Lesson 3: Building a Big Data Infrastructure Part 3
Real-time Processing & Messaging
Messaging is Good For
Long Running Tasks
Managing Bursty Events
Decoupling Services
Denormalizing or Copying Data
Messaging Semantics
Producers or Publishers
Brokers, Queues
Consumers or Workers
First in, First out
Kafka
Distributed - Brokers, Producers, Consumers
Persistent
High Throughput
Queues through Topics
Open Source - Built at LinkedIn
Other Messaging Systems
RabbitMQ
Kestrel
Redis
ZeroMQ
←
→
/
#