1. Active - Active
- Allow Servers to have dedicated Data Stores
- These Data stores replicate
Problems: CAP –> ref: CAP Theorem | Consistency & Availability
2. Active - Passive
- Has Data recovery system, that ensures Partition Fault Tolerance
- Helps for recovery in disaster management.
Active - Active Techniques
Since this is more used for scaling in distributed DBs, we deep dive here:
1. Two - Phase Commits
- Make sure all nodes are replicated (Synchronous Copy) - ( ref: Consistency Models )
- Once all gives back ACK –> response send to server
2. Multi - Master
- Record to be updated in one of many possible clustered nodes
- Serious limitations, when it comes to data consistency.
- Sharded Database –> Sharding Databases
Comparison:
