Overview
Explore the implementation of java.util.concurrent APIs for distributed coordination in this conference talk from Devoxx Poland 2019. Delve into Hazelcast's new implementation built on the Raft consensus algorithm, examining how Java locks, semaphores, and other concurrency primitives can be effectively utilized in distributed environments prone to partial failures. Learn about the evolution of distributed coordination tools, the benefits of high-level APIs, and the challenges faced during development and testing. Gain insights through code samples and practical demonstrations, including a look at adding redundancy. Understand the importance of replicated state machines, CP sessions, and the advantages of using established frameworks for coordination tasks. Discover the roadmap for future developments and access valuable resources to further your knowledge in this critical area of distributed systems.
Syllabus
Intro
Hazelcast
Challenges
Distributed Coordination Systems
A Simple Locking Recipe for Zookeeper
An Opinionated & High-Level Framework
Why Raft?
Replicated State Machines
CP Sessions
DEMO 12: Adding Redundancy
Recap Avold writing your own implementations for coordination
Roadmap
Resources
Taught by
Devoxx