Hello everyone, I'm Anvil Yanarachi. Thank you for joining my session on simplify network services for real world cloud native applications with ballerina. Without further ado, let's dive in. Ballerina, at its core is an open source cloud native programming language designed to make integration a breeze. It is a product of neutered by the Fox at WSO since 2016 and it officially released in February 2022. And that's not all. Barina comes with a vibrant ecosystem offering a plethora of network protocols, data formats and connectors. What's exciting is that you can craft your code the way you like it, whether through text or visually, using a sequence diagram flowcharts for that added quality, Marina brings in built in, user friendly and user efficient concurrency, all backed up with safety features, ensuring your development journey is smooth and secure. In modern programming, everything is an endpoint. May it be a database, a security, handling or even Internet of Thing Android device, everything can be an endpoint. Interesting thing is that the application that we are building or cloud native apps are increasingly depending on these endpoints. So effectively what we are building is an application which is talking of the network with massive number of endpoints. Integration is the discipline of resilient communication between endpoints. It isn't easy, you know that there are a lot of technology and techniques have designed to help build system like compensations, transaction events, circuit breakers, discovery, protocol handling and mediation. These are all hard problems to solve. In the past we had two ways of solving this left hand side. We have used and using systems like eases eais or enterprise service plus or enterprise integrations business process management to solve this problem. These things are understand integration. It helps to do integration simple and they have one big challenge and that is these not agile. In other hand you can use general programming like Java Node and these challenge of this they don't understand the integration, they are not integration simple. So developer has to take these responsibility for either solving hard problem or they have to find a suitable framework to support that. Camel and spring integration framework are some common framework people are using. These are complex bolton framework and don't necessarily be integration simple and still have high learning curve and complexities. So we kind of came to a conclusion that we call it integration gap which is you can either be agile but not integration simple or integration simple but not agile. As an integration company WSO two, we have been working for more than a decade to solve these integration problems and we have been working on more than 1000 integration projects and almost every project end up in one side of that integration gap now let's see how ballerina is minimizing this integration gap. Ballerina is a cloud native language, so why it is cloud native language? The language network primitives simplify writing services and deploying them in cloud native environments. These primitives make it eases to handle network related tasks, which is crucial for cloud native application development. Ballerina is a flexibly typed language, so structural types with support powerfulness are a key feature in ballerina. They serve two purposes, one enhances static typing and two describing service interfaces. This flexibility is valuable for designing and maintaining complex systems. Barina allows for typesafe decorative processing of various data formats, including JSON and XML and tabular data, so thus making the data processing over network much simpler. WSO language integrated queries simplify data manipulation, enhancing productivity and code quality. Balina programs offer both textual and graphical representations. These graphical form based on sequence diagrams provide a visual way to understand and design program flaws. Also, it makes your code documentation much simpler. Davino excels in providing easy and efficient concurrency management. The use of sequence diagrams and language managed threads simplifies concurrent programming without the complexities often associated with synchronous functions. Barina also enhance program reliability and maintain boot to several means. Explicitly handling static typing and concurrency safety contribute to most robust applications you can develop. This is all achieved while maintaining a familiar and readable syntax, making it easier for developers to work and understand the code base. Let's see ballerina in action. When it comes to ballerina, the preferred IDE plugin is visual studio code. It also offers support for various IDE plugins for visual studio code such as docker, Kubernetes, Opentelemetry, Corio, Copilot and GitHub. Recently we have added support for Asia function as a service deployment as well. So when you are using vs code, everything will be everything that you need to compile, debug, run, observe and monitor. Your application will be with visual studio code. Wso as I mentioned, Barina is a graphical programming language as well as syntax programming language. So this is a sample of what marina looks like. In visual studio code you have a simple program with the main function and on the left side you can disview the code. On the right side you can see the sequence diagram that simplifies what this code does. So what this code does is actually connect to the GitHub API and then get the open pull request and then using the Google Sheet connector it add those pull requests to the Google sheet. So you can simply check this sequence diagram and understand the code better. You can also edit the sequence diagrams to generate code as well. It works both ways, similar to the previous side which shows a sequential program. It also works with an integration designer as well. This is a sample ballerina service, a rest service that we have run, HTTP service. So it also displayed the service the types that it has declared and you can also view the open API configurations of this service as well. You can navigate to the SQL diagram view and then in that view you can also view the open API of the service that you write as well. So vs code also support autocompletion of ballerina programs. It also knows the libraries that ballerina has. So we have a set of standard libraries for doing the standard functionalities of a programming language. So those functions will be simply you can add it, simply export it and you can just use them in your IDE. The IDE will automatically complete the values there for you. Marina has also cracked the challenge of mapping one kind of data into another kind of data, which is a common scenario in integration. You can do that in code as well as our graphical editor as shown in this picture. So it is very easy when you do the graphical one, you can just add a functions to convert value from one end to another value. Balino also support data persistence. So when you write a code or declare your objects in Balarina those are called records. When you declare your records, Balina will automatically map the entirety and relation diagram between them. You can simply generate the SQL code that you want to execute to run this Barina program. So that also works graphically as well as textually. Marina also supports debugging your application via vs code. You can remote debug your application and check the values that you receive for your inputs and check your logics in the debug mode, simply add debugging pointers to the code and start the application in debug mode. If you are using multiple services in your application, you can picture or graphically view them using the architecture view of the vs code plugin. Balina vs code plugin so this is a complex system that handles multiple services. You can simply go to the architecture view to see the connections between those services and you can read down on each component and see fine details as well. Bavina has inbuilt support for multiple tools that is required essential in integration. So one is these OpenaPI tool which generates client and skeleton for the open API specification that you receive. GraphQl tool which generate client skeletons in ballerina for your graphQL endpoints. Async API tool which generate ballerina service and listener skeletons for an async app contract and strand and dump tool which dumps and inspect current available strands of a ballerina program which is used for performance testing then health tool which generate fhir HL seven profile to client and stub generator tool of ballerina. We also have support EDI tool, the set of command tools that provide to work with EDI files in Balrina. This is just most used tools. There are many more tools that you can use with ballerina. Now let's look at how ballerina is the deployment of the program that you have developed using unique features of ballerina. Bell command will generate an executable jar file which you can use to run with bell tool or using the Java ADK version as well. You can also simply do bell run in your program which will compile and run your program in your machine. So both of these are supported with Balrina. Balrina also support native compilation using GalVM. You can simply say bell build minus minus GalvM which will compile a machine code according to your machine architecture which runs on your machine as an executable. You can also build a docker container using bellbill command which packs your application to a docker image. So ballerina compiler is aware of your application and it will automatically generate docker file and docker image for your application. You can simply say bell build minus minus cloud declare docker which will generate docker file and a docker image for your application. You can also build GalVM compatible docker image as well. You can simply say bell build minus minus cloud declared docker and minus minus Galvm which will compile a docker image into a galvm. So these are the stats of some popular frameworks and ballerina in using GalVM. So as you can see, ballerina pitimus has the same or better experience compared to these values in other framework. Not only that, ballerina also generates Kubernetes artifacts for you as well. When you just write a service or a main function or anything, you can simply say bell build minus minus cloud equal ketest which will generate yamls that are required to deploy your application in kubernetes. This also build a docker image for you as well and that docker image will be added to your deployment yaml and you can just say Kubectl apply and give that Yaml folder. It will automatically deploy your application into Kubernetes cluster as well. Balina also supports function as a service. You can write a Balarina function, either make it either deploy it in azure functions or AWS lambda. There are built in support for these things. You can simply add annotation, or you can simply write an Asia function executable and generate an Asia function executable to deploy in Asia functions as well. Well, you can now deploy ballerina as a Java file docker, image Galvm, Galvm plus docker and kubernetes. Now let's see what we can do with observability in ballerina. Every ballerina programming is automatically observable by any telemetry tool. You can view all the visibility and codes, behavior and performance automatically will be published in open telemetry syntax and you can simply add them and view them in any open telemetry supported tool distributed login also supported by ballerina. You can simply say no hub bal run and add this one and redirect your output to a ballerina log. Or then you can tail this log as well. So you can view these log value outputs in elasticsearch as well. All right, so let me explain about my setup a bit here. I'm using vs code with Balarina extension installed and my current ballerina version is Swan Lake update seven let me explain about the scenario a little bit. What I'm going to do is I'm going to write a service, HTTP service, which will provide the location and weather information when you provide an IP address to that one. In order to do that, I'm using two API endpoints. One is IP API, the other one is weather API. So this IP API endpoint, what it does is it will return the location data. These you provide an IP address for this one. These location data contains latitude and longitude belonged into that IP address. So now I'm going to extract those two location data, latitude and longitude, and then pass that one into the weather API endpoint where I will be passing that latitude and longitude value to get the current weather information for that particular location. Then I'm going to combine the results of these two and provide an aggregated weather data update for that particular location which contains following items. So it contains last updated time of the current weather, temperature condition fields like weather information. Then it will also contain the location information, city, country and these ISP. So I have exposed this one as a service called with base eases geo data which runs on port 99 and I have defined a resource function which does all of this which accept IP address as a path parameter and returns the weather data or a bad request or internal server depending on the results that we receive. Let me show the graphical view of this program real quick so you can simply click on this icon which will explore the graphical view and it will give you a glance of what you have written. Here we have this get two functions, one service, one record and three module level variables. So what these functions, these is this one get geodata, the other one get weather data. And this service, it contains two resources and it features this weather data and I have used the user key which is the key for the weather app that I need to invoke. So I am reading this value from the environment variable when we run and it is configurable in ballerina. So let's take a look at the service view of this graphical view. You can see the resource functions that I have written here graphically similar to that open API spec level visualizer. You can also view the cloud code view of this one which explain these logic that is written inside this resource endpoint. So we have this endpoint which returns which is very clear and easy to understand. If you want to document your code you can simply copy and paste this value and be done with it. All right now let's try and invoke this service in my local machine. So I'm going to use the terminal for this one. In the terminal I have created two terminal instances, one to compile and one to invoke this services I'm going to do just do a bell run on this folder which will compile my program and then also run it in this 90 90 port. In these terminal I'm going to use curl to invoke this one and get the output. All right it compiling. Let me type the curl real fast until that we can do HTTP and it's running on my local machine in port 90 90. My base path is your data and the weather is these endpoint and it accept IP address as a path parameter. Now it is running on port 90 90. I can simply knock this one as you can see. Now you can see the output of this program and the logs that I have added in this program which prints the location information and the weather information. And finally we can see the output on this window which contains the details for my current IP address location. Let me invoke another IP address as well. Let's say one seven two here. All right, it's from United Kingdom and it's heavy raining at the moment. All right that's one way of testing. Let me show you another way. Using the graphical way you can simply click on this try it button which will invoke the swagger editor for this API which contains all the results, services and the port you can provide. And I can just say get weather IP and try it out and enter an IP address here. Let me say one 7210 21 and execute. It will give a nice output in graphical view as well. So United States partially cloudy and these are these temperature information. That's the graphical view. I want to just give a glass. You can do lot more with this graphical view. Now since we have running this, let's run this one in Kubernetes. So I'm going to do this without writing a single YAML file or building a docker image. All right I have stopped the service for now but before doing that I want to add another opensource to this one to check the readiness of this service. Let's do it by graphical way. I'm going to go to service and going to add the resource and it will be a get resource and my path will be health readiness. So this is the endpoint that I will invoke in my readiness probe and the response for that will be 200. Okay and I'm going to do it and then save. As you can see the code for this one got automatically invoked. Now I'm going to do a return HTTP okay here. Okay now we are ready to deploy this one in Kubernetes. So in order to modify the Kubernetes artifacts generated by compiler we can use this cloud TML file. Here I have added a cloud TML file to configure the name of the docker image that we are building and the tag that we are providing for that image. Also as you can remember we have a configurable variable which requires the user key to operate and eases it to the weather API. So that for that one I have already created a config map which contains this user key and I want to use that in my deployment. So I am saying that use this info config map and use that key refer its key user key in my deployment. Now I want to add the readiness probe to this one. As you can see we can simply type readiness and it will give you the suggestions here. I have to give the path. So my path will be health readiness, it should be geodata health readiness and my port will be 99. All right now we have added the readiness probe, let's compile. So I'm going to use bell build minus minus cloud equal Kubernetes command which will compile this program. Then pack that giant to a docker image and the docker file will be generated along with the YAMl file. Let's see. All right, we got an error. In addition. All right, as I said earlier, the compiler is aware of the resources and these YAml files, so we can see that health actually contain an EZ here. So I have missed that one. Let's add it and try again. So until it bears, let me explain. By my Kubernetes cluster, I'm using rancher for desktop along local cluster, so I can use the same docker registry in my local machine in this Kubernetes cluster as well. All right, now it's generating executable. Likewise it generates Kubernetes artifacts that are mentioned here along with the Docker image. So these things should be in there and it will finally print the commands that you need to run this program. If you go to the target folder and Kubernetes file, you can see that the file is already generated and it has created a service with port 90 90 of cluster ip. Then the deployment with the same labels which will match these service labels. And as I mentioned earlier, we have this config value parameter which will read from this one and inject it into the container that we have provided. These name is Geodeeployment and the port is GeopRC one. And it also got that readiness trip configured as well with these initial delay of 30 seconds. We also generate an horizontal auto scale for dotoscale. If you look at the Docker file, it contains the jars that you require to run this one and the docker image as well. So now I'm going to do, let me quickly run a Docker images command to show whether the Docker image has built. I'm going to do docker images. As you can see, we have this geo weather data v one 10 just created about a minute ago. Now I'm going to run this command which we got from this output of this ballerina one kubectl create minus f and the part to the ML file. All right, everything got created. Let's see the status. Okay, still it is running. I think this is because of that initial delay of 32nd we have added. Let's wait for 30 seconds and see. Let's see. Other resources also got created. Get SVC. Sorry these, we got this network services and get HPA as well. Okay, we got this HPA as well. Still it's not running. Let's check these status describe pod. Okay, it's still waiting for that readiness pop. Okay, now it's running. Now we want to expose this service as a node port in our cluster. So I'm going to type the command that we already retrieved from that. Build output, expose deployment, your deployment as a node port. Okay, now this will create a node port service which we can use to invoke this service in the Kubernetes cluster. So it's exposing this 90 90 via this 32 two four cluster. Now let's invoke that one and see whether we get the same result we got from curl. Before doing that, let's log the pods, tail the logs of the pods as well. We are going to do this log. Okay, now we just do same, but with the port that we got for the node. What's the port? Let me quickly get the port. It's 32 two four. I'm going to use that in the curl. Okay. All right, we got the log and the result. Let's try another IP address as well this time like one, one two. Okay, we got our local results. As you can see, the pod logs are also being printed. Okay, now that's all for the demo that I want to show. This is just a scratch of the surface of what ballerina can do. In this demo, we have exposed Ballerina's network primitives, the dual textual and graphical syntax, and we also experienced the ease of concurrency management and how easy it is to deploy in Kubernetes or docker environments. The power of ballerina does not stop here. It's a dynamic language that keeps evolving and there's always more to discover and leverage in your cloud native applications. I encourage you to take what you have learned today and apply it in your projects. Furthermore, you can experiment, collaborate and innovate with ballerina as you build your scalable, efficient and relevant cloud native applications. If you have any questions or would you like to explore ballerina further, please contact us with one of these channels, these websites I have mentioned in this slide you. Thank you once again for joining with us. Have a fantastic day ahead.