Design scalable notification system
Overview
This question evaluates a candidate's ability to design scalable, low-latency, multi-channel notification systems, testing competencies in distributed systems architecture, reliability and fault tolerance, data modeling, API design, capacity planning, message routing, monitoring, and cost trade-offs.
Airbnb
Jul 29, 2025, 8:05 AM
Software Engineer
Onsite
System Design
22
0
System Design: Low-Latency, Multi-Channel Notification Platform
You are asked to design a scalable, reliable notification system that can send messages to millions of users with low latency across multiple channels (email, SMS, push).
Assume a large, consumer-facing product operating globally with both transactional (real-time) and bulk/marketing use cases.
Cover the following:
-
Requirements Gathering
- Functional, non-functional, traffic assumptions, SLAs/latency targets, compliance.
-
High-Level Architecture
- Core components and data flow for real-time, scheduled, and bulk sends.
-
Data Model
- Key entities (templates, preferences, messages, attempts, providers, etc.).
-
API Design
- Producer APIs, admin APIs, idempotency, status callbacks/webhooks.
-
Message Prioritization
- Priority levels, queueing, fairness, rate limits, quotas.
-
Deduplication
- Idempotency keys, content-based dedup, time windows.
-
Retries and Failure Handling
- Backoff, dead-letter queues, poison-pill handling, fallback channels, circuit breaking.
-
Scaling Strategies
- Partitioning, horizontal scaling, multi-region, autoscaling triggers.
-
Monitoring and Alerting
- SLIs/SLOs, metrics, logs, traces, runbooks.
- Cost Considerations
- Unit economics by channel, routing, batching, budgets, frequency caps.
State reasonable assumptions where needed and explain trade-offs.
Solution
Login required
Show