First Study: CAP Theorem && Consistency & Availability
Strong Consistency
Idea is to make sure all copies are synchronously updated, and after ACKs, return data
- Server executes a write in primary DB
- Primary Db update -> replica DB
- Replica DB ack -> Primary DB
- Then Primary DB data -> Server
Eventual Consistency
Aynsc Update (May or may not pass - No ACKs Wait)
- Server writes -> primary DB
- Primary DB updates -> replica DB
- Primary no wait for ACK, send data -> server
3. Weak consistency
Used for:
- High availability, low consistency (read >> write)
- low latency
- Server write -> Cache
- Cache pass write -> MQ & then send ACK -> server
- Db consumes write operation -> perform write