Which option reflects availability-oriented choices in distributed databases?

Which option reflects availability-oriented choices in distributed databases?

• A. Use a single, strongly consistent relational database for high availability.
• B. Cassandra, DynamoDB in multi-AZ, or Redis clusters.
• C. A single-node database with no replication.
• D. Any database can achieve high availability without replication.

Answer: (B)

Availability-oriented design means keeping the system online even if some parts fail, which is achieved by replicating data across multiple nodes and often across multiple availability zones. This replication lets the system continue serving requests when a node or zone goes down. That's why distributed databases like Cassandra, DynamoDB in multi-AZ deployments, and Redis clusters are the best examples: they are built to run as replicated clusters that survive failures and maintain access to data. A single-node database with no replication cannot maintain availability during failures, so it doesn’t embody availability-focused architecture. Likewise, claiming any database can achieve high availability without replication ignores the redundancy needed to tolerate component outages. The emphasis here is on distributed, replicated setups that provide continuity in the face of failures.

Availability-oriented design means keeping the system online even if some parts fail, which is achieved by replicating data across multiple nodes and often across multiple availability zones. This replication lets the system continue serving requests when a node or zone goes down. That's why distributed databases like Cassandra, DynamoDB in multi-AZ deployments, and Redis clusters are the best examples: they are built to run as replicated clusters that survive failures and maintain access to data.

A single-node database with no replication cannot maintain availability during failures, so it doesn’t embody availability-focused architecture. Likewise, claiming any database can achieve high availability without replication ignores the redundancy needed to tolerate component outages. The emphasis here is on distributed, replicated setups that provide continuity in the face of failures.