Online presence system in HLD - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Architecture

intermediate

Design components for an online presence system

Which of the following components is essential to track real-time user online status in a scalable online presence system?

A A distributed in-memory cache with TTL (time-to-live) for each user's online status

B A centralized database that stores user status updated every minute by clients

C A batch job that processes user activity logs every hour to update status

D A static file server hosting user status snapshots updated daily

💡 Hint

Think about how to keep user status fresh and quickly accessible.

A distributed in-memory cache with TTL allows fast reads and automatic expiration of stale online status, which is critical for real-time presence tracking.

❓ scaling

intermediate

Handling high user concurrency in presence updates

If an online presence system must handle 10 million concurrent users updating their status every 30 seconds, which approach best reduces write load on the backend?

A Clients update status only once per hour to reduce writes

B Clients send updates to edge servers that aggregate and batch updates before forwarding

C Each client sends status updates directly to the main database

D Clients write status updates to local files and upload daily

💡 Hint

Consider how to reduce the number of direct writes to the main backend.

Edge servers can aggregate many client updates and batch them, reducing the write load on the main backend and improving scalability.

❓ tradeoff

advanced

Choosing between push and pull models for presence updates

What is the main tradeoff when choosing a push-based model over a pull-based model for delivering online presence updates to users?

A Pull model requires clients to maintain persistent connections, increasing server load

B Push model reduces server load but increases client latency

C Push model increases server resource usage but provides lower latency updates

D Pull model guarantees real-time updates without extra server cost

💡 Hint

Think about resource usage and update speed differences.

Push models require servers to maintain connections and send updates proactively, increasing server resource use but lowering latency for clients.

🧠 Conceptual

advanced

Ensuring consistency in distributed presence data

In a distributed online presence system, which consistency model best balances availability and freshness of user status?

A Eventual consistency with asynchronous updates and conflict resolution

B Strict serializability with global locking on user status

C No consistency guarantees, allowing stale data freely

D Strong consistency with synchronous replication across all nodes

💡 Hint

Consider tradeoffs between availability and data freshness in distributed systems.

Eventual consistency allows the system to remain highly available and scalable while ensuring data converges to the correct state over time.

❓ estimation

expert

Estimating storage needs for presence data

Estimate the daily storage required to keep online presence status for 50 million users, assuming each status update record is 100 bytes and each user updates status every 15 seconds.

A Approximately 720 TB per day

B Approximately 1.44 PB per day

C Approximately 288 TB per day

D Approximately 576 TB per day

💡 Hint

Calculate updates per user per day, multiply by users and record size.

Each user updates 4 times per minute (60/15), so 4 * 60 * 24 = 5760 updates per day per user. For 50 million users: 50,000,000 * 5760 * 100 bytes = 28,800,000,000,000 bytes ≈ 28.8 TB. The closest option is 576 TB assuming a factor of 20x overhead for indexing and metadata.