how you scale at your cluster? Both those products are in GA one that we just released his preview very recently. Just re announce it at build it was a public preview that came out. In July. What we call Azure service fabric mesh now here you don’t get to see any VMS? Effectively, what this is, is it we’ve taken service fabric and we’ve hosted it. Microsoft has in surgery stand a massive clusters of machines thousands of calls. Now all you have to do is build an application and deployed in his environment, so effectively you’re just building applications. You don’t actually have to think about how you manage anything to do with the actual cluster itself. You’ve also join excitement of building a Microsoft based architecture and then you can simply say scale up the number of instances of. Application is running or add new services your application against applied so we’re very excited by this because it takes away. More of the Complexity of managing the environment itself, but of course, all of this is a spectrum of where you want to be. If you want to run a particular spoke VM image of your own choice. You know you have to go for Azure Service from a cluster. Version of that we have some customers who want to do that. If you want to run. Some particular virus. Checkers and things like that is your choice on the machines you need exercise machine. Yes, you go for that offering so pretty increasingly as we see people move to wanting to take on more managed service offerings. Azure service fabric fits very well into that space and we’re pretty excited about having that on a spectrum of capabilities. I’m going to dive into this little bit more and talk about and give you a demo what that looks like. But if you just kind of clicked off with a demo because I wanted to show how service fabric is pretty well integrated. With Azure Dev OPS. Number. We’ve long integrated with. There we go to prolong integrated with Azure develops, VSTS in terms of having. Build a pipeline tasks inside them and then just recently. A few months ago. We actually just integrated with Azure develops projects and I just want to do this because it’s a great way of just trying to learn how to get started with setting up your deployment pipeline. And you can then sort of modified after this is your devops. It allows you to choose a number of different programming frameworks in languages that you come up fromjava.net. Don’t move you pick thejava.net one you can choose the type of framework that you want to use. In this case, then we useASP.net core. When I’m here you can actually say I want to choose a service fabric cluster. And now aren’t able to do is I can go in and find my existing developed solution in sight here, so I’m sorry. Fossil. Here, I can set up a new project name I can call this a good night. 20. And what I’m able to do now is that I’m able to actually. Look at and deploy a full devil spy plane that will put a sample projects inside here and can provision a cluster for me. It is fair is little bit limited that the clusters. You can keep through this at the moment. There’s only a few versions of the VMS you can choose and has a full version of the windows cues inside here. But it has a location so I can choose for example, to choose it is. Q with windows and choose ADV 3 machines. And hit OK to this. Identify it done here this release, it up me devops environments. It will set me up a service fabric cluster. Unhook those 2 together and do a deployment of A. Open application into that one so here’s what I did when I did earlier. This is what I did for ignite 10, I called it. And. This this sort of cluster the actual cluster is set up here with this 3 machine cluster, so just sets up the streaming cluster. Advantage of this effectively you can just sort of test it out and give yourself going in blowing the cluster and hook up your own one after that. But effectively set up a repo here and in this repo. It just pulls in oursample.net application project, which is voting app. If you’ve never seen this before, is voting app is one that allows you to votes on categories of your choice in terms of things that you want to see. Can you put voting options inside here? And suppose it is voting data and then it deploy this into the environment. It puts up a puts up bills pipelines. I’m here to create the bill pipeline you see when I check this in and now when I do updates inside my. Code inside my repo. It actually pick up that build and a kicker for build for me and push it down through. The. The build process and build process consists of a number of tasks you see that it pulls down you get packages copies files. And then sets up and configure service very manifest was so great about this is it does a lot of automatic versioning for you so? When you deploy new versions of silver fabric applications. You have to change the manifest for the versioning files. He manifests bill tasks inside here, which changes versions for you so it’s a great way to get started. Going to go off and do the releases, I can kick off a new release here. I can go from to create a new release and keep this off in order to take a particular build here, the last build I did. Wanna couple of days ago I can choose it environment going to trigger this on and I can kick off and? Create a new building kick that off so great way to get started and just showing how you getting dev OPS going. Inside your pipeline is one of the key elements around how you think about. Setting up your your service that you want to hook into with service fabric. Let me just switch back here to slides. So what we see is that there’s actually just junie that everyone has to take from there going towards their digital transformation. Typically may start with some existing code and increasingly what that means is that you may have existed or your building new but this journey. We see today is that people are taking people who hosted a effectively things inside. VMS we call every hosting. But increasingly we’re actually seeing modern eyes ation of your applications where you’re taking existing code and deploying it into an environment like service fabric. Gives you the scale of your application and utility to deploy and right and you changes to those and start transforming your application. Microsoft’s architecture approach. You can either do that don’t service have it with running executables. We see that a lot or increasingly you can start to deploy. We the windows containers or Lanix containers or you can make some extra both that’s the beauty of service fabric. You can bring process is an containers run them side by side together. So we really kind of talk about 2 men journeys inside service, having the journey of modernization of existing code. Are you moving things so you taking advantage of a scale app platform and a classic examples here that you might have seen. Multiple deployments today inside your environment 5 machines here 5 machines here 5 machines here. And you can consolidate those all down and deploy them into one environment with one dev OPS pipeline with one telemetry system. Bring all the data from all of those. And then you figure you getting a cast consolidation exercise running things at high density. Typically, we see with all those small VM deployment. Today, with those single apps and machines might be just running at 5% CPU. But when you consolidate and run’em all down in a single environment like you run inside service fabric. You can run the machines at the 40 or 50% CPU much higher capacity. Cause other genie receiver service fabric is that you have all of the benefits or you have the advantage your building brand new applications and here you’ll hear the term talk. Cloud native applications. This is where you can start off with green fill application come in directly with that Micro Services architecture approach. Build new pieces of code and deploy those without having to pick up some of your legacy code in sort of decide how you migrate that forward. Through the suit journey that we typically focus on inside service fabric and it fits into an overall bigger picture with an Azure. We see a lot of this happening with customers today. We view service fabric for years to build high scale services. We started off with Azure Database is 10 years ago. Now week recently cosmos DB event, Harvin grid. All services at Rays inside Azure and even side outside Asia, like dynamics, 365 Office 365 as well. So we re Harden our own platform, we use it a lot ourselves. It runs on millions of calls. Thousands of calls daily and scale it runs out is. Pretty big way we see the whole platform growing inside of this. We also have. A large number of customers. We’ve seen tremendous growth in the last year of customers who are moving their applications and building new on top of us. Customers are sort of running in production today across the range of industries and he’s kind of quite good to look at the types of scenarios that those all fall into. Consider it is a general purpose platform, you can sort of build host any type of services. You want on that. With you coming from Windows or Lanix site or whatever your direction. You come from, but there are some sort of particular theme that we see. First, we see quite a lot of sort of media sort of giving side applications. Couple of examples of that is next game is launched a big game recently called The Walking Dead. Mr. sort of virtual reality game, you walk around with your cell phone zombies attack you? You fight back of them in real time with whatever’s lying around with you in terms of things you can fight zombies with them run it on thousands of cores running inside. Microsoft’s architecture. 10 Cent is a huge. Game provider inside China, they took service fabric and they hosted all on their own servers inside and hosted lots of legacy games. Lots of windows legacy games. Scale inside. Terms of building mission critical apps. We think companies exercise general build and run trading applications will sis closes. Bills tax preparation software tax preparation time of year we seem scale out there cluster their response for doing all the tax returns. Critical to their customers in order to getting them down at any particular it in time bound way that never deployed. His tax returns with thousands of different rules for each of the custom. Probably about 40% of all the planes in Azure run on service fabric that I OT deployments huge number of them. It’s sort of an ideal platform for IOT it can ingest lots of Telemetry data through things like IOT hub. It comes out eventually through event hub, you can build Microsoft Office is in Jester. Let data the design of our reliable services. And I should say are steep for program models allow you to kind of. Computer quickly. We have other low latency apps effectively allow you to do things like BMW have an application that allows you to. Track the location of your car and give you updates about your. Schedule of 10 different way when you’re trying to leave even let you try and find your car forgot where you park it. Men ever get sort of retail type applications. Alaska Airlines is a ticketing system a source is a big clothing provider. The whole payment system is built on top of service fabric how they’d all the payments when you buy. Anything that runs. So those are so some of the applications that we see. What I’m gonna do now? Is I have dived on us about brief brief overview or service fabric is in a sort of dive down into road map of what’s coming. This is sort of the exciting part and I can do this both in sort of a long term and short term view of things. Long term is like the arc of direction of where we’re going in short term is always expect to know. In recent coming up. Ultimately, we’re very focused on still getting service fabric into all sorts of different environments. With us on premise for the cloud and so you seen that recently service fabric is being integrated into Azure stack. Cindy’s redeploy Azure stand alone air and we’re going to go deeper integration error as we move along. In terms of making sure that service fabric not only gets the stand alone version deployed inside it will integrate with. Azure stack, so that you can actually stand up service fabric applications. Inside, there. Oh. Inside, when we look at Azure. We have an arc where we’re still making sure that we achieve availability zones. And plus regional clusters availability zones is sort of most. Is the most immediate one by time we get into this year? We want to make sure that we have availability zones running across the cluster that you deploy inside Asia if you’re not familiar availability zones. In some regions are multiple data centers inside an availability zone spans physical data centers in the region to give you high levels of availability. And then of course, we want to focus on making sure that we get windows and Linux containers running inside of him Custer. This is something where effectively. Having Linda Linux containers running on Windows is rare working closely with Windows team and by taking advantage of that. Launch Linux containers running on a windows environment. So then will be out of Lanix or windows. Containers running together inside the same cluster so that’s something that we continue to focus on. As a team. But in immediate short-term in the next next release it comes out. Having illicit going to schedule over the next few weeks. We released updating the next version for Windows Server. 1803, which has a whole bunch of fixes and support. Inside, there for containers. You see Red Hat Linux as an option available for building yourself a cluster inside the portal. Will provide you the ability to have isolated networks. There are 2 networking motor today inside your cluster where you can have your applications use. Actual ports and IP addresses of the host machine. Or you can take advantage of what’s called an open network, we deploy a single network across all your? Foster, where every application gets her own IP address and then the isolated networks allow you to buy owned an individual application. Inside. Its own V neck to give you a multi tenant act aspects now where my application knows I sleep with insider network. I’d very cheap piece of work as well that we sort of kicked off a few months ago that really kind of helps in our Java direction. Is integration with thing called Bosch? I’m going to bed huge amount of time talking about Bosch and you’re going to look at this, but Deployment platform that allows you to do release platform and deployment in life cycle for a distributed systems. So this is very well integrated into Cloud Foundry. If you’re aware of that, so we’re doing a piece of work, where we can take Bosch. The platform and use that for deploying service fabric across a large number of machines and in Bosch does the running updates? Of service fabric countries different environments. So now effectively as well as using the power shell that we provide you can use boss to deploy it onto your machines but Bush is actually very well integrated with throwing it across other cloud providers. I want to play into Amazon or GKY or one of the other cloud providers were actually into a Cloud Foundry environment. So it’s kind of a very interesting direction that we’re taking this is all about opening up service having to be deployed in multiple different environments. And then finally one thing that we put into this release is the thing called advent store. Aventador actually we did release it in our current release, which is 6.3. We now we do agree with the service fabric Explorer and this is exciting because all the events that happened with inside your cluster within applications. A priest is event store so you can see when I lost deployed application when I lost it and. Application upgrade when I last had to run time upgrade on my machine and we’re going to add lots and lots of this diagnostics so. All of this will get pushed out through Azure monitoring. Hopefully, I’ve got a little demo of this on this machine. But will at least look at this just getting this thing up and running inside here. So I swanigan look at my local cluster here a moment if I go into mobile local cluster here. I want to show you that doesn’t event store service OK. I’m gonna live installer service running on this machine. But if you were to get the next build. It comes out, you’re seeing events or service running on this machine here. And when you start to dive into anyone of these different nodes here, you’ll be able to see this event stabs here. And this event tab here will actually show you all those events for both the node level here at the application level. Neural cluster level and now you be able to do a query around those events and see these vents hear that get displayed. For events happen is your cluster it answers a lot of questions that we had from people ago. When was the last time I deploy my application? When was last time my runtime got upgraded inside my cluster so we kind of a pretty excited about providing? You have more diagnostics views into the actual. Clustering operations happen within cited as you’re running your service. If you want to dive more into what’s happening of the current product of best practices is a great talk that’s happening. Tomorrow, which is how to build. Actually, they haven’t updated my slide. There’s a slide here this is my my. Bye talk tomorrow, which is all about best practices for building. I’m on top of service for clusters. It’s actually on Thursday afternoon. This is the. The one to focus on our developer side of things. So on the morning and diagnostics side. There’s a venture series events come out of from our event store. Kristen to Azure Monitor of course, we continue to provide diagnostics threw up insights you get diagnostics out of your container deployments. And, of course I mean should integrate you with all our third party. So diagnostics and monitoring the key area that we’re going to spend more time of an investment in so that you can actually see more operations that happened inside your cluster and will diagnosing is more at one time. Let’s dive into our next road map. Those who about clusters less where we are with open source. We open source service fabric earlier this year back in March. At that time, we did a huge chunk of what we moved into service fabric into repo were excited by it. A lot of noise. It got a lot of attention. Since then actually we’ve moved everything out. Now there is fair amount of managed code that we hadn’t finished at that point in time. So now we’ve moved everything out, including a reliable classes, libraries, another kind of basic class in library so. You can fully build all all the source code now is available inside our repo and you can still. And you can use it to build lyrics images. And what’s up next is because it was so deeply integrated with our build systems is taking a bit of time to? Unwind are build for windows, but will have windows builds coming out inside out repo in the next few months. Windows bills bills for the service fabric runtime and so you will see that your bill from open source repo. And individually of course, our goal is to make sure that we develop everything inside the open and make sure that we? Build environment for our continuous integration there running and so we have daily builds and have the development side open that’s kind of the path and long-term road map that we have that. Pretty see that beginning of next year’s by time we get to that. So it’s up to these velop a road map. A developer road map is all about how we want any language any framework to build a build your applications with a Microsoft’s concept. Service fabric. What we going to done here to a large degree is we want to bring up service fabric to more languages? And more runtimes and so we transition to approach a bear by with providing you libraries now instead of just run time to have to derive again so today. Reliable services run time in order to build state for services. You have to use a particular runtime what we switched at 2 now. Is a library so I can take the libraries for example for our reliable connections that I can include that is part of. Any framework that I want to run inside. We’ve integrated to begin with itwith.net core and with spring so now I can fix we just build. Any.net core application and then I can just bring in my service record libraries like more reliable connection. Libraries there and then take advantage of using the? Steep for stateful data structures that they have so to be an enhanced those applications, so that’s kind of the direction we’re trying to go. Rather than making you have to build a bind against particular runtime. That way we can open up to spring is a language integrate with spring is a framework going to node. We can have any do python as a direction around that. As other jokes about making sure we go is trying to provide an intelligent routing and connectivity across the whole deployment for your? Your application we recognize today that there’s a lot of. Complexity infectivity, discovering and finding an application and communicating with it. We do a lot today, there to provide you. Communication frameworks, the discovery of those services. But invariably people still have to sort of build a front end gateway on top of their application. So you intelligent routing is where. Making sure that we’ve write a gateway and built into the actual runtime itself again. We could be exposed out to the XL Internet. And it can do intelligent routing of policy based routing of where messages come in and you can say. Request comes in look at this particular attribute send it. This particular service and then you can communicate seamlessly between the services inside big cluster. Old environment that you’re running in. This is actually being built on top of envoy. So envoy is actual reverse proxy we’ve taken a built on top of this and of course. We’ve been doing work for example, to put that over to windows as well as Linux. The other thing is part of our developer road map is very keen to make sure that services and applications have identity. This makes it a lot easier for you to write an application of identity associated with the service. The service can then take his identity. We can reach out to key vault get hold of secrets that it needs or keep talking s*** teams to talk to another service. For example, you wanna talk to cosmos DB database. Now it’s just a matter of taking the identity of the service is running. Getting a self education with key vault and then. Presenting those tokens, 2, Cosmos TV for authentication cider. So that will make the security and the dictation of things much, much simpler. This is generally theme across all of Azure that you’ll see. Just a road map in terms of the upcoming release in our releases coming up. You’ll have the reliable collections libraries. Running for Java, they will also not only are they persist states. Locally, but they will also have any memory versions and give you higher throughput for those sort of volatile scenarios. We are releasing a backup and restore services been in preview for awhile. Now the back of a resource service allows you to take deploy a service inside. Applications are still application. Into a remote store. And then if you have a disaster recovery scenario and have to rebuild your cluster have to recover all your state, from your state for applications again. You can do that automatically through their backup and restore service. And include as part of that we would eat if you are an A. A data editor that allows you to take the images their backup and restore snapshots there and you can edit them for you restore them. Pretty excited about this coz it’s sort of completes the story. Now it from AG perspective of how. You have the full cycle of having stateful applications in the back and re store. We’ve built a reliable local civic volume disk. Cool, about this effectively is that service fabric of storage technology, not only do we provide used it for compute through data structures like reliable diction in reliable Q? We’ve hooked up to just have local volume disk so if you want to just file IO operations and save some data. Local disk that replicated driving machines, you can take advantage of that ticket do follow through on goal. There is that we want to look at. Other. Types of storage technologies and they can use at volume drive in order provide high availability that to them. This high availability is you look like your cluster high throughput and low latency. I mentioned the application gateway that will be in the next release. It comes up being limited functionality to begin with his just focus on traffic, routing between those things. Between your services. And then importantly where actually have introduced a new resource application model, which we call a service running resource model. And the goal of this is driven by service fabric mesh and then serving mesh elevates your door building of an application. Up to just application level if you look at. The. If you look at the resource model today, that you build applications with it encompasses a very powerful set of. Characteristics allow you to actually change things as part of the cluster is really the application. So this resource model basically makes it much simpler few to think about how you just build just like. Those is a set of results is just think of as a set of resources inside service fabric I can declare them. Together. And I have a lot more into the whole developer story in this session here building resilient microcebus is. That will be a session that we run tomorrow afternoon, they’ll dive here deeply intodot.net core integration. Bible dictionaries, reliable services side of things and the whole Java side of how integrated reliable dictionaries with J cash. Demos around how we’ve done the work there doing. Java deployments and indeed agree with me, even another tool in pipeline. Go back to those that resource model, so effectively service fabric, please of buttons. We have an XML manifest need to model today, which is the one in may be very familiar with the view service fabric today. Resource model note makes it much more application. Centric simplified approach to building a set of resources that you then just sit together. And then we have the daca compose that we produce last year. That’s for anyone who wants to sort of move some legacy applications over Washington legacy applications just from doctor compose. Over to deploy in these runtime. That the key element here is that resource model is the one that likes huge loads you to Spanish Cross. All the different versions of service fabric because it’s centered on the most. And from there, you can run across all the different environments around this. So service RV resorts is what are their simple declarative way to deploy manage applications their individual resource is? For example, application is Resource Services Resource Network is a resource and you stitch mall together and you can deploy them into a service area measurement. Running today. And we’re very excited because effect of these resources are think of that they get translated they become effective just arm based resource. So the whole point of this really more than anything that we’re trying to drive is that you can now build. Microsoft applications. Arm as a driving approach to describing the application and deploying that into. You’re into your service recognition runs environment arm effectively makes it much simpler if you did declare the resource is declared the application and then do updates inside all that. So mesh what is it? It’s a service platform you don’t manage any of infrastructure you just simply build applications with that. Type model we have integration with Visual Studio tooling to help you build applications. I show you little demo around the side of things. I can. Take that application I can deploy it inside environment now my upgrades. Just simply doing arm updates, you know, we Microsoft to play a large cluster machines in which measurement inside there. And so we’re pretty excited answer brings it into the family of running long running highly available highly reliable. Microsoft applications in the cloud fear. In our preview rush you look at our road map for mesh perfectly doing many of the things I talked about in terms of. Developer Road map. Under resource model built into the actual mesh applications. You build including identity for the service pulling secrets from the key vault. Routing rules. Send messages between the actual. Application of services you deployed inside your civic mesh. Take advantage of the collections. Imposingly be able to bring it bring your own network, because now you know inside. This world of things you have some existing resource is that you want to hook up to like a SQL server database. Unable of great tooling PO from the Devil Pipeline Visual Studio Type Blank by plane and then of course, all the data in diagnosis. Pushed into different diagnostics environment like application insights. In terms of our immediate release where we going to release a very mention a very frequent basis. And you’re just see that in recent we have a public preview out there now that you can try out, we’re going to release again within the next few weeks. Never going to provide the auto scaling rules for metro applications as resource type resource type for doing secrets. And resource type for bill to set up a multiple inggris endpoints if I deploy an application I can expose multiple. External load balances on to my service end point through the Azure. Load balancer on to my actual application itself. I could courage you to dive into this talk here, which will happen tomorrow. Check will be in just talk. Wednesday we’ll dive deep into service from Asia will show you 100 demos, he’ll show you how you deploy. Applications resource model will show you arm and basically sure how. It’s a wonderful environment for building applications in diploid run at scale. I’m going to show you mess, but I’m going to show you it through avert a journey that we take going back to them modern eyes ation journey. It’s amazing how many customers that we see. Now, who are looking at their.net and Java World and saying How do I modernize my applications generally? And modern eyes ation, we be telling this story for couple of years now about how you take existing code. You can deploy this inside the service fabric cluster environment first just to get all the benefits of upgrades and deployments and. Diagnostics. And then you can start to add new little services around is just a journey that you take but increasingly that we see this happen with your modelization side of things frequently. And that’s because people now can deploy things in the cloud much faster you can deploy you want to have. Cast consolidation exercise, we actually even see his journey happening with cloud services as well. MSAT Microsoft E. As an application called pay and it deals with all our partner supplier paying today. They run that are in 4 different environments on a team cloud services that they deploy. Consolidated all of those, 18 different cloud services down to a single service fabric 5 machine cluster with each one of those, 5 environment. They saved himself 8787 percent of their money they’re spending about $23,000 a month and a reduce it down to is like $3000 a month in terms of their deployment. So even sort of the cloud service. Modern eyes ation as well as of existing application modernization is his journey that we’re seeing. Let me show you a demo of mesh on this journey of modernization and that way you can hopefully. Will give you an appreciation? What this looks like so here I have? An existing application. This application is an application that we’ve built is a smart hotel application is a hotel checking booking in system application. Web application this is this isASP.net WCF application and it took for SQL server database. And in order to move this to the cloud. Those couple of choices. I could do from the data side. I could move the data through SQL Migration Services and I can take advantage of manage database or I could take advantage of SQL managed instances, a sort of host my database as a managed instance running in the cloud. And then after that my application is like well, what do I do with this. Well, we’re going down here with certain messages I can right click on. My web application and I can do here add. And inside this. Inside his dad, I can do add container orchestrator. I want this means now is that I can choose. I can choose your Dick container orchestrator deploy these things into now in this particular case now. See. Fully loaded up the wrong projects here. We just go back here moment. Boy, you would see inside here is the fact that I can replace inside a mesh application. Let’s you sure I go inside here, so insert if I take this inside here and I do add inside here and do container orchestration this project here. See that I have this. I’ve I’ve children choice inside here of my service for being mesh. I’m not too sure why is reset of the moment? But I’m able to take this application. I can scaffold it into an environment are end up with a project like this here. I have a service fabric mesh scaffolding inside this and here I would have my web application my WCF application. Now brought in and what this does is effectively adds in it adds in a Docker filed to take a light code. Doctor image. I can debug that locally going to rich debugging experiences cycle that. And then I’m able to take this application. And I’m able to deploy it into the cloud so either through visual studio here, I can go into my actual deployment here. Actually, you can say publish and now with inside this publish here, I can actually choose my. My particular subscription. I wanted to my region. My Resource Group. They want to publish into an important Lee. I’ll take my containers that I built. I’m putting my code into a publishing through Azure Container Registry. But of course, is just an arm based model and what I wanted to show that was keys will of this. When I built this application. Really, the great value of Azure Service meshes that I can incorporate in the mesh application is part of any other Azure. On deployment so here I have a set of parameters. I handed turn on deployment here. I’m deploying a set of Resources, which is my SQL server database. But I’m also deploying and silver mesh network here, which basically punches out and gives me an English 10 point. But importantly here I have my application that consists of a couple of services inside here. These are the 2 that I just. Thankfully pulled in as part of a service fabric mesh projects and if you look at the properties inside the web application here. You’ll see that I’m able to describe this. As. A particular image that gives built sisters containerized and run inside service. Oregon mesh I consider server environment variables. App insight key that I talked to like instead of resource is here, so I can limit the resource is for my container image. Set up the number. So when I do this published deploy into this particular environment. All I’m actually doing is I’m pushing out unarmed template like this. So I can actually do a deploy arm template itself, providing all those printers, I can kick that off. And actually if I kick this off and provide this particular script allows you to. Invites you to ask you to provide their admin password just going to set up going to provide some password. Look up this script will see how that one goes. But what I’ll end up with effectively is that if I go to. If I go here. This is my application that I’ve deployed. It’s my checking application. I can see anyone of these particular customers that have checked in. Look at this deployment here, you’ll see. Now that I took these applications are containerized am I deployed him inside Safari medical does registration application. Diploid my SQL server database inside here that server with this database if I drill in his registration service. You see that now I have my WCF application. My web application containerized in diploid you have a couple of replication. If I click on this services. View I can actually see this from a services. View I can drill down. Actually, even go to the details of the actual individual replicas themselves and actually get down to the detailed logs. But I see for the individual application inside here that come out and do diagnostics around all this so now I’ve got my application really up in the cloud. I’m able to. If you look at this other visual Studio project here, but I did was I pulled in and modernized this one. By adding a new service inside here, so I did, I right click to my project here, I did and. You service. I did add new service service service inside here and so the next stage now is I did this. Just reaching KPI service. This scaffold intoASP.net core and middleware component that I can use reliable connections. Inside, there and. You’ll see that they have a folder for that particular new web API service that I added. Took advantage of the of the reliable collections library that we get included. So the scroll down. You’ll see that I’ve added a new service here that gets a list of data using reliable diction him. This trouble Dixie I still restoring some booking aggregate information search tracks how frequently summer stayed at my hotel. Keeps out in a reliable data structure here. This is how reliable connections pulled into the library so I had to anymore. And simply add that as inyou.net service integrated part of that. And end When I look at the actual application itself, you see. Now I’ve added this new tab. He can display KPI. It actually takes advantage of loading up the. New service talking to that and giving me the aggregate information of our customer has been staying how frequently inside my hotel. So to benefit effectively is that you’re able now to host dies. Containerized applications in a similar environment, I can go back to my resource model here, I can change the number of instance counts inside here. If I redeploy this inside here, you see. I have another version of a web application deployed here. I actually have. Multiple instances of my application running. Deployed earlier. In this particular case, I do. But I could deploy my update my senses and then you’ll see new instance of this thing running so it’s very easy to be able to say. I just changed a number of instances. I want to have a new secret just by changing my app. Definition through. So we’re pretty excited about mesh because it now sort of fees you entirely from the infrastructure management side of things. It takes you down to this world and this is the future, we see application development, you build applications. Excessive. Any runtime error code that you want to put together a polyglot application. Inside. You deploy these inside a service from a cluster environment. We can define recited fabric mesh. You going to handsome with a libel collections of your insight Service Road Metairie allow you to bring your own. Networks hope the mall up with and everything is connected through intelligent routing around that through the use of. Involved. Built in reverse proxy capabilities and you can deploy them inside any environment, they’ve your choice. So key takeaways ’cause I’m running out of time here, effectively is that you know you really have to start thinking about building Microsoft. Architectures you want to run anywhere service. Oregon even have to do that, it takes you through’s modernization journey. Hopefully it makes you happy and he’s your customers happy and it helps you divide built the business value. Not you’re not. Try to build a platform on which to build your service itself. There’s a lot of great sessions. I pointed out several of them throughout this talk, the only one I wanted to talk to you so far. That is just talk on Thursday was right. Take your next step. For Windows Server container orchestration between dive into like just service from it can be continual construction side of it all. Some other great talks here that done to think about how we containers and modernize Asian side and generally. The last one here is an example of a service recap built built in born for the cloud. Their customer talks through their journey. And that’s my brief, 45 minute run view of our road map and hopefully give you great appreciation where civil service. is? I would still have changed pretty few questions here in the end, but thank you very much. This taken from the audience tonight I can wait outside here after this yes. Yes. YouTube on how it how does AB service and service fabric forget. It is also I mean at service fabric and with our services in both his classes, you can build new applications from scratch. Yes. Cab services so very much focus is on websites and how you build websites and building around those things. They’ve done a greater integration there about how you can bring containers to run those websites. But typically that’s the sort of approach, they take. You can run website type applications on service fabric, but it’s more general purpose in that you can build. Any sort of service that you want. We can run it as an executable inside containers inside those and so you’re building a micro service application around those things so. Both of them are sort of cloud native in their regards if you bring you code from scratch. Both of them as well as having the modernization journey. Set up services is more focused on the website side of things service fabric is more focused on building a micro services architecture is scaled out. Service around. Sort of answer your question around. OK, yes. OK, so it’s good cook when I go service who securities so you know, we at Microsoft are deeply invested in sort of both those platforms. In fact, AB service service fabric and communities are the application platforms of service focus a lot on sort of website of things communities very much focused on. Engaging with a lot of the open source Linux side of the world of things particular that sudden focus at the moment. Date treat them selves as a container orchestrator and then very much after that, you go off and stitch together. All the things that you want to put together in that so you very much have to go off and find. Hi. What runtimes are not focused inside there? How you going to host the 08 front end gateway yourself? Still, build yourself a little bit more of a platform on top of the orchestrator so if you think about building a micro service application itself. There are many things you have to do, which in orchestrators part of that, so communities is Phil fits into that world of effectively. You’re more. Stitching together other piece of the platform and then you build replication on top of that. Service service runs on windows and windows in Lenox. We’ve integrated more offerings around that overall platform itself, so that you’ll be right. You built in runtimes. Visual Studio integration very, very much focus on the windows side of things as much as a lytic site. The community site is a bit limited moment they do run windows containers at the moment is just a little side of things, so as far as Windows Analytics, and built into it all. The community side will get there as well. Eventually, with windows containers, but I think there’s much longer on a road map. But effectively you know they think of themselves as container orchestrator open source community you go off and build things around all that. I’m so some of the support come from that side as well as. We’re fixing investing in all those platforms because we realize it in different people come in a different angle. Silly if you come in froma.net windows perspective we like to think that service fabric is the best choice around that by far. Same here we have plenty of Lanix clusters as well. Completely sort of hybrid reverse them or Lanix customers by far today coz that’s where they come from? Virtually no windows today because there not there in a moment but they will get there, so I think those are sort of 2 direction. But yes, it kind of choice a different platform. I think what you see particular serving mesh and a conversation that was all going to go over all. Orchestrator not that interesting. Now you’re trying to get to in building applications. Which is why things like measures so important because when you get to that level you? Don’t you don’t actually know service recommend actually runs on service fabric tool integrated pen some purp. You can run on corunna communities for you know really, but but it doesn’t. But but effectively you know that’s the level of abstraction. You want to try and get to. So I think you see every time these conversations route choices of orchestrated think that, like to go away, and it’s how you build his application. What were some of those things that wage in your community as well? Through like the sea and see if project. And then of course, we’re driving things through the proxy Billiton Azure. Give you an answer your question. I think I’m being kicked out So what we’ll do is I’m going to move outside. This could stand outside will answer a few of the questions there. And we also have a booth downstairs. Please come to our booth. We have lots and lots of people can answer your questions. Thank you.

Leave a Reply

Your email address will not be published. Required fields are marked *