Notification to all parties in LLD - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Architecture

intermediate

Design a scalable notification system for multiple parties

You need to design a system that sends notifications to all parties involved in a transaction. The system must handle millions of notifications per day and ensure delivery even if some parties are temporarily offline.

Which architectural component is essential to ensure reliable delivery and scalability?

A A message queue to buffer notifications and retry delivery

B A single synchronous API call to notify all parties at once

C A direct database update without any intermediate layer

D A batch job that sends notifications once a day

💡 Hint

Think about how to handle temporary failures and high volume.

A message queue helps buffer notifications and retry sending if a party is offline, ensuring reliability and scalability.

❓ scaling

intermediate

Estimating capacity for notification delivery

Your notification system must support 10 million users, each receiving an average of 5 notifications per day. Estimate the number of notifications your system must handle per second.

A Approximately 50 notifications per second

B Approximately 5000 notifications per second

C Approximately 600 notifications per second

D Approximately 10000 notifications per second

💡 Hint

Calculate total notifications per day and divide by seconds in a day.

10 million users * 5 notifications = 50 million notifications per day. 50 million / 86400 seconds ≈ 578 notifications per second, rounded to 600.

❓ tradeoff

advanced

Choosing between push and pull notification models

For notifying all parties, you can either push notifications immediately or let parties pull notifications when they check. What is a key tradeoff of using a push model?

A Push is simpler to implement than pull and requires no client changes

B Push reduces server load but increases latency for users

C Push eliminates the need for retries and guarantees delivery

D Push provides real-time updates but requires more server resources to maintain connections

💡 Hint

Consider resource usage and immediacy of updates.

Push notifications deliver updates instantly but need servers to keep connections open, increasing resource use.

❓ component

advanced

Identify the component responsible for handling notification retries

In a notification system, which component should handle retrying failed notification deliveries to ensure eventual consistency?

A The message queue or broker with dead-letter queue support

B The client application requesting notifications

C The frontend UI displaying notifications

D The database storing user information

💡 Hint

Think about where failed messages can be stored and retried automatically.

Message queues with dead-letter queues can hold failed messages and retry delivery, ensuring no notifications are lost.

🧠 Conceptual

expert

Ensuring notification ordering across multiple parties

You must guarantee that all parties receive notifications in the exact order events occurred, even under high load and failures. Which design approach best supports this requirement?

A Send notifications in parallel without ordering guarantees to improve speed

B Use partitioned message queues with ordering keys and idempotent consumers

C Store notifications in a relational database and poll clients for updates

D Use a simple broadcast without tracking delivery order

💡 Hint

Consider how to keep order while scaling and handling retries.

Partitioned queues with ordering keys ensure messages for the same key are processed in order. Idempotent consumers prevent duplicates.