CAP Theroem & Consistency Model

CAP Theorem

Only two of these goals can be Strongly Achieved in a Distributed System #21

1. Consistency
1. All Clients will read the latest value #23
2. Availability
1. Availabality of Data(or Some part of Data) in case of node/network failure #25, #22
2. All Clients will be able to read and write data
3. Partition Tolerance
1. Database is split into Multiple machines. Will not go down even if a network segmentation breaks #23

Consistency Models

Note : Cassandra provides Tuneable Consistency( to move b/w Strict & Eventual ) to the Client Application #22

1. Strict Consistency #21
1. Read gets the latest update data(ie add, update, delete)
2. Causal Consistency
1. Reads get the latest update data only for operations that are causally related
3. Eventual Consistency(Weak Consistency)
1. Read might not get the latest update data. But future reads will eventually get the latest upate (ie all replicas of data will become Eventually Consistent)