1. Apache EventMesh(incubating): 事件驱动分布式应用运行时 陈广胜 Apache EventMesh(incubating) Founder&PPMC 微众银行技术专家

2.

3. • Why Event-Driven? • What is EventMesh? • Why EventMesh? • EventMesh core features • EventMesh in webank • EventMesh in other scenarios • EventMesh ecosystem • EventMesh future

4. Message Vs Event • Messages - Updates that are directly sent from one service to another. • Events - Updates that occur at a specific time and are not tied to any specific recipient or client.

5. Real-time intelligence and 'lossless' business activity analysis

6. EventMesh Architecture

7. REST-Driven Vs Event-Driven • • • • • • • Pros REST-Driven Systems Cons Tooling and documentation • Built on request-reply communication assumptions. Synchronous in Industry adoption and support nature. Simplicity of implementation • Point-to-point communication through the adoption of a client-server architecture and reliance on HTTP leading to tight coupling in microservices architectures resulting in a distributed monolith. • REST lock-in. If microservices communicate using REST, that will be the main form of communication preventing interoperability with non-REST applications. Event-Driven Systems Pros Cons Eventual consistency. The reason why I see this as a pro is because an • Complexity of handling asynchronous messages when using primitive event-driven system promotes a guaranteed high availability of the data event brokers. This complexity can be minimized by using an advanced where eventually all microservices in the system will receive the last event broker. updated value of the data. • Coupling with the messaging infrastructure. Reduced coupling between microservices; a message sender does not need to know about any consumer. The only coupling in EDA is a topic hierarchy structure. Loose coupling of applications instead of point-to-point relationships, resulting in easier scaling of the system. Load leveling and shock absorbing though queuing mechanisms.

8. ServiceMesh Vs EventMesh

9. Using Both Service Mesh and Event Mesh for Microservices

10. EventMesh makes simple

11. Feature: Pub/Sub+

12. Feature: Choreography

13. CPU is Not a Good Metric for Scaling Event-Driven Microservices

14. Feature: Better Metric for Scaling

15. Feature: Event Streaming

16. Feature: State Abstraction

17. Feature: Schema

18. Feature: Wasm runtime

19. EventMesh in WeBank Message Bus https://github.com/webankfintech/defibus APP APP APP APP EventMesh APP APP MQ APP APP APP EventMesh APP APP APP EventMesh APP APP APP APP APP APP

20. Other Scenario 1: Event-Driven Data Mesh

21. Other Scenario 2: EventMesh for Data Lake Hydration

22. Other Scenario 3: Decentralized Data Economy

23. Other Scenario 3: Decentralized Data Economy

24. EventMesh Ecosystem

25. EventMesh Future • • gRPC/CloudEvents Dynamic Routing, federated, serverless workflow standard transfer and codec protocol • • • Choreography • Event Streaming Support real time event processing • • OpenTelemetry Tracing,Logging,Metrics • • OpenSchema Avro\JSON\ProtoBuf…… • WASM runtime • Multiple language SDK • c\go\python …… • More Protocols • MQTT\AMQP\WebSocket …… • Support more event store • Kafka/Redis/Pulsar, blockChain, cold storage ……

26.

27. EventMesh助手 EventMesh公众号 • https://github.com/apache/incubator-eventmesh • https://gitee.com/WeBank/EventMesh • https://github.com/webankfintech/defibus • https://gitee.com/WeBank/DeFiBus