HA and FT

High Availability Vs Fault Tolerance

HA : H/w , s/w and configuration allowing a system to RECOVER quickly in event of failure. FT: System designed to operate through a failure with NO user impact. More expensive and complex to achieve. 
 Planes are fault tolerant. They are having redundant electrical, power , etc systems.
 Multiple independent systems.

Highly available (HA) vs Fault Tolerant (FT).

HA means that there is some access to your application in the event of failure. It’s perf may be degraded. FT architecture is NO performance degradation during the outage. You can design either HA or FT. If you are doing FT, you will have extra redundancy and more cost.

DynamoDB and R53 are fault tolerant AWS services. There is so much redundancy in those services. If there is an outage, no DNS lookup or DB performance are impacted.

Disaster recovery

RPO (Recovery Point Objective) How much a business can tolerate to lose, expressed in time. Maximum time between Failure and last successful backup.

RTO (Recovery Time Objective) Get the system up and running. How long the system takes to recover?

Both RTO and RPO need to be as low as possible and expressed in time.

Scaling Ability of the system to meet the demand. Vertical Scaling 
- Additional resources by adding CPU, memory etc. Horizontal scaling.

• Adding more instances etc.

Tiered Architecture

Presentation tier : Interacts with the consumer/customer. Logic Tier : Business logic that does things and takes decisions. Data tier : controls interaction to db and data.

Encryption

Symmetrical process : same key is used for both encryption and decryption Asymmetrical process: Where different keys called public and private keys are used.