News feed generation in HLD - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Architecture

intermediate

Design components for a scalable news feed system

Which of the following component sets best supports a scalable news feed generation system that handles millions of users with low latency?

A User service, Feed generator, Cache layer, Database, Message queue

B User service, Monolithic feed generator, Single database, No cache

C User service, Feed generator, Cache layer, Multiple databases without queue

D User service, Feed generator, Cache layer, Message queue, No database

💡 Hint

Think about components that help with asynchronous processing and fast data retrieval.

Option A includes a message queue for asynchronous feed generation, a cache for fast access, and a database for persistence, which together support scalability and low latency.

❓ scaling

intermediate

Handling high write throughput in news feed updates

When millions of users post updates simultaneously, which approach best handles the write load for news feed generation?

A Batch updates once a day to reduce write frequency

B Directly write all updates synchronously to the main database

C Store updates only in cache without persistence

D Use a distributed message queue to buffer updates and process them asynchronously

💡 Hint

Consider how to smooth out spikes in write traffic.

Using a distributed message queue buffers incoming updates and allows asynchronous processing, preventing database overload and improving scalability.

❓ tradeoff

advanced

Tradeoffs between push and pull models in news feed generation

Which statement correctly describes a tradeoff between push and pull models for generating user news feeds?

A Pull model reduces latency but increases storage cost; push model reduces storage but increases latency

B Push model reduces latency but increases storage cost; pull model reduces storage but increases latency

C Push model increases latency and storage cost; pull model reduces both latency and storage

D Push and pull models have identical latency and storage costs

💡 Hint

Think about when feed data is prepared and stored.

Push model precomputes and stores feeds, reducing latency but using more storage. Pull model computes feeds on demand, saving storage but increasing latency.

🧠 Conceptual

advanced

Impact of fan-out on news feed system design

What is the main challenge caused by high fan-out (users with millions of followers) in news feed generation?

A It causes excessive read load on the database

B It reduces the number of messages in the queue

C It causes excessive write amplification and storage overhead

D It simplifies feed generation logic

💡 Hint

Consider what happens when one user’s post must be delivered to many followers.

High fan-out means one post must be copied to many feeds, increasing write operations and storage needs, known as write amplification.

❓ estimation

expert

Estimating storage for a news feed system

Estimate the daily storage needed if 10 million users each post 2 updates per day, and each update is stored in the feeds of 500 followers. Each feed entry is 1 KB.

A Approximately 10 TB per day

B Approximately 5 TB per day

C Approximately 1 TB per day

D Approximately 20 TB per day

💡 Hint

Calculate total updates, multiply by followers, then by size.

10M users * 2 updates = 20M updates/day. Each update stored in 500 feeds: 20M * 500 = 10B entries. Each 1 KB = 10B KB = 10 TB.