Database sharding strategies in HLD - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Horizontal vs Vertical Sharding

Which statement correctly describes the difference between horizontal and vertical sharding?

A Horizontal sharding splits data by columns across multiple databases; vertical sharding splits data by rows across databases.

B Horizontal sharding splits data by rows across multiple databases; vertical sharding splits data by columns or tables across databases.

C Horizontal sharding duplicates the entire database across servers; vertical sharding partitions data by user location.

D Horizontal sharding stores data in a single large database; vertical sharding stores data in multiple identical copies.

💡 Hint

Think about whether data is divided by rows or columns.

Horizontal sharding divides data by rows (e.g., user IDs) across databases. Vertical sharding divides data by columns or tables (e.g., user info in one shard, orders in another).

❓ Architecture

intermediate

Choosing a Sharding Key for User Data

You are designing a sharded database for a social media app. Which sharding key would best distribute user data evenly across shards?

A User's unique user ID (numeric or UUID)

B User's geographic location (country or city)

C User's last login timestamp

D User's account creation date

💡 Hint

Consider which key is unique and evenly distributed.

User ID is unique and usually evenly distributed, making it a good sharding key to balance load. Geographic location or timestamps can cause uneven shard sizes.

❓ scaling

advanced

Handling Hotspots in Sharded Databases

Your sharded database has a hotspot where one shard receives most of the traffic due to popular users. Which approach best mitigates this issue?

A Increase the hardware resources only on the hotspot shard without changing sharding strategy.

B Use vertical sharding to split user data by columns instead of rows.

C Move all popular users to a single dedicated shard to isolate traffic.

D Re-shard the database using a hash of user ID to distribute load evenly.

💡 Hint

Think about redistributing data to balance load.

Hash-based sharding redistributes users evenly, reducing hotspots. Simply upgrading hardware or isolating popular users does not solve uneven load distribution effectively.

❓ tradeoff

advanced

Tradeoffs Between Range-Based and Hash-Based Sharding

Which statement best describes a key tradeoff between range-based and hash-based sharding?

A Range-based sharding allows efficient range queries but risks uneven data distribution; hash-based sharding distributes data evenly but makes range queries inefficient.

B Range-based sharding evenly distributes data but requires complex hashing; hash-based sharding is simple but causes hotspots.

C Range-based sharding is only used for vertical sharding; hash-based sharding is only used for horizontal sharding.

D Range-based sharding duplicates data across shards; hash-based sharding stores data in a single shard only.

💡 Hint

Consider query efficiency versus data balance.

Range-based sharding groups similar keys together, making range queries fast but can cause uneven shard sizes. Hash-based sharding evenly distributes data but scatters related keys, making range queries harder.

❓ estimation

expert

Estimating Number of Shards for Growing User Base

Your application currently has 10 million users stored in 5 shards. Each shard can handle up to 3 million users efficiently. You expect user growth to 50 million in 2 years. How many shards should you plan for to maintain performance?

A At least 10 shards

B At least 20 shards

C At least 17 shards

D At least 15 shards

💡 Hint

Divide expected users by max users per shard and round up.

50 million users / 3 million per shard = 16.66, so at least 17 shards are needed to keep each shard under capacity.