Design a key-value store in HLD - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Architecture

intermediate

Identify the main components in a key-value store architecture

Which of the following lists correctly represents the essential components of a scalable key-value store system?

A Client, Load Balancer, Cache, Storage Nodes, Metadata Service

B Client, Load Balancer, File System, DNS Server

C Client, API Gateway, SQL Database, Cache

D Client, Web Server, Database, Authentication Service

💡 Hint

Think about components that handle requests, speed up access, and store data reliably.

A scalable key-value store typically includes clients to send requests, a load balancer to distribute them, a cache for fast access, storage nodes for persistent data, and a metadata service to track data locations.

❓ scaling

intermediate

Scaling a key-value store for high read traffic

If a key-value store experiences very high read traffic but low write traffic, which scaling strategy is most effective to improve read performance?

A Add more storage nodes and replicate data across them

B Reduce the number of replicas to save storage space

C Use a single powerful server with more CPU cores

D Increase the size of each storage node's disk

💡 Hint

Think about how to serve many read requests simultaneously without bottlenecks.

Replicating data across multiple storage nodes allows read requests to be distributed, improving read throughput and reducing latency.

❓ tradeoff

advanced

Choosing consistency model for a distributed key-value store

Which consistency model choice best balances availability and partition tolerance in a distributed key-value store used globally?

A No consistency guarantees, allowing stale reads

B Strong consistency with synchronous replication

C Strict serializability with global locking

D Eventual consistency with asynchronous replication

💡 Hint

Consider the CAP theorem and the challenges of global distribution.

Eventual consistency with asynchronous replication allows the system to remain available and tolerate network partitions, which is important for global systems.

🧠 Conceptual

advanced

Understanding request flow in a key-value store

In a typical key-value store, what is the correct sequence of steps when a client requests a value for a key?

A 2,1,3,4

B 1,3,2,4

C 1,2,3,4

D 1,2,4,3

💡 Hint

Think about the order of checking cache before storage and how the load balancer routes requests.

The client sends the request to the load balancer, which forwards it to the cache. If the cache misses, the storage node is queried. The storage node returns the value to the cache and then to the client.

❓ estimation

expert

Estimating storage requirements for a key-value store

A key-value store holds 1 billion keys. Each key is 32 bytes, and each value is 256 bytes. The system keeps 3 replicas for fault tolerance. Estimate the total storage needed (in terabytes) including 20% overhead for metadata and indexing.

A Approximately 2.8 TB

B Approximately 1.1 TB

C Approximately 1.5 TB

D Approximately 1.8 TB

💡 Hint

Calculate raw data size, multiply by replicas, then add overhead.

Raw data per key-value pair: 32 + 256 = 288 bytes. For 1 billion keys: 288 × 1,000,000,000 = 288 GB. With 3 replicas: 288 GB × 3 = 864 GB. Including 20% overhead: 864 GB × 1.2 ≈ 1,037 GB ≈ 1.1 TB.