Vending Machine — 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 vending machine 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 Dozens of slots · coins/cards · offline operation.
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: every object transition is legal and atomic at one aggregate boundary.

↑ top

2. Capacity assumptions

State assumptions before choosing technology. The supplied workload is Dozens of slots · coins/cards · offline operation. Convert it into peak throughput, retained records, bandwidth, and failure budget; then connect each number to a component.

Dimension	Working estimate	Design implication
Traffic	Dozens of slots · coins/cards · offline operation	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

INSERT_MONEY(value) / TAP_CARD(token)

SELECT(slot)validate stock/price

DISPENSE(slot, vend_id)

REFUND(session) / COMPLETE(vend_id)

Authoritative records and indexes

slot(slot_id PK, product_id, price, count, version)

cash_bin(denomination, count, version)

vend(vend_id, slot_id, amount, payment_state, dispense_state)

fault(fault_id, component, code, observed_at)

Command and event envelopes

# Command envelope
operation: INSERT_MONEY(value) / TAP_CARD(token)
request_id: req_01J...
idempotency_key: idem_7f2...
aggregate_key: vending-machine_123
expected_version: 17
payload: {...}

# Durable event
event_id: evt_01J...
aggregate_key: vending-machine_123
event_type: vending_machine_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
Display / Keypad	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Accept money into a session and maintain refundable balance separately from machine cash.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents the external caller from also owning the contract and failure behavior of Machine Controller.	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 “Motor runs but item does not drop” 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 Display / Keypad boundary before extracting or scaling it.
Machine Controller	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Validate selected slot inventory and current price; calculate feasible change from actual coins.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Display / Keypad from also owning the contract and failure behavior of Session State.	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 “Exact change unavailable after payment” to Display / Keypad.	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 Machine Controller boundary before extracting or scaling it.
Session State	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Authorize card or reserve cash, then claim one slot item with expected version.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Machine Controller from also owning the contract and failure behavior of Product Catalog.	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 “Two inputs race one machine session” to Machine Controller.	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 Session State boundary before extracting or scaling it.
Product Catalog	Preserves durable ordering, replay, and consumer checkpoints. In this design it participates in: Command dispenser and confirm drop sensor before committing inventory/payment.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Session State from also owning the contract and failure behavior of Slot Inventory.	Call every downstream consumer synchronously. That couples user latency and availability to analytics, notification, indexing, and audit dependencies. Here that simpler choice would couple “Card authorization expires” to Session State.	Costs eventual consistency, schema evolution, replay tooling, idempotent consumers, and lag operations. Earn it when bursts or downstream failure must not block acknowledgment. Measure the latency, saturation, lag, or mismatch at the Product Catalog boundary before extracting or scaling it.
Slot Inventory	Owns authoritative versioned state or immutable evidence. In this design it participates in: On jam/timeout refund or mark uncertain vend for operator reconciliation.	Defines the source of truth and commit boundary. It prevents Product Catalog from also owning the contract and failure behavior of Selection Policy.	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 “Refill changes inventory during vend” to Product Catalog.	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 Slot Inventory boundary before extracting or scaling it.
Selection Policy	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Accept money into a session and maintain refundable balance separately from machine cash.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Slot Inventory from also owning the contract and failure behavior of Coin Acceptor.	Embed policy in every caller. Rules then drift by deployment version and emergency changes require fleet-wide redeployment. Here that simpler choice would couple “Motor runs but item does not drop” to Slot Inventory.	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 Selection Policy boundary before extracting or scaling it.
Coin Acceptor	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Validate selected slot inventory and current price; calculate feasible change from actual coins.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Selection Policy from also owning the contract and failure behavior of Cash Inventory.	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 “Exact change unavailable after payment” to Selection Policy.	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 Coin Acceptor boundary before extracting or scaling it.
Cash Inventory	Owns authoritative versioned state or immutable evidence. In this design it participates in: Authorize card or reserve cash, then claim one slot item with expected version.	Defines the source of truth and commit boundary. It prevents Coin Acceptor from also owning the contract and failure behavior of Card Terminal.	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 “Two inputs race one machine session” to Coin Acceptor.	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 Cash Inventory boundary before extracting or scaling it.
Card Terminal	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Command dispenser and confirm drop sensor before committing inventory/payment.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Cash Inventory from also owning the contract and failure behavior of Payment Strategy.	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 “Card authorization expires” to Cash Inventory.	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 Card Terminal boundary before extracting or scaling it.
Payment Strategy	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: On jam/timeout refund or mark uncertain vend for operator reconciliation.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Card Terminal from also owning the contract and failure behavior of Change Solver.	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 “Refill changes inventory during vend” to Card Terminal.	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 Payment Strategy boundary before extracting or scaling it.
Change Solver	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Accept money into a session and maintain refundable balance separately from machine cash.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Payment Strategy from also owning the contract and failure behavior of Dispenser Motor.	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 “Motor runs but item does not drop” to Payment Strategy.	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 Change Solver boundary before extracting or scaling it.
Dispenser Motor	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Validate selected slot inventory and current price; calculate feasible change from actual coins.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Change Solver from also owning the contract and failure behavior of Drop Sensor.	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 “Exact change unavailable after payment” to Change Solver.	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 Dispenser Motor boundary before extracting or scaling it.
Drop Sensor	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Authorize card or reserve cash, then claim one slot item with expected version.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Dispenser Motor from also owning the contract and failure behavior of Fault / Refill Console.	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 “Two inputs race one machine session” to Dispenser Motor.	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 Drop Sensor boundary before extracting or scaling it.
Fault / Refill Console	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Command dispenser and confirm drop sensor before committing inventory/payment.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Drop Sensor from also owning the contract and failure behavior of operators and repair tooling.	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 “Card authorization expires” to Drop Sensor.	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 Fault / Refill Console boundary before extracting or scaling it.

Decision 01

Chosen boundary vs simpler monolith

Use a state machine for idle, collecting, selected, dispensing, refunding, and out-of-service.

A single deployable with one transactional store is the right starting point for a small vending machine workload. The proposed boundary becomes justified when use a state machine for idle, collecting, selected, dispensing, refunding, and out-of-service. Separating it contains the named failure “Motor runs but item does not drop” without placing unrelated work on the critical path.

Decision 02

Correctness vs availability

Abstract payment and change-making strategies; calculate change from actual coin inventory.

The design coordinates only where every object transition is legal and atomic at one aggregate boundary. 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

Commit inventory only when dispensing succeeds, and expose reconciliation for jams or power loss.

This choice adds dashboards, ownership, schema compatibility, capacity planning, and recovery runbooks. Start with the minimum topology that supports “Authorize card or reserve cash, then claim one slot item with expected version.” 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 every object transition is legal and atomic at one aggregate boundary. Durable projections, notifications, indexing, analytics, and repair are isolated so their load or failure cannot corrupt the acknowledged operation.

Correctness boundary

Use a state machine for idle, collecting, selected, dispensing, refunding, and out-of-service.

Scale boundary

Abstract payment and change-making strategies; calculate change from actual coin inventory.

Operational boundary

Commit inventory only when dispensing succeeds, and expose reconciliation for jams or power loss.

↑ top

5. Critical flow and algorithm

Accept money into a session and maintain refundable balance separately from machine cash.

Validate selected slot inventory and current price; calculate feasible change from actual coins.

Authorize card or reserve cash, then claim one slot item with expected version.

Command dispenser and confirm drop sensor before committing inventory/payment.

On jam/timeout refund or mark uncertain vend for operator reconciliation.

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

Add a cache only for a measured repeated read; do not place mutable correctness state in a disposable layer.

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

every object transition is legal and atomic at one aggregate boundary. 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
Motor runs but item does not drop	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.
Exact change unavailable after payment	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.
Two inputs race one machine session	Resolve at source of truth with uniqueness, expected version, lease generation, or monotonic cursor.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Card authorization expires	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.
Refill changes inventory during vend	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: Motor runs but item does not drop
operation_id: op_01J...
aggregate_key: vending-machine_123
last_durable_version: 18
projection_checkpoint: partition-07@offset-482901
retry_policy: same_identity / bounded_backoff
reconciliation_status: pending
owner: Fault / Refill Console

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: Motor runs but item does not drop — 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 is exact change calculated?

Use a state machine for idle, collecting, selected, dispensing, refunding, and out-of-service. Walk the concrete lifecycle step “Accept money into a session and maintain refundable balance separately from machine cash.” Then test it against “Motor runs but item does not drop”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

What if dispensing jams?

Abstract payment and change-making strategies; calculate change from actual coin inventory. Walk the concrete lifecycle step “Validate selected slot inventory and current price; calculate feasible change from actual coins.” Then test it against “Exact change unavailable after payment”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do card payments differ?

Commit inventory only when dispensing succeeds, and expose reconciliation for jams or power loss. Walk the concrete lifecycle step “Authorize card or reserve cash, then claim one slot item with expected version.” Then test it against “Two inputs race one machine session”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How does an operator refill stock?

Use a state machine for idle, collecting, selected, dispensing, refunding, and out-of-service. Walk the concrete lifecycle step “Command dispenser and confirm drop sensor before committing inventory/payment.” Then test it against “Card authorization expires”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do promotions work?

Abstract payment and change-making strategies; calculate change from actual coin inventory. Walk the concrete lifecycle step “On jam/timeout refund or mark uncertain vend for operator reconciliation.” Then test it against “Refill changes inventory during vend”. 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	every object transition is legal and atomic at one aggregate boundary.
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 every object transition is legal and atomic at one aggregate boundary. The online path stays bounded, authoritative state is versioned and partitioned, and asynchronous work is replayable and observable.”

↑ top

Display / Keypad

Machine Controller

Session State

Product Catalog

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