Systems Atlas
Master system design from first principles to production.
From foundational mental models to principal-level architectural judgment, for senior, staff, and principal engineers. 326 topics and 32 system dissections, each with mechanisms, capacity math, failure analysis, runnable examples, and interview preparation.
Written and validated so far: 10 of 326 topics, 0 of 32 system dissections. Items without a link are planned and not yet written.
0%Overall completion
0 hEstimated study time remaining
A. Design method and quantitative reasoning
- Functional and non-functional requirements
- Constraints, assumptions, and architecture drivers
- Workload characterization and traffic modeling
- Latency
- Throughput
- P99 latency
- Tail latency
- Back-of-the-envelope capacity estimation
- Latency budgets and critical-path analysis
- Concurrency, utilization, saturation, and bottlenecks
- Storage, bandwidth, memory, and IOPS calculations
- Little’s Law and practical queueing theory
- Availability and reliability mathematics
- Failure-budget and risk calculations
- Architecture fitness functions
- Decision criteria and trade-off matrices
- Architecture decision records
- Reversible versus irreversible decisions
B. Computer systems
- Processes, threads, coroutines, and event loops
- Blocking versus non-blocking I/O
- Async I/O and scheduling
- File descriptors, sockets, and resource limits
- System calls and context switching
- CPU caches, locality, false sharing, and NUMA
- Serialization, compression, and zero-copy
- Race conditions
- Thread safety
- Deadlocks
- Atomicity, visibility, and memory ordering
- Livelock, starvation, and lock-free programming
- Memory leaks
- Garbage collection
- CPU, heap, lock-contention, and I/O profiling
C. Networking and web foundations
- TCP versus UDP
- DNS
- HTTP semantics, methods, status codes, and headers
- HTTP caching and conditional requests
- Connection establishment, keep-alive, and draining
- Congestion control and head-of-line blocking
- NAT, firewalls, routing, anycast, and BGP fundamentals
- Packet loss, jitter, retransmission, and MTU problems
- TLS
- Certificate authorities, PKI, and mutual TLS
- HTTP/2 and HTTP/3
- Reverse proxies
- Forward proxies, tunnels, ingress, and egress
- Load balancing
- L4 versus L7 load balancing
- Gateways and routers
- CORS
- WebSockets
- Server-Sent Events
- Long polling
- CDNs
- Edge caching
- Origin shielding and CDN request collapsing
D. API and contract design
- REST resource modeling
- Pagination: offset, cursor, and keyset
- Filtering, sorting, sparse fields, and bulk operations
- API error models and partial success
- Idempotency keys
- Deadlines and cancellation propagation
- Long-running operation APIs
- Webhooks
- Webhook replay protection and signature verification
- API gateways
- Backend-for-frontend architecture
- GraphQL challenges
- GraphQL federation and query-complexity controls
- gRPC
- Protocol Buffers compatibility
- API versioning
- Semantic versioning
- Compatibility, deprecation, and sunset policies
- Consumer-driven contract testing
E. Data systems and storage internals
- Relational modeling and normalization
- Database indexing
- Query optimization
- N+1 queries
- Connection pooling
- Transactions and ACID
- Isolation levels and transaction anomalies
- MVCC and serializability
- Optimistic locking
- Pessimistic concurrency control
- Composite, covering, and secondary indexes
- Denormalization and materialized views
- Access-pattern-driven NoSQL modeling
- B-trees and LSM trees
- Write-ahead logs, checkpoints, and recovery
- Buffer pools, pages, and storage layout
- Compaction and read/write amplification
- Read replicas
- Replication
- Partitioning
- Sharding
- Consistent and rendezvous hashing
- Hot keys, hot partitions, and rebalancing
- Database migrations
- Schema versioning
- Online schema changes and large backfills
- Change data capture
- Search and inverted indexes
- Time-series databases
- Graph databases
- Vector databases and approximate-nearest-neighbor indexes
- Columnar and analytical storage
- Blob and object storage
- Polyglot persistence and data ownership
F. Distributed systems
- Partial failure and failure detection
- Network partitions
- CAP theorem
- PACELC
- Eventual consistency
- Linearizable, sequential, causal, and eventual consistency
- Read-your-writes and monotonic-read guarantees
- Read/write quorums
- Tunable consistency
- Clock skew
- Lamport clocks
- Vector clocks and version vectors
- Hybrid logical clocks
- CRDTs and conflict resolution
- Gossip, membership, and anti-entropy
- Leader election
- Raft and Paxos concepts
- Consensus safety and liveness
- Leases, fencing tokens, and split brain
- Distributed locks
- Distributed transactions
- Two-phase commit
- Atomic commit and blocking failure modes
- Exactly-once claims versus end-to-end effects
- Invariant-first distributed-system design
G. Messaging, streaming, and workflows
- Message queues
- Publish/subscribe
- Delivery semantics
- Ordering and partition keys
- Consumer groups and parallel processing
- Acknowledgements and visibility timeouts
- Dead-letter queues
- Poison messages, quarantine, and replay
- Event-driven architecture
- Transactional outbox and inbox patterns
- Dual-write failure and reconciliation
- Event schemas and schema registries
- Event notification versus event-carried state transfer
- Event sourcing
- CQRS
- Stream-processing windows
- Watermarks and late-arriving events
- Cron jobs
- Durable timers and duplicate execution
- Saga patterns
- Workflow engines and durable execution
- Orchestration versus choreography
H. Performance and resilience
- Timeouts
- Retries
- Exponential backoff
- Retry budgets
- Circuit breakers
- Bulkheads and isolation pools
- Rate limiting
- Backpressure
- Load shedding and admission control
- Graceful degradation and brownouts
- Caching strategies
- Cache invalidation
- Cache stampedes and request coalescing
- Cache admission and eviction algorithms
- Probabilistic data structures
- Batching, pipelining, and parallelism
- Load, stress, spike, and soak testing
- Benchmark design and coordinated omission
- Health checks
- Graceful shutdown and connection draining
- Horizontal scaling
- Vertical scaling
- Autoscaling
- Cold starts
- Serverless limits
- Blast-radius containment
- Single points of failure and correlated failure
- Safe defaults, kill switches, and operational guardrails
- Failure-mode and effects analysis
- Chaos engineering
I. Observability and operations
- Logging
- Structured logging and correlation identifiers
- Metrics
- Metric cardinality and histogram design
- Distributed tracing
- Trace sampling and context propagation
- OpenTelemetry and telemetry pipelines
- Monitoring
- Observability
- Golden Signals, RED, and USE methods
- SLAs, SLOs, and SLIs
- Error budgets
- Alerts
- Multi-window burn-rate alerts
- Alert fatigue and symptom-based alerting
- Production incidents
- Incident command and incident communication
- On-call engineering
- Runbook design and operational ownership
- Sustainable on-call rotations
- Postmortems
- Production-readiness reviews
J. Security, privacy, and multi-tenancy
- Threat modeling and abuse cases
- Authentication versus authorization
- Identity and access management
- RBAC versus ABAC
- Least privilege and separation of duties
- OAuth
- OpenID Connect and PKCE
- JWT rotation
- Workload identity
- Secrets management
- SQL injection
- Cross-site scripting
- Cross-site request forgery
- Server-side request forgery
- Request smuggling and insecure deserialization
- Encryption in transit
- Encryption at rest
- Envelope encryption
- KMS, HSMs, and crypto agility
- Web application firewalls
- DDoS protection
- Tenant isolation and noisy-neighbor controls
- Supply-chain security, SBOMs, and provenance
- Audit logs and tamper evidence
- Data classification, retention, and deletion
- Data residency and sovereignty
- Privacy-by-design
- Security incident containment
K. Platform and delivery engineering
- Docker
- Container image layers and multi-stage builds
- Kubernetes
- Helm charts
- Kubernetes controllers and reconciliation
- Kubernetes requests, limits, QoS, and disruption budgets
- Liveness and readiness probes
- Persistent storage in Kubernetes
- Service discovery
- Service mesh architecture and trade-offs
- CI/CD
- Build caching
- Reproducible and hermetic builds
- Dependency management and dependency hell
- Infrastructure as code
- Terraform
- GitOps
- Policy as code and admission control
- Artifact promotion and environment parity
- Feature flags
- Rolling deployments
- Blue-green deployments
- Canary releases
- Progressive delivery and automated canary analysis
- Rollbacks and roll-forwards
- Platform engineering and paved roads
- Developer experience as a platform concern
L. Multi-region and continuity
- Regions, availability zones, and failure domains
- Multi-region deployments
- Active-passive versus active-active architecture
- Regional traffic steering
- Cross-region replication and write ownership
- Failover and failback
- RTO and RPO
- Business-impact analysis
- Backups and restoration
- Immutable backups and cyber recovery
- Restore verification
- Disaster recovery
- Regional evacuation and capacity reservation
- Control-plane versus data-plane survivability
- Cell-based architecture
- Disaster-recovery exercises
M. Architecture and evolution
- Cohesion, coupling, modularity, and information hiding
- Modular monoliths
- Microservice boundaries and service sizing
- Hexagonal and ports-and-adapters architecture
- Domain-driven design and bounded contexts
- Synchronous versus asynchronous composition
- Strangler migrations
- Branch by abstraction
- Dual reads, dual writes, and migration reconciliation
- Anti-corruption layers
- Multi-tenancy architecture models
- Control-plane and data-plane separation
- Evolutionary architecture
- Technical-debt portfolio management
- Build versus buy versus managed services
N. Testing and correctness
- Test portfolios and confidence boundaries
- Unit, integration, component, and end-to-end testing
- Contract and compatibility testing
- Property-based testing
- Fuzz testing
- Invariant testing
- Deterministic distributed-system simulation
- Concurrency and race testing
- Fault-injection testing
- Shadow traffic and traffic replay
- Differential testing
- Migration and rollback testing
- Restore and disaster-recovery testing
- Formal methods and model checking
- Production verification
O. Economics and principal-level leadership
- Cloud unit economics
- Cost per request, tenant, and workload
- Cost optimization
- Total cost of ownership
- Operational burden and opportunity cost
- Capacity commitments and elasticity
- Architecture narratives and review documents
- Executive communication
- Cross-team contracts and ownership boundaries
- Dependency governance
- Technical strategy and sequencing
- Migration economics
- Standardization versus autonomy
- Risk management and compliance evidence
- Decision facilitation, dissent, and escalation
- Mentoring architecture reasoning
- Measuring architecture outcomes
Systems dissected
- URL shortener / Bitly
- Cloud file storage / Dropbox
- Local delivery service
- Ticket booking / Ticketmaster
- Social news feed
- Dating and matching service / Tinder
- Online judge / LeetCode
- Messaging platform / WhatsApp
- Distributed rate limiter
- Video platform / YouTube
- Live comments
- Top-K video analytics
- Ride-sharing platform / Uber
- Distributed web crawler
- Advertisement click aggregator
- Social-post search
- Local business discovery / Yelp
- Photo-sharing platform / Instagram
- Fitness activity platform / Strava
- Distributed cache
- Online auction
- Distributed job scheduler
- News aggregator
- Price-tracking service
- Notification system
- Retail brokerage platform / Robinhood
- Collaborative document editor / Google Docs
- Payment system
- Metrics-monitoring platform
- Online chess platform
- LLM conversational platform / ChatGPT
- Flash-sale platform
Cross-cutting patterns (not yet written) · All system dissections · Traceability matrix