Scaling & Data Replication

Fault (Vs) Failure :

Fault is the CAUSE , Failure is the EFFECT

Fault Tolerance – It will replicates the server , if any one of the server fails also the other replicated server will give the Response to the Customer . it will applicable for DB as well if one DB goes down the other replica DB will respond.

Types of Fault :

Transient Fault : Occurs for a very small duration , Hard to locate

Permanent Fault : Continues until fixed , Easily Identifiable

Horizontal Scaling & Vertical Scaling :

Vertical Scaling : When you increase the Capacity of the Existing Resource its Vertical Scaling

Horizontal Scaling : Increase the Number of Resources its Horizontal Scaling

Scaling :

Increase The Load

System not Complex

Performance Should not take ahead

Performance Should Increase Performance

Database Replication :

To have an Copy of an Existing Database

The Existing DB its called as Primary DB / Master DB , The Replicated DB of Existing DB its called Secondary DB / Slave

Replication will reduce the latency by many Replication will be keeps around the Globe

Replication lag - is that time take the value could be Copy to Primary DB to Replicated DB

Synchronous Replication – All Replicas have to be updated before host is acknowledged

Asynchronous Replication – Host is acknowledged after Primary DB is updated , Replicas update Asynchronously

Semi Synchronous Replication – Here Its an combination of Synchronous & Asynchronous Replication , Here the Primary DB write will happen so it will update the values in all Replica DB so the Primary DB will except any one of the Replica DB acknowledge that The write has been replicated Successfully so the Primary DB will Confirms the Write Operation from the HOST

Snapshot – Its allow us to take the states of the Data

Ex : T1 = 6 table 10k rows (First Snapshot State)

T1+6 = 6 table 12k rows (Second Snapshot State After 6 hours)