Beyond the CAP Theorem - Consistency Without Consensus With CRDTs

Explore the world of distributed systems and learn about Commutative and Convergent Replicated Data Types (CRDTs) in this 41-minute conference talk from Devoxx. Dive into the trade-offs involved in building distributed systems, focusing on the balance between availability and strong consistency as outlined in the CAP Theorem. Discover how CRDTs provide strong eventual consistency without requiring complex consensus protocols like Paxos, RAFT, or Zookeeper. Examine the internal structure of CRDTs and how they enable safe concurrent and distributed updates without information loss. Cover topics such as distributed computing, data consistency, dynamic systems, semantic resolution, vector clocks, and practical applications of CRDTs in real-world scenarios like Cassandra and SoundCloud. Gain insights into advanced concepts like CQRS and garbage collection in distributed systems.

Intro
Distributed computing
CAP theorem
Data consistency
Data structures
Dynamic systems
Semantic resolution
Vector clocks
What are CRs
Consistency with data structures
C oddities
Delay
Fancy words
Joint simulators
Convergence ERD
GCounter
AP Encounter
Tombstone state items
Last working element sets
Cassandra columns
Jepsen
SoundCloud
CQRS
Garbage