Vert.x Event Driven Non Blocking Reactive Toolkit
Download as PPTX, PDF0 likes108 views
Vert.x is an open-source reactive toolkit for building asynchronous network applications, supporting multiple JVM languages and event-driven, non-blocking scalability using an event bus. It utilizes a multi-threaded event loop and includes features like verticles for deployment and various common libraries for functionality such as logging, http handling, and messaging. Libraries can be configured in multiple formats, with capabilities for high availability and communication across multiple JVMs.
Vert.x Event Driven Non Blocking Reactive Toolkit
- 1. Vert.x – Event Driven Non Blocking Reactive Toolkit Brian S Paskin, Senior Application Architect, R&D Services, IBM Cloud Innovations Lab 6 April 2020 WebSphere Liberty Lunch and Learn
- 2. Introduction 2 Vert.x is an open source project by the Eclipse Foundation Is a reactive toolkit that can used as a serverless implementation or in an application Used for writing asynchronous network applications Supported under multiple Java based languages (polyglot) – Java, Kotlin, JavaScript, Groovy, Ruby, Scala Has two programming methods, with or without Verticles (spelled that way) There are different libraries for databases, monitoring, logging, et cetera Non blocking and event driven that is scalable using an Event Bus – Uses an Event Loop to process requests – Event Bus to handle communications between Verticles – Can span multiple JVMs and systems Can be used with Quarkus and GraalVM Requires Maven Shade to create a fat jar
- 4. Event Loop 4 Implementation of the multi Reactor design pattern – Continuously checking for new events and pass it to the proper Handler – Normal Reactor is single threaded, but Vert.x is multi-threaded Number of threads for the Event Loop depends on the number of processors – Is equal to 2 * number of processors – Not based on the number of Verticles deployed – Threads are assigned in a round robin method for Verticles Verticles will always execute on the assigned thread Worker Verticles have their own thread pools The Context is passed to the Verticle from the Event Loop – All Handlers in an Event Loop run using the same Context – Each Context has its own error handler – Maintains thread safety Never block the event loop
- 5. Event Loop, Contexts, and Verticles 5
- 6. Verticles 6 A Verticle is a unit of deployment Assigned a Thread in an Event Loop Each Verticle has an isolated classloader and cannot share global state Verticles can be passed configuration information 2 Different Types of Verticles – Standard – Workers A standard Verticle should be quick and a non blocking unit of work A worker Verticle is like a standard Verticle but using a different Thread pool – Can call blocking code – Used for limiting number of executions of code (ie. calling backend system) Can be run in high available mode – Deploys new Verticle if it dies abruptly
- 7. Event Bus 7 The Event Bus is used for Verticles to communicate with each other by sending messages – Single instance per Vert.x instance – Asynchronous Can be used with Verticles in the same JVM or across multiple JVMs and systems – Outside of a single JVM Hazelcast is used to transport messages Can be used to communicate with JavaScript Three types of messaging – Publish and subscribe – Point to point – Request – response Failures in all or part of the Event Bus may lead to lost messages
- 9. Common Libraries Highlights 9 Core – Deploy Verticles – HTTP Server – Basic Event Bus – Logging – JSON and Parser – Datagrams – Metrics – High availability and failover – Timer events – Buffers Web – Route requests to Handlers and sub routers – Routing based on criteria (HTTP Method, MIME type, etc.) – Authentication (OAuth, Basic) – Serve web pages – CORS – Uploads – Cookie Handling – Form handling – Virtual hosts – SockJS support
- 10. Common Libraries Highlights 10 Configuration – Multiple formats (JSON, Properties, YAML, HOCON, System Props, Environment Entries, etc.) – Can read default configuration file or specify file – Different configurations can be passed to different Verticles JDBC Client MongoDB Client SQL Common – Used in conjunction of one of the other data access libraries Reactive – Uses RxJava Circuit Breaks for Micro Services MQTT Kafka Client Mail Client
- 11. More information 11 Sample code on GitHub Libraries Build a starter application
Editor's Notes
- #3: Serverless implementation – an App Server/Transaction processor is not needed. SpringBoot is serverless. Reactive programming is a programming paradigm, associated with asynchronous streams, which respond to any changes or events. Vert.x uses an event bus, to communicate with different parts of the application and passes events, asynchronously to handlers when they available. a verticle is a unit of deployment Vert.x does not follow the JEE / Jakarta EE standards Uses Netty and Hazelcast under the covers Need Maven Shade to create a fat jar.
- #5: https://medium.com/@alexey.soshin/understanding-vert-x-event-loop-46373115fb3e https://vertx.io/blog/an-introduction-to-the-vert-x-context-object/ https://vertx.io/docs/vertx-core/java/
- #6: https://vertx.io/blog/an-introduction-to-the-vert-x-context-object/
- #7: https://vertx.io/docs/guide-for-java-devs/
- #8: \
- #10: https://vertx.io/docs
- #11: https://vertx.io/docs