Design a notification system in HLD - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Architecture

intermediate

Identify the core components in a notification system architecture

Which of the following lists contains the essential components needed to build a scalable notification system?

A Notification queue, Image processor, Delivery service, User database

B User database, Notification queue, Delivery service, Analytics module

C User database, Notification queue, Video streaming service, Analytics module

D User interface, Payment gateway, Notification queue, Cache

💡 Hint

Think about components that handle user data, message queuing, sending notifications, and tracking results.

A notification system requires storing user info, queuing notifications for delivery, sending them out, and tracking analytics. Payment gateways or video streaming are unrelated.

❓ scaling

intermediate

Scaling the notification delivery for millions of users

You need to design the delivery service to handle millions of notifications per minute. Which approach best supports this scale?

A Distribute notifications across multiple worker servers using a message queue

B Use a single server with multi-threading to send all notifications

C Send notifications directly from the database server to reduce latency

D Batch all notifications and send them once per hour to reduce load

💡 Hint

Consider how to handle high volume and avoid bottlenecks.

Distributing work across multiple workers with a message queue allows parallel processing and better fault tolerance. Single servers or direct DB sending create bottlenecks. Hourly batching delays notifications.

❓ tradeoff

advanced

Choosing between push and pull notification models

What is a key tradeoff when choosing push notifications over pull notifications in a mobile app?

A Push notifications provide real-time updates but require persistent connections

B Push notifications reduce server load but increase client battery usage

C Pull notifications guarantee delivery but cause higher latency

D Pull notifications are faster but consume more network bandwidth

💡 Hint

Think about how push notifications work compared to clients requesting updates.

Push notifications deliver messages instantly but need persistent connections or platform services. Pull notifications require clients to check periodically, causing latency but less connection overhead.

🧠 Conceptual

advanced

Ensuring notification delivery guarantees

Which strategy best ensures 'at least once' delivery of notifications in a distributed system?

A Send notification once and do not retry on failure

B Store notifications only on client devices to avoid server retries

C Send notifications in batches without tracking individual status

D Use a message queue with acknowledgement and retry on failure

💡 Hint

Consider how to handle failures and retries in message delivery.

Using a message queue that requires acknowledgement and retries on failure ensures notifications are delivered at least once. Sending once without retry risks loss. Batching without tracking loses reliability.

❓ estimation

expert

Estimating storage needs for notification logs

Your system sends 10 million notifications daily. Each notification log entry requires 1 KB of storage. How much storage is needed to keep logs for 30 days?

A 30 TB

B 3 TB

C 300 GB

D 300 TB

💡 Hint

Calculate total notifications over 30 days and multiply by size per notification.

10 million notifications/day * 30 days = 300 million notifications. Each 1 KB means 300 million KB ≈ 300 GB.