Shelf · System Design
System Design
Scale, caching, queues, consistency and the interview framework, always with numbers.
1
The method
PlannedDifficulty
Priority
- Design interview method: the eight-step flowMediumP5Know it exists
- Clarify the user journey and scopeEasyP2Talk about it
- Turn nonfunctional goals into decisionsMediumP2Talk about it
- Back-of-envelope estimates with unitsMediumP2Know cold
- Draw and narrate one complete requestEasyP2Talk about it
- Manage a 45-minute design conversationMediumP5Know it exists
2
Building blocks
PlannedDifficulty
Priority
- DNS and CDN on the request pathEasyP2Talk about it
- L4/L7 load balancers and reverse proxiesMediumP2Talk about it
- App tiers and resource saturationEasyP2Talk about it
- Cache and queue as supporting componentsEasyP2Talk about it
- Database and object store responsibilitiesMediumP2Talk about it
- Search as a rebuildable read modelMediumP3Talk about it
3
Scaling
PlannedDifficulty
Priority
- Vertical versus horizontal scalingEasyP2Talk about it
- Sharding and hot keysHardP3Talk about it
- Stateless services still have stateMediumP2Talk about it
- Replication is not sharding or backupMediumP2Talk about it
- Read replicas and replica lagMediumP2Talk about it
- Find the next bottleneck before adding nodesMediumP2Talk about it
4
Caching
PlannedDifficulty
Priority
- Cache patterns and invalidation contractsMediumP2Know cold
- TTL and eviction solve different problemsEasyP2Know cold
- Invalidation under concurrent reads and writesHardP3Talk about it
- Prevent a thundering herd or stampedeMediumP3Talk about it
- Match Redis structures to the access patternMediumP3Talk about it
- What not to cacheEasyP2Talk about it
5
Async, queues & streams
PlannedDifficulty
Priority
- Messaging semantics: queue versus logMediumP2Talk about it
- Ordering, partitions and consumer progressHardP3Talk about it
- End-to-end idempotency and effect boundariesHardP2Know cold
- DLQs and controlled redriveMediumP3Talk about it
- Transactional outbox and relay failureHardP3Talk about it
- Backpressure and fan-out with bounded workMediumP2Talk about it
6
Distributed fundamentals
PlannedDifficulty
Priority
- Distributed consistency: CAP and PACELCHardP3Talk about it
- Consistency models and quorumsHardP3Talk about it
- Leader election and split-brain preventionHardP4Talk about it
- Distributed locks: leases and fencingHardP4Talk about it
- Clock skew and ordering without wall-time trustMediumP4Talk about it
- Two-phase commit versus sagaHardP3Talk about it
7
Reliability engineering
PlannedDifficulty
Priority
- Retry and timeout budgetsHardP2Know cold
- Circuit breakers and recovery probesMediumP3Talk about it
- Bulkheads and graceful degradationMediumP3Talk about it
- Rate limiting and overload enforcementMediumP2Talk about it
- Load shedding before queues become failureMediumP3Talk about it
- Health checks that support recoveryMediumP2Talk about it
8
APIs & service boundaries
PlannedDifficulty
Priority
- API contract design across REST, gRPC and GraphQLMediumP2Know cold
- Pagination styles with stable orderingMediumP2Know cold
- Versioning and compatibility contractsMediumP2Talk about it
- Service boundaries: monolith or microservicesHardP2Talk about it
- Decomposition without a distributed monolithHardP3Talk about it
- BFF and gateway responsibilitiesMediumP3Talk about it
9
Data-intensive system design
PlannedDifficulty
Priority
- Batch and streaming designMediumP3Talk about it
- Orchestration as durable run stateMediumP2Talk about it
- Data quality gates before publishingMediumP2Know cold
- Backfill and replay with late dataHardP2Talk about it
- Event time, windows and late-result policyHardP3Talk about it
- Exactly-once in pipelines has boundariesHardP3Talk about it
10
Case studies, end to end
PlannedDifficulty
Priority
- URL shortener, end to endMediumP3Talk about it
- Rate limiter, end to endHardP3Talk about it
- Notification service, end to endHardP3Talk about it
- S3 upload/processing service, end to endHardP2Talk about it
- Job scheduler, end to endHardP3Talk about it
- Chat, end to endHardP3Talk about it
- News feed, end to endHardP3Talk about it
- Metrics/logging system, end to endHardP3Talk about it
- Multi-tenant ETL platform, end to endHardP2Talk about it
11
Security & operability in design + the review checklist
PlannedDifficulty
Priority
- Tenant isolation and PII boundariesHardP3Talk about it
- Place authn/authz at every trust boundaryMediumP2Know cold
- Follow sensitive data through its lifecycleMediumP2Talk about it
- Cost as a design constraintMediumP4Talk about it
- Design the operating and recovery pathMediumP2Talk about it
- The what-did-I-forget design reviewMediumP5Know it exists
No cards match these filters.