33 / 60 · Low-Level Design

Library Management — Production Architecture

All designs

Manage catalog items, copies, members, holds, loans, fines, and notifications.

1M titles · 50 branches · concurrent holds
1. High-level architecture and animated call flow

Choose a path, then play or step through exactly how the call moves between components.

← / → step · Space play
CLOUD / MULTI-AZ DEPLOYMENT LIBRARY MANAGEMENT · PRODUCTION TOPOLOGY 14 named components · solid = request/data dependency · warm curve = control, event, or repair path MULTI-ZONE · VERSIONED · REPLAYABLE 01 · REQUEST & CONTROL02 · DOMAIN & COORDINATION03 · DATA, ASYNC & OPERATIONSrequestauthenticateroutevalidateread / reservecommitpublishconsumeprojectobserverepairreconcileinvalidatepolicy snapshotevent streamtelemetry 01 EXTERNAL Member/Staff UI domain subsystem 02 STREAM Catalog Search durable ordered events 03 SERVICE Membership Service routing & ownership 04 SERVICE Circulation Service domain subsystem 05 SECURITY Policy Engine decision & coordination 06 STREAM Title Catalog durable ordered events 07 DATA Copy Inventory authoritative durable state 08 SERVICE Loan Aggregate domain subsystem 09 STREAM Hold Queue durable ordered events 10 SERVICE Branch Transfer domain subsystem 11 SERVICE Fine Calculator domain subsystem 12 SERVICE Payment Adapter domain subsystem 13 SERVICE NotificationService domain subsystem 14 SERVICE Digital LicenseAdapter domain subsystem externaledgeservicedatastreamsecurityops
Step 1 / 6Member / Staff UI

The caller starts the operation at Member / Staff UI with a stable request and retry identity.

  1. 01 · Member / Staff UIThe caller starts the operation at Member / Staff UI with a stable request and retry identity.
  2. 02 · Circulation ServiceValidate membership and circulation policy for the exact physical/digital copy.
  3. 03 · Policy EngineConditionally transition AVAILABLE→LOANED and create loan with computed due date.
  4. 04 · Payment AdapterA hold enqueues by title/branch policy; return assigns first eligible member atomically.
  5. 05 · Copy InventoryRenew only if policy permits and no higher-priority hold blocks the copy.
  6. 06 · Title CatalogFine calculator consumes overdue/loss transitions; notifications are asynchronous and idempotent.
2. Critical request and recovery sequence
MUTATION / COMMITMember/StaffUIACTOR 01Catalog SearchACTOR 02CirculationServiceACTOR 03Loan AggregateACTOR 04ONLINE / READ + RECOVERYMember/StaffUIACTOR 01MembershipServiceACTOR 02Title CatalogACTOR 03DigitalLicenseACTOR 04Validate membership and circulation / policy for the exactConditionally transition / AVAILABLE→LOANED and create loanA hold enqueues by title/branch / policy; return assigns firstcommitted result + eventA hold enqueues by title/branch / policy; return assigns firstRenew only if policy permits and no / higher-priority hold blocks theFine calculator consumes / overdue/loss transitions;cached/read result + cursortimeout → retry same identity → read committed truth → reconcile

Member / Staff UI

  • Owns one bounded domain decision, transition, or operational capability.
  • Partitioned by aggregate/key and guarded by version, uniqueness, or fencing.

Catalog Search

  • Preserves durable ordering, replay, and consumer checkpoints.
  • Append-only offsets and independently checkpointed consumers.

Membership Service

  • Owns one bounded domain decision, transition, or operational capability.
  • Partitioned by aggregate/key and guarded by version, uniqueness, or fencing.

Circulation Service

  • Owns one bounded domain decision, transition, or operational capability.
  • Partitioned by aggregate/key and guarded by version, uniqueness, or fencing.

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 library management 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 1M titles · 50 branches · concurrent holds.
  • 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 1M titles · 50 branches · concurrent holds. Convert it into peak throughput, retained records, bandwidth, and failure budget; then connect each number to a component.

DimensionWorking estimateDesign implication
Traffic1M titles · 50 branches · concurrent holdsSeparate online compute, durable state, and asynchronous consumers.
Peak factor5–10× averageAdmission control, autoscaling headroom, and hot-key protection.
Stored statewrite rate × retention × record bytesPartition count, indexes, compaction, backup, and archival tier.
Networkpeak operations × payload sizeBatching, compression, edge placement, and egress budget.
Availability99.9–99.99%Multi-zone replicas, tested failover, and dependency error budgets.
RecoveryRPO/RTO by invariantAck 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

POST /v1/loanscheckout copy
POST /v1/holdsqueue title request
POST /v1/returnsreturn and assign hold
POST /v1/loans/{id}/renewpolicy transition

Authoritative records and indexes

title(title_id PK, isbn, metadata)

copy(copy_id PK, title_id, branch_id, state, version)

loan(loan_id, copy_id, member_id, due_at, state)

hold(title_id, member_id, position, pickup_branch, expires_at)

Command and event envelopes

# Command envelope
operation: POST /v1/loans
request_id: req_01J...
idempotency_key: idem_7f2...
aggregate_key: library-management_123
expected_version: 17
payload: {...}

# Durable event
event_id: evt_01J...
aggregate_key: library-management_123
event_type: library_management_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.

ComponentOwned responsibilityWhy separate itRejected alternativeCost and evolution trigger
Member / Staff UIOwns one bounded domain decision, transition, or operational capability. In this design it participates in: Validate membership and circulation policy for the exact physical/digital copy.Separates a distinct scale, ownership, policy, or failure boundary. It prevents the external caller from also owning the contract and failure behavior of Catalog Search.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 librarians checkout same copy” 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 Member / Staff UI boundary before extracting or scaling it.
Catalog SearchPreserves durable ordering, replay, and consumer checkpoints. In this design it participates in: Conditionally transition AVAILABLE→LOANED and create loan with computed due date.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Member / Staff UI from also owning the contract and failure behavior of Membership Service.Call every downstream consumer synchronously. That couples user latency and availability to analytics, notification, indexing, and audit dependencies. Here that simpler choice would couple “Hold expires while member arrives” to Member / Staff UI.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 Catalog Search boundary before extracting or scaling it.
Membership ServiceOwns one bounded domain decision, transition, or operational capability. In this design it participates in: A hold enqueues by title/branch policy; return assigns first eligible member atomically.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Catalog Search from also owning the contract and failure behavior of Circulation Service.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 “Return at different branch” to Catalog Search.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 Membership Service boundary before extracting or scaling it.
Circulation ServiceOwns one bounded domain decision, transition, or operational capability. In this design it participates in: Renew only if policy permits and no higher-priority hold blocks the copy.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Membership Service from also owning the contract and failure behavior of Policy Engine.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 “Renewal races new hold” to Membership Service.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 Circulation Service boundary before extracting or scaling it.
Policy EngineOwns one bounded domain decision, transition, or operational capability. In this design it participates in: Fine calculator consumes overdue/loss transitions; notifications are asynchronous and idempotent.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Circulation Service from also owning the contract and failure behavior of Title Catalog.Embed policy in every caller. Rules then drift by deployment version and emergency changes require fleet-wide redeployment. Here that simpler choice would couple “Digital license count exhausted” to Circulation Service.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 Policy Engine boundary before extracting or scaling it.
Title CatalogPreserves durable ordering, replay, and consumer checkpoints. In this design it participates in: Validate membership and circulation policy for the exact physical/digital copy.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Policy Engine from also owning the contract and failure behavior of Copy 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 “Two librarians checkout same copy” to Policy Engine.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 Title Catalog boundary before extracting or scaling it.
Copy InventoryOwns authoritative versioned state or immutable evidence. In this design it participates in: Conditionally transition AVAILABLE→LOANED and create loan with computed due date.Defines the source of truth and commit boundary. It prevents Title Catalog from also owning the contract and failure behavior of Loan Aggregate.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 “Hold expires while member arrives” to Title 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 Copy Inventory boundary before extracting or scaling it.
Loan AggregateOwns one bounded domain decision, transition, or operational capability. In this design it participates in: A hold enqueues by title/branch policy; return assigns first eligible member atomically.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Copy Inventory from also owning the contract and failure behavior of Hold Queue.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 “Return at different branch” to Copy 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 Loan Aggregate boundary before extracting or scaling it.
Hold QueuePreserves durable ordering, replay, and consumer checkpoints. In this design it participates in: Renew only if policy permits and no higher-priority hold blocks the copy.Decouples bursts and failure from user acknowledgment. It prevents Loan Aggregate from also owning the contract and failure behavior of Branch Transfer.Call every downstream consumer synchronously. That couples user latency and availability to analytics, notification, indexing, and audit dependencies. Here that simpler choice would couple “Renewal races new hold” to Loan Aggregate.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 Hold Queue boundary before extracting or scaling it.
Branch TransferOwns one bounded domain decision, transition, or operational capability. In this design it participates in: Fine calculator consumes overdue/loss transitions; notifications are asynchronous and idempotent.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Hold Queue from also owning the contract and failure behavior of Fine Calculator.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 “Digital license count exhausted” to Hold Queue.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 Branch Transfer boundary before extracting or scaling it.
Fine CalculatorOwns one bounded domain decision, transition, or operational capability. In this design it participates in: Validate membership and circulation policy for the exact physical/digital copy.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Branch Transfer from also owning the contract and failure behavior of Payment Adapter.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 librarians checkout same copy” to Branch Transfer.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 Fine Calculator boundary before extracting or scaling it.
Payment AdapterOwns one bounded domain decision, transition, or operational capability. In this design it participates in: Conditionally transition AVAILABLE→LOANED and create loan with computed due date.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Fine Calculator from also owning the contract and failure behavior of Notification Service.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 “Hold expires while member arrives” to Fine Calculator.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 Adapter boundary before extracting or scaling it.
Notification ServiceOwns one bounded domain decision, transition, or operational capability. In this design it participates in: A hold enqueues by title/branch policy; return assigns first eligible member atomically.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Payment Adapter from also owning the contract and failure behavior of Digital License Adapter.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 “Return at different branch” to Payment Adapter.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 Notification Service boundary before extracting or scaling it.
Digital License AdapterOwns one bounded domain decision, transition, or operational capability. In this design it participates in: Renew only if policy permits and no higher-priority hold blocks the copy.Separates a distinct scale, ownership, policy, or failure boundary. It prevents Notification Service 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 “Renewal races new hold” to Notification Service.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 Digital License Adapter boundary before extracting or scaling it.
Decision 01

Chosen boundary vs simpler monolith

Separate bibliographic titles from physical or digital copies; circulation happens on copies.

A single deployable with one transactional store is the right starting point for a small library management workload. The proposed boundary becomes justified when separate bibliographic titles from physical or digital copies; circulation happens on copies. Separating it contains the named failure “Two librarians checkout same copy” without placing unrelated work on the critical path.

Decision 02

Correctness vs availability

Represent loan and hold lifecycles explicitly, including expiry, pickup branch, renewal, and loss.

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

Use a FIFO or policy-based hold queue with transactional copy assignment.

This choice adds dashboards, ownership, schema compatibility, capacity planning, and recovery runbooks. Start with the minimum topology that supports “A hold enqueues by title/branch policy; return assigns first eligible member atomically.” 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

Separate bibliographic titles from physical or digital copies; circulation happens on copies.

Scale boundary

Represent loan and hold lifecycles explicitly, including expiry, pickup branch, renewal, and loss.

Operational boundary

Use a FIFO or policy-based hold queue with transactional copy assignment.

↑ top

5. Critical flow and algorithm

01

Validate membership and circulation policy for the exact physical/digital copy.

02

Conditionally transition AVAILABLE→LOANED and create loan with computed due date.

03

A hold enqueues by title/branch policy; return assigns first eligible member atomically.

04

Renew only if policy permits and no higher-priority hold blocks the copy.

05

Fine calculator consumes overdue/loss transitions; notifications are asynchronous and idempotent.

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.

OperationConsistencyReason
Authoritative mutationConditional / serializableProtect uniqueness, ownership, money, inventory, order, or legal state transition.
Read after own writePrimary, session token, or cache warmThe caller must observe the result it just committed.
Ordinary read projectionBounded stale where safeReplicas and caches improve latency when version/freshness is visible.
Analytics and operationsEventual + replayFreshness 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

FailureRuntime behaviorRecovery proof
Two librarians checkout same copyResolve 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.
Hold expires while member arrivesStop 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.
Return at different branchStop 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.
Renewal races new holdResolve 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.
Digital license count exhaustedStop 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: Two librarians checkout same copy
operation_id: op_01J...
aggregate_key: library-management_123
last_durable_version: 18
projection_checkpoint: partition-07@offset-482901
retry_policy: same_identity / bounded_backoff
reconciliation_status: pending
owner: Digital License Adapter

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: Two librarians checkout same copy — alert before the user-visible invariant is violated.

Regional evolution

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

10. Interviewer follow-ups

How do inter-branch transfers work?

Separate bibliographic titles from physical or digital copies; circulation happens on copies. Walk the concrete lifecycle step “Validate membership and circulation policy for the exact physical/digital copy.” Then test it against “Two librarians checkout same copy”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How are renewals blocked by holds?

Represent loan and hold lifecycles explicitly, including expiry, pickup branch, renewal, and loss. Walk the concrete lifecycle step “Conditionally transition AVAILABLE→LOANED and create loan with computed due date.” Then test it against “Hold expires while member arrives”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How are fines calculated?

Use a FIFO or policy-based hold queue with transactional copy assignment. Walk the concrete lifecycle step “A hold enqueues by title/branch policy; return assigns first eligible member atomically.” Then test it against “Return at different branch”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do e-books differ?

Separate bibliographic titles from physical or digital copies; circulation happens on copies. Walk the concrete lifecycle step “Renew only if policy permits and no higher-priority hold blocks the copy.” Then test it against “Renewal races new hold”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do librarians override policy safely?

Represent loan and hold lifecycles explicitly, including expiry, pickup branch, renewal, and loss. Walk the concrete lifecycle step “Fine calculator consumes overdue/loss transitions; notifications are asynchronous and idempotent.” Then test it against “Digital license count exhausted”. 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

TimeWhat to cover
0–5 minClarify scope, workload, SLOs, and the correctness invariant.
5–10 minEstimate throughput/storage and define concrete APIs and records.
10–20 minDraw online, durable-write, and asynchronous/repair paths.
20–30 minExplain the exact lifecycle, partition key, and concurrency control.
30–38 minDiscuss caching, hot keys, replication, and backpressure.
38–43 minRun failure drills, security, observability, and regional evolution.
43–45 minSummarize the five most important choices and trade-offs.

Final decision summary

Invariantevery object transition is legal and atomic at one aggregate boundary.
PartitioningStable aggregate/key hash with versioned ownership and measured hot-key isolation.
CommitExpected version or uniqueness at the authoritative state boundary.
Async workDurable outbox/log, idempotent consumers, checkpoints, and replay.
RecoverySame 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