app-designs¶

Useful links and key notes¶

Basics of request-response pattern and Event based¶

Request-response pattern¶

- Point to point messaging
- Suitable for synchronous integration use cases
- Not suitable when request continuously flows in (like a stream). 
    REST / SOAP web services do not provide characteristics to build a scalable, reliable real time infrastructure for a high throughput of events.
- Technologies: HTTP, REST/SOAP, gRPC

** API technologies that use asynchronous communication: WebSocket, MQTT, Server-side Events (SSE), Kafka Protocol

Event based pattern¶

- Messaging / Pub Sub (sending data from A to B , C)
- Suitable when request continuously flows in (continuous data processing)
- Often asynchronous 
- Event-driven, supporting patterns like Event Sourcing and CQRS (command Query request segregation)
- Technologies: Kafka, MQ, PubSub

** Some vendors provides Messaging (ESB), Storage, Processing capabilities.

Interesting articles¶

Relationship between API and event streaming ? 
    https://www.kai-waehner.de/blog/2020/05/25/api-management-gateway-apache-kafka-comparison-mulesoft-kong-apigee/

Why synchronous and asynchronous communication together ?
    https://www.enterpriseintegrationpatterns.com/ramblings/18_starbucks.html

Youtube video Martin Fowler – Microservices¶

https://www.youtube.com/watch?v=2yko4TbC8cI

- Microservices thinks interms of breaking Application into components. 
    Idea of benefit is that a component can be replaced or thrown away if required without affecting other components.
- Smart endpoints and dumb pipes
    SOA has systems that communicate with very smart middleware/infrastructure area (ESB - enterprise bus). The smart pipes figure out how to communicate.
    In Microservices the network / communication is dumb and intelligence is moved to end points. This is how microservices differ from SOA
- Decentralised decision making
    Monoliths has 1 database for entire Application. 
    Microservice and SOA mandates each service gets its own database and they cannot be shared.
- Design for failure
    Being ready for failure
    Netflix has chaos monkey that goes around and randomly brings down services