Maps and Navigation — System Design Academy

Complete System Design Tutorial

The design separates the latency-sensitive online path, the correctness-critical durable transition, and asynchronous projection or repair work. Each chapter below explains one decision shown in the diagrams.

1. Requirements and scope

Functional requirements

Execute the primary maps and navigation workflow and return an observable result.
Expose the concrete operations shown below with stable retry identities.
Allow operators to inspect, disable, repair, or replay ambiguous state.
Keep derived analytics and notifications off the correctness-critical path.

Non-functional requirements

Design for 1B route requests/day · continental graph · p99 < 1 s.
Protect the named business invariant under concurrency and partial failure.
Scale by an explicit aggregate/partition key without one global mutex.
Degrade intentionally when cache, queue, store, or regional ownership is unavailable.

Primary invariant: each accepted maps and navigation operation has one stable result and retries do not repeat its business effect.

↑ top

2. Capacity assumptions

State assumptions before choosing technology. The supplied workload is 1B route requests/day · continental graph · p99 < 1 s. Convert it into peak throughput, retained records, bandwidth, and failure budget; then connect each number to a component.

Dimension	Working estimate	Design implication
Traffic	1B route requests/day · continental graph · p99 < 1 s	Separate online compute, durable state, and asynchronous consumers.
Peak factor	5–10× average	Admission control, autoscaling headroom, and hot-key protection.
Stored state	write rate × retention × record bytes	Partition count, indexes, compaction, backup, and archival tier.
Network	peak operations × payload size	Batching, compression, edge placement, and egress budget.
Availability	99.9–99.99%	Multi-zone replicas, tested failover, and dependency error budgets.
Recovery	RPO/RTO by invariant	Ack boundary, replay retention, reconciliation, and operator runbooks.

Do not cargo-cult the numbers: show the arithmetic, name the uncertain variable, and say which production metric will replace the interview estimate.

↑ top

3. API and data model

GET /v1/geocode?q=&bbox=

POST /v1/routeswaypoints/mode

GET /v1/tiles/{z}/{x}/{y}

POST /v1/trafficsigned observations

Authoritative records and indexes

road_edge(edge_id, from, to, geometry, restrictions)

graph_partition(id, boundary_nodes, version, uri)

traffic(edge_id, speed, confidence, observed_at)

route(request_id, graph_version, path_edges, eta)

Command and event envelopes

# Command envelope
operation: GET /v1/geocode?q=&bbox=
request_id: req_01J...
idempotency_key: idem_7f2...
aggregate_key: maps-navigation_123
expected_version: 17
payload: {...}

# Durable event
event_id: evt_01J...
aggregate_key: maps-navigation_123
event_type: maps_navigation_committed
schema_version: 3
trace_id: tr_8ac...

The authoritative aggregate and idempotency record share a transactional or conditional-write boundary. Event streams, search indexes, caches, analytics, and dashboards are rebuildable projections.

↑ top

4. Why this architecture and what it costs

The boxes are not a shopping list. Each boundary exists because it owns a different invariant, scaling dimension, data lifecycle, security policy, or failure domain. The table makes those choices explicit and shows when the design should remain simpler.

Component	Owned responsibility	Why separate it	Rejected alternative	Cost and evolution trigger
Maps Client	Owns client protocol, request identity, retry, and resume behavior. In this design it participates in: Geocode endpoints and snap waypoints to routable graph edges.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents the external caller from also owning the contract and failure behavior of Tile CDN.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Road closure arrives during route” to the external caller.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Maps Client boundary before extracting or scaling it.
Tile CDN	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Choose hierarchy/landmarks and route through relevant graph partitions.	Removes repeated work from the latency-critical path. It prevents Maps Client from also owning the contract and failure behavior of Geocoding API.	Read authoritative storage for every request. Simpler initially, but latency and source-of-truth load scale together and a hot key becomes a database incident. Here that simpler choice would couple “Long route crosses many partitions” to Maps Client.	Costs invalidation, staleness, stampede protection, and memory. Add or split it only after the workload (1B route requests/day · continental graph · p99 < 1 s) proves repeated reads dominate. Measure the latency, saturation, lag, or mismatch at the Tile CDN boundary before extracting or scaling it.
Geocoding API	Owns authentication, validation, quotas, deadlines, and routing. In this design it participates in: Overlay fresh traffic, closures, turn restrictions, and mode-specific costs.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Tile CDN from also owning the contract and failure behavior of Place Index.	Put protocol, authentication, quotas, and business transitions in one process. Acceptable at low scale, but it couples external traffic risk to domain releases. Here that simpler choice would couple “Traffic data is sparse or malicious” to Tile CDN.	Costs another hop, policy distribution, and a high-blast-radius fleet. Introduce it when many clients or services need one protocol and protection boundary. Measure the latency, saturation, lag, or mismatch at the Geocoding API boundary before extracting or scaling it.
Place Index	Serves a low-latency, disposable projection with authoritative fallback. In this design it participates in: Compute alternatives, ETA, instructions, and return graph/traffic versions.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Geocoding API from also owning the contract and failure behavior of Route API.	Read authoritative storage for every request. Simpler initially, but latency and source-of-truth load scale together and a hot key becomes a database incident. Here that simpler choice would couple “Graph versions differ at boundaries” to Geocoding API.	Costs invalidation, staleness, stampede protection, and memory. Add or split it only after the workload (1B route requests/day · continental graph · p99 < 1 s) proves repeated reads dominate. Measure the latency, saturation, lag, or mismatch at the Place Index boundary before extracting or scaling it.
Route API	Owns authentication, validation, quotas, deadlines, and routing. In this design it participates in: Aggregate privacy-safe telemetry to update traffic and schedule base graph rebuilds.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Place Index from also owning the contract and failure behavior of Graph Partition Router.	Put protocol, authentication, quotas, and business transitions in one process. Acceptable at low scale, but it couples external traffic risk to domain releases. Here that simpler choice would couple “Offline client uses stale restrictions” to Place Index.	Costs another hop, policy distribution, and a high-blast-radius fleet. Introduce it when many clients or services need one protocol and protection boundary. Measure the latency, saturation, lag, or mismatch at the Route API boundary before extracting or scaling it.
Graph Partition Router	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Geocode endpoints and snap waypoints to routable graph edges.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Route API from also owning the contract and failure behavior of Road Graph Shards.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Road closure arrives during route” to Route API.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Graph Partition Router boundary before extracting or scaling it.
Road Graph Shards	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Choose hierarchy/landmarks and route through relevant graph partitions.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Graph Partition Router from also owning the contract and failure behavior of Hierarchy / Landmarks.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Long route crosses many partitions” to Graph Partition Router.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Road Graph Shards boundary before extracting or scaling it.
Hierarchy / Landmarks	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Overlay fresh traffic, closures, turn restrictions, and mode-specific costs.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Road Graph Shards from also owning the contract and failure behavior of Traffic Overlay.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Traffic data is sparse or malicious” to Road Graph Shards.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Hierarchy / Landmarks boundary before extracting or scaling it.
Traffic Overlay	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Compute alternatives, ETA, instructions, and return graph/traffic versions.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Hierarchy / Landmarks from also owning the contract and failure behavior of Closure Feed.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Graph versions differ at boundaries” to Hierarchy / Landmarks.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Traffic Overlay boundary before extracting or scaling it.
Closure Feed	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Aggregate privacy-safe telemetry to update traffic and schedule base graph rebuilds.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Traffic Overlay from also owning the contract and failure behavior of Route Worker.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Offline client uses stale restrictions” to Traffic Overlay.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Closure Feed boundary before extracting or scaling it.
Route Worker	Executes idempotent asynchronous work under lease and retry policy. In this design it participates in: Geocode endpoints and snap waypoints to routable graph edges.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Closure Feed from also owning the contract and failure behavior of ETA Model.	Perform the work before acknowledging the caller. This gives simpler consistency but unbounded tail latency and fragile retries. Here that simpler choice would couple “Road closure arrives during route” to Closure Feed.	Costs leases, checkpoints, retries, DLQs, and lag operations. Separate once execution time or burstiness no longer fits the request deadline. Measure the latency, saturation, lag, or mismatch at the Route Worker boundary before extracting or scaling it.
ETA Model	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Choose hierarchy/landmarks and route through relevant graph partitions.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Route Worker from also owning the contract and failure behavior of Alternative Router.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Long route crosses many partitions” to Route Worker.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the ETA Model boundary before extracting or scaling it.
Alternative Router	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Overlay fresh traffic, closures, turn restrictions, and mode-specific costs.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents ETA Model from also owning the contract and failure behavior of Offline Bundle Builder.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Traffic data is sparse or malicious” to ETA Model.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Alternative Router boundary before extracting or scaling it.
Offline Bundle Builder	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Compute alternatives, ETA, instructions, and return graph/traffic versions.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Alternative Router from also owning the contract and failure behavior of Telemetry / Map Feedback.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Graph versions differ at boundaries” to Alternative Router.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Offline Bundle Builder boundary before extracting or scaling it.
Telemetry / Map Feedback	Owns authoritative versioned state or immutable evidence. In this design it participates in: Aggregate privacy-safe telemetry to update traffic and schedule base graph rebuilds.	Defines the source of truth and commit boundary. It prevents Offline Bundle Builder from also owning the contract and failure behavior of operators and repair tooling.	Let each service persist its own interpretation. That removes one clear commit boundary and makes retries, support, and reconciliation ambiguous. Here that simpler choice would couple “Offline client uses stale restrictions” to Offline Bundle Builder.	Costs schema migration, replication, backup, and shard ownership. Split storage only when access patterns, retention, or compliance boundaries materially differ. Measure the latency, saturation, lag, or mismatch at the Telemetry / Map Feedback boundary before extracting or scaling it.

Decision 01

Chosen boundary vs simpler monolith

Map tiles are immutable/versioned CDN objects; route computation uses a compact road graph and cannot scan raw geographic records per request.

A single deployable with one transactional store is the right starting point for a small maps and navigation workload. The proposed boundary becomes justified when map tiles are immutable/versioned cdn objects; route computation uses a compact road graph and cannot scan raw geographic records per request. Separating it contains the named failure “Road closure arrives during route” without placing unrelated work on the critical path.

Decision 02

Correctness vs availability

Precompute contraction hierarchies or landmarks for fast routes, then overlay recent closures and traffic without rebuilding the base graph.

The design coordinates only where each accepted maps and navigation operation has one stable result and retries do not repeat its business effect. Everywhere else it accepts bounded staleness or asynchronous propagation. The alternative—strong coordination for every projection—simplifies the mental model but increases tail latency and turns partitions into broad write unavailability.

Decision 03

Operational cost and evolution

Partition geographically with boundary-node exchange; long routes use hierarchical planning to avoid fan-out across every shard.

This choice adds dashboards, ownership, schema compatibility, capacity planning, and recovery runbooks. Start with the minimum topology that supports “Overlay fresh traffic, closures, turn restrictions, and mode-specific costs.” and extract another service only when measured throughput, blast radius, data lifecycle, or team ownership changes independently.

Why this design: the synchronous path contains only the decisions required to protect each accepted maps and navigation operation has one stable result and retries do not repeat its business effect. Durable projections, notifications, indexing, analytics, and repair are isolated so their load or failure cannot corrupt the acknowledged operation.

Correctness boundary

Map tiles are immutable/versioned CDN objects; route computation uses a compact road graph and cannot scan raw geographic records per request.

Scale boundary

Precompute contraction hierarchies or landmarks for fast routes, then overlay recent closures and traffic without rebuilding the base graph.

Operational boundary

Partition geographically with boundary-node exchange; long routes use hierarchical planning to avoid fan-out across every shard.

↑ top

5. Critical flow and algorithm

Geocode endpoints and snap waypoints to routable graph edges.

Choose hierarchy/landmarks and route through relevant graph partitions.

Overlay fresh traffic, closures, turn restrictions, and mode-specific costs.

Compute alternatives, ETA, instructions, and return graph/traffic versions.

Aggregate privacy-safe telemetry to update traffic and schedule base graph rebuilds.

Retry-safe mutation skeleton

identity = authenticate(request)
prior = idempotency.get(identity, request.key)
if prior: return prior.result

current = authoritative_store.read(partition_key)
next_state, events = domain.transition(current, request)
authoritative_store.compare_and_set(current.version, next_state)
idempotency.commit(request.key, next_state.result)
outbox.append(events)
return next_state.result

Exactly-once business effect: the network can duplicate requests. Stable identities, conditional state transitions, immutable event IDs, and reconciliation make the user-visible effect happen once.

↑ top

6. Caching, hot keys, and backpressure

Cache policy

Cache immutable or slowly changing reads with versioned keys, TTL jitter, short negative entries, and request coalescing.

Hot partition or tenant

Measure per-key and per-tenant skew. Replicate safe reads, isolate the offender, salt only when ordering permits, and enforce budgets.

Backpressure

Bound queues, buffers, connection pools, and concurrency. Shed low-priority work and scale from lag/oldest-item age rather than CPU alone.

Stampede and retry storms

Use single-flight/coalescing, jittered expiry, bounded retries, exponential backoff, and circuit breakers so one dependency failure does not multiply traffic.

Cache outage rule: never allow every miss to fall through simultaneously. Use local protection, request coalescing, circuit breakers, bounded pools, and controlled fallback.

↑ top

7. Partitioning and consistency

partition = stable_hash(primary_or_aggregate_key) → shard owner
write = conditional_update(expected_version)
read_projection = versioned and rebuildable

Strong boundary

each accepted maps and navigation operation has one stable result and retries do not repeat its business effect. Use a uniqueness constraint, serializable/conditional write, expected version, or fencing token at this boundary.

Eventual boundary

Analytics, notifications, search, read models, and operational dashboards may lag when freshness is observable and replay exists.

Operation	Consistency	Reason
Authoritative mutation	Conditional / serializable	Protect uniqueness, ownership, money, inventory, order, or legal state transition.
Read after own write	Primary, session token, or cache warm	The caller must observe the result it just committed.
Ordinary read projection	Bounded stale where safe	Replicas and caches improve latency when version/freshness is visible.
Analytics and operations	Eventual + replay	Freshness is less important than isolating the online path.

Route every mutation for one aggregate to one owner/partition epoch.
Use cursors rather than offsets for changing ordered collections.
Retain tombstones until replication and repair windows close.
Move shard ownership through versioned maps and dual-read verification.

↑ top

8. Failure drills and recovery

Failure	Runtime behavior	Recovery proof
Road closure arrives during route	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Long route crosses many partitions	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Traffic data is sparse or malicious	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Graph versions differ at boundaries	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Offline client uses stale restrictions	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.

Recovery checkpoint

# Recovery checkpoint
scenario: Road closure arrives during route
operation_id: op_01J...
aggregate_key: maps-navigation_123
last_durable_version: 18
projection_checkpoint: partition-07@offset-482901
retry_policy: same_identity / bounded_backoff
reconciliation_status: pending
owner: Telemetry / Map Feedback

Every dependency needs a timeout, retry budget, circuit policy, and explicit fail-open, fail-closed, stale, or read-only mode. “Highly available” is incomplete until degraded behavior is named.

↑ top

9. Security, regions, and observability

Security and abuse

Authenticate at ingress and authorize the exact resource/tenant on every operation.
Encrypt transport and durable data; tokenize or isolate sensitive material.
Rate-limit by tenant, actor, resource, and network; detect enumeration and anomalous fan-out.
Redact secrets and personal data before logs or analytical storage.
Audit privileged mutations, support access, policy changes, and forced recovery.

Signals that prove correctness

User-visible success rate and p50/p95/p99 latency.
Timeouts, circuit state, saturation, and replica health by dependency.
Queue lag, oldest-item age, retries, and dead-letter growth.
Conflict, duplicate-suppression, and reconciliation mismatch counts.
Business drill: Road closure arrives during route — alert before the user-visible invariant is violated.

Regional evolution

Begin multi-zone with tested failover, backups, and reconciliation.
Add global stateless ingress and regional read projections.
Keep one write owner per aggregate or partition where coordination is expensive.
Move ownership with epochs/fencing; avoid silent simultaneous writers.
Use expand/contract schemas, canaries, shadow reads, and reversible cutovers.

↑ top

10. Interviewer follow-ups

How do traffic updates affect routes?

Map tiles are immutable/versioned CDN objects; route computation uses a compact road graph and cannot scan raw geographic records per request. Walk the concrete lifecycle step “Geocode endpoints and snap waypoints to routable graph edges.” Then test it against “Road closure arrives during route”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How are map tiles generated?

Precompute contraction hierarchies or landmarks for fast routes, then overlay recent closures and traffic without rebuilding the base graph. Walk the concrete lifecycle step “Choose hierarchy/landmarks and route through relevant graph partitions.” Then test it against “Long route crosses many partitions”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do you route across partitions?

Partition geographically with boundary-node exchange; long routes use hierarchical planning to avoid fan-out across every shard. Walk the concrete lifecycle step “Overlay fresh traffic, closures, turn restrictions, and mode-specific costs.” Then test it against “Traffic data is sparse or malicious”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How does offline navigation work?

Map tiles are immutable/versioned CDN objects; route computation uses a compact road graph and cannot scan raw geographic records per request. Walk the concrete lifecycle step “Compute alternatives, ETA, instructions, and return graph/traffic versions.” Then test it against “Graph versions differ at boundaries”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do you support multiple travel modes?

Precompute contraction hierarchies or landmarks for fast routes, then overlay recent closures and traffic without rebuilding the base graph. Walk the concrete lifecycle step “Aggregate privacy-safe telemetry to update traffic and schedule base graph rebuilds.” Then test it against “Offline client uses stale restrictions”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

Answer pattern: state the invariant, choose a mechanism, walk one failure, name the recovery signal, and explain the cost of the alternative.

↑ top

11. How to present this in 45 minutes

Time	What to cover
0–5 min	Clarify scope, workload, SLOs, and the correctness invariant.
5–10 min	Estimate throughput/storage and define concrete APIs and records.
10–20 min	Draw online, durable-write, and asynchronous/repair paths.
20–30 min	Explain the exact lifecycle, partition key, and concurrency control.
30–38 min	Discuss caching, hot keys, replication, and backpressure.
38–43 min	Run failure drills, security, observability, and regional evolution.
43–45 min	Summarize the five most important choices and trade-offs.

Final decision summary

Invariant	each accepted maps and navigation operation has one stable result and retries do not repeat its business effect.
Partitioning	Stable aggregate/key hash with versioned ownership and measured hot-key isolation.
Commit	Expected version or uniqueness at the authoritative state boundary.
Async work	Durable outbox/log, idempotent consumers, checkpoints, and replay.
Recovery	Same operation identity, durable truth lookup, and business reconciliation.

Closing summary

CLOSING “The design protects each accepted maps and navigation operation has one stable result and retries do not repeat its business effect. The online path stays bounded, authoritative state is versioned and partitioned, and asynchronous work is replayable and observable.”

↑ top

Maps Client

Tile CDN

Geocoding API

Place Index

Complete System Design Tutorial

1. Requirements and scope

Functional requirements

Non-functional requirements

2. Capacity assumptions

3. API and data model

Authoritative records and indexes

Command and event envelopes

4. Why this architecture and what it costs

Chosen boundary vs simpler monolith

Correctness vs availability

Operational cost and evolution

5. Critical flow and algorithm

Retry-safe mutation skeleton

6. Caching, hot keys, and backpressure

7. Partitioning and consistency

Strong boundary

Eventual boundary

8. Failure drills and recovery

Recovery checkpoint

9. Security, regions, and observability

Security and abuse

Signals that prove correctness

Regional evolution

10. Interviewer follow-ups

11. How to present this in 45 minutes

Final decision summary

Closing summary