Data replication strategies in HLD - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding replication consistency models

Which data replication consistency model ensures that all replicas return the most recent write for a given data item?

A Eventual consistency

B Causal consistency

C Strong consistency

D Read-your-writes consistency

💡 Hint

Think about the model that guarantees immediate visibility of writes across all replicas.

Strong consistency means every read reflects the latest write, ensuring all replicas are synchronized immediately.

❓ Architecture

intermediate

Choosing replication topology for low latency

You want to design a data replication system for a global application that requires low read latency for users worldwide. Which replication topology is best suited?

A Single primary with multiple asynchronous replicas distributed globally

B Single primary with multiple synchronous replicas

C Multi-primary (master-master) replication

D No replication, single centralized database

💡 Hint

Consider a topology that allows fast reads worldwide but may tolerate some delay in writes.

Asynchronous replicas distributed globally allow local reads with low latency, while writes propagate asynchronously to replicas.

❓ scaling

advanced

Scaling write throughput with replication

Which replication strategy best helps scale write throughput in a distributed database system?

A Multi-primary (master-master) replication

B No replication, single node writes

C Single primary with asynchronous replicas

D Single primary with synchronous replicas

💡 Hint

Think about allowing multiple nodes to accept writes simultaneously.

Multi-primary replication allows multiple nodes to accept writes, increasing write throughput by distributing load.

❓ tradeoff

advanced

Tradeoffs in synchronous vs asynchronous replication

What is the main tradeoff when choosing synchronous replication over asynchronous replication?

A Asynchronous replication guarantees no data loss but increases write latency

B Synchronous replication improves write latency but risks data loss

C Asynchronous replication improves write latency and guarantees strong consistency

D Synchronous replication guarantees no data loss but increases write latency

💡 Hint

Consider how waiting for replicas affects write speed and data safety.

Synchronous replication waits for replicas to confirm writes, increasing latency but preventing data loss.

❓ estimation

expert

Estimating replication lag impact on read freshness

A system uses asynchronous replication with an average replication lag of 200ms. If the system receives 1000 writes per second, what is the expected number of writes not visible on a replica at any moment?

A Approximately 50 writes

B Approximately 200 writes

C Approximately 100 writes

D Approximately 500 writes

💡 Hint

Calculate lag in seconds multiplied by writes per second.

200ms = 0.2 seconds; 0.2 * 1000 writes/sec = 200 writes lagged on average.