Package Registry — 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 package registry 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 downloads/day · 100M versions · strong namespace integrity.
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 package registry 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 downloads/day · 100M versions · strong namespace integrity. 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 downloads/day · 100M versions · strong namespace integrity	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

POST /v1/packages/{name}/uploads

POST /v1/uploads/{id}/publish

GET /v1/packages/{name}/{version}

POST /v1/resolvelockfile result

Authoritative records and indexes

package(namespace, name, visibility, owner)

version(package_id, version, blob_digest, metadata, state)

blob(digest PK, size, uri, scan_state)

tombstone(package_id, version, reason, effective_at)

Command and event envelopes

# Command envelope
operation: POST /v1/packages/{name}/uploads
request_id: req_01J...
idempotency_key: idem_7f2...
aggregate_key: package-registry_123
expected_version: 17
payload: {...}

# Durable event
event_id: evt_01J...
aggregate_key: package-registry_123
event_type: package_registry_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
Package Client	Owns client protocol, request identity, retry, and resume behavior. In this design it participates in: Authorize namespace and create resumable content upload session.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents the external caller from also owning the contract and failure behavior of Registry Edge.	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 “Version metadata publishes before blob replication” 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 Package Client boundary before extracting or scaling it.
Registry Edge	Owns authentication, validation, quotas, deadlines, and routing. In this design it participates in: Verify digest/signature, scan bytes, and store immutable content-addressed blob.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Package Client from also owning the contract and failure behavior of Registry API.	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 “Namespace typo enables dependency confusion” to Package Client.	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 Registry Edge boundary before extracting or scaling it.
Registry API	Owns authentication, validation, quotas, deadlines, and routing. In this design it participates in: Atomically publish version metadata pointing only to ready blob.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Registry Edge from also owning the contract and failure behavior of Namespace Auth.	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 “Malware verdict changes after download” to Registry Edge.	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 Registry API boundary before extracting or scaling it.
Namespace Auth	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Resolve constraints into exact versions/digests and cache immutable downloads at CDN.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Registry API from also owning the contract and failure behavior of Upload Session.	Embed policy in every caller. Rules then drift by deployment version and emergency changes require fleet-wide redeployment. Here that simpler choice would couple “Popular package overloads origin” to Registry 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 Namespace Auth boundary before extracting or scaling it.
Upload Session	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Replicate blobs, index metadata, and enforce tombstone/retention policy with audit.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Namespace Auth from also owning the contract and failure behavior of Metadata DB.	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 “Maintainer tries to overwrite version” to Namespace Auth.	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 Upload Session boundary before extracting or scaling it.
Metadata DB	Owns authoritative versioned state or immutable evidence. In this design it participates in: Authorize namespace and create resumable content upload session.	Defines the source of truth and commit boundary. It prevents Upload Session from also owning the contract and failure behavior of Content Blob Store.	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 “Version metadata publishes before blob replication” to Upload Session.	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 Metadata DB boundary before extracting or scaling it.
Content Blob Store	Owns authoritative versioned state or immutable evidence. In this design it participates in: Verify digest/signature, scan bytes, and store immutable content-addressed blob.	Defines the source of truth and commit boundary. It prevents Metadata DB from also owning the contract and failure behavior of Checksum / Signature.	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 “Namespace typo enables dependency confusion” to Metadata DB.	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 Content Blob Store boundary before extracting or scaling it.
Checksum / Signature	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Atomically publish version metadata pointing only to ready blob.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Content Blob Store from also owning the contract and failure behavior of Malware Scanner.	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 “Malware verdict changes after download” to Content Blob Store.	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 Checksum / Signature boundary before extracting or scaling it.
Malware Scanner	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Resolve constraints into exact versions/digests and cache immutable downloads at CDN.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Checksum / Signature from also owning the contract and failure behavior of Dependency Resolver.	Embed policy in every caller. Rules then drift by deployment version and emergency changes require fleet-wide redeployment. Here that simpler choice would couple “Popular package overloads origin” to Checksum / Signature.	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 Malware Scanner boundary before extracting or scaling it.
Dependency Resolver	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Replicate blobs, index metadata, and enforce tombstone/retention policy with audit.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Malware Scanner from also owning the contract and failure behavior of Search Index.	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 “Maintainer tries to overwrite version” to Malware Scanner.	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 Dependency Resolver boundary before extracting or scaling it.
Search Index	Serves a low-latency, disposable projection with authoritative fallback. In this design it participates in: Authorize namespace and create resumable content upload session.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Dependency Resolver from also owning the contract and failure behavior of CDN.	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 “Version metadata publishes before blob replication” to Dependency Resolver.	Costs invalidation, staleness, stampede protection, and memory. Add or split it only after the workload (1B downloads/day · 100M versions · strong namespace integrity) proves repeated reads dominate. Measure the latency, saturation, lag, or mismatch at the Search Index boundary before extracting or scaling it.
CDN	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Verify digest/signature, scan bytes, and store immutable content-addressed blob.	Removes repeated work from the latency-critical path. It prevents Search Index from also owning the contract and failure behavior of Replication Worker.	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 “Namespace typo enables dependency confusion” to Search Index.	Costs invalidation, staleness, stampede protection, and memory. Add or split it only after the workload (1B downloads/day · 100M versions · strong namespace integrity) proves repeated reads dominate. Measure the latency, saturation, lag, or mismatch at the CDN boundary before extracting or scaling it.
Replication Worker	Executes idempotent asynchronous work under lease and retry policy. In this design it participates in: Atomically publish version metadata pointing only to ready blob.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents CDN from also owning the contract and failure behavior of Deletion / Tombstone.	Perform the work before acknowledging the caller. This gives simpler consistency but unbounded tail latency and fragile retries. Here that simpler choice would couple “Malware verdict changes after download” to CDN.	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 Replication Worker boundary before extracting or scaling it.
Deletion / Tombstone	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Resolve constraints into exact versions/digests and cache immutable downloads at CDN.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Replication Worker from also owning the contract and failure behavior of Audit / Provenance.	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 “Popular package overloads origin” to Replication 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 Deletion / Tombstone boundary before extracting or scaling it.
Audit / Provenance	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Replicate blobs, index metadata, and enforce tombstone/retention policy with audit.	Proves correctness and supports controlled recovery. It prevents Deletion / Tombstone from also owning the contract and failure behavior of operators and repair tooling.	Infer correctness from request logs. Logs do not prove durable business state, detect silent drift, or safely repair ambiguous outcomes. Here that simpler choice would couple “Maintainer tries to overwrite version” to Deletion / Tombstone.	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 Audit / Provenance boundary before extracting or scaling it.

Decision 01

Chosen boundary vs simpler monolith

Package blobs are content-addressed and immutable; mutable metadata points versions/tags to verified hashes.

A single deployable with one transactional store is the right starting point for a small package registry workload. The proposed boundary becomes justified when package blobs are content-addressed and immutable; mutable metadata points versions/tags to verified hashes. Separating it contains the named failure “Version metadata publishes before blob replication” without placing unrelated work on the critical path.

Decision 02

Correctness vs availability

Publish uses a staged upload and atomic metadata commit so resolvers never observe a version with missing bytes.

The design coordinates only where each accepted package registry 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

Dependency resolution is client or service policy, but every result pins exact versions/hashes to make builds reproducible.

This choice adds dashboards, ownership, schema compatibility, capacity planning, and recovery runbooks. Start with the minimum topology that supports “Atomically publish version metadata pointing only to ready blob.” 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 package registry 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

Package blobs are content-addressed and immutable; mutable metadata points versions/tags to verified hashes.

Scale boundary

Publish uses a staged upload and atomic metadata commit so resolvers never observe a version with missing bytes.

Operational boundary

Dependency resolution is client or service policy, but every result pins exact versions/hashes to make builds reproducible.

↑ top

5. Critical flow and algorithm

Authorize namespace and create resumable content upload session.

Verify digest/signature, scan bytes, and store immutable content-addressed blob.

Atomically publish version metadata pointing only to ready blob.

Resolve constraints into exact versions/digests and cache immutable downloads at CDN.

Replicate blobs, index metadata, and enforce tombstone/retention policy with audit.

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 package registry 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
Version metadata publishes before blob replication	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.
Namespace typo enables dependency confusion	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.
Malware verdict changes after download	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.
Popular package overloads origin	Isolate the key/tenant/priority, apply admission control, and preserve durable work.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Maintainer tries to overwrite version	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: Version metadata publishes before blob replication
operation_id: op_01J...
aggregate_key: package-registry_123
last_durable_version: 18
projection_checkpoint: partition-07@offset-482901
retry_policy: same_identity / bounded_backoff
reconciliation_status: pending
owner: Audit / Provenance

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: Version metadata publishes before blob replication — 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 you prevent dependency confusion?

Package blobs are content-addressed and immutable; mutable metadata points versions/tags to verified hashes. Walk the concrete lifecycle step “Authorize namespace and create resumable content upload session.” Then test it against “Version metadata publishes before blob replication”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How are packages scanned?

Publish uses a staged upload and atomic metadata commit so resolvers never observe a version with missing bytes. Walk the concrete lifecycle step “Verify digest/signature, scan bytes, and store immutable content-addressed blob.” Then test it against “Namespace typo enables dependency confusion”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

Can a version be deleted?

Dependency resolution is client or service policy, but every result pins exact versions/hashes to make builds reproducible. Walk the concrete lifecycle step “Atomically publish version metadata pointing only to ready blob.” Then test it against “Malware verdict changes after download”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How does CDN caching work?

Package blobs are content-addressed and immutable; mutable metadata points versions/tags to verified hashes. Walk the concrete lifecycle step “Resolve constraints into exact versions/digests and cache immutable downloads at CDN.” Then test it against “Popular package overloads origin”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do private namespaces work?

Publish uses a staged upload and atomic metadata commit so resolvers never observe a version with missing bytes. Walk the concrete lifecycle step “Replicate blobs, index metadata, and enforce tombstone/retention policy with audit.” Then test it against “Maintainer tries to overwrite version”. 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 package registry 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 package registry 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

Package Client

Registry Edge

Registry API

Namespace Auth

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