Neo4j Community, Enterprise, Aura, and Infinigraph: Capability, Responsibility, and Cost Boundaries

Learning outcomes

AtlasMart’s proof of concept now works on one Community container. Procurement asks whether production should use Community, self-managed Enterprise, Aura, Neo4j Desktop, or Infinigraph. Treating those names as “free, paid, cloud, and bigger” would hide the engineering boundary: each option changes available capabilities, licensing/terms, who operates the DBMS, how failures are handled, and what evidence can be reproduced locally.

Five names, five different boundaries

The most important word in the table is boundary. Cypher data access can look similar across deployments while administration, topology, backup, security, observability, plugin availability, and failure ownership differ substantially. Application portability is therefore not the same as operational portability.

Community: excellent learning baseline, intentionally limited production surface

Community provides the core graph database: property graph storage, Cypher, ACID transactions, indexes/Community-supported constraints, Bolt, Browser, Cypher Shell, and official drivers. It is suitable for the course because the majority of modeling/query semantics can be learned without a commercial license. The current Cypher documentation distinguishes Community from Enterprise in multi-database and security behavior: Community has the system database and one user database, and it supports multiple users but not Enterprise role-based privilege separation; users have implied administrator-level access.

This matters for labs. Creating a second Community user does not demonstrate least privilege. Likewise, SHOW DATABASES and SHOW SERVERS belong to Enterprise administration surfaces and must not be mandatory evidence in a Community-only lab. Instead, Community evidence should use commands that exist there, such as CALL dbms.components(), SHOW CURRENT USER, relevant SHOW SETTINGS, and normal graph queries.

CALL dbms.components() YIELD name, versions, edition RETURN name, versions, edition;SHOW CURRENT USER;SHOW SETTINGS 'db.query.default_language' YIELD name, value RETURN name, value;

Enterprise: more capability means more operator responsibility, not less

Self-managed Enterprise is still self-managed. It adds features needed by many production systems: clustering/failover, online backup, RBAC and broader security controls, multi-database administration, and other enterprise operations. But AtlasMart still owns machine/VM/container security, supported JVMs, certificates, storage, capacity headroom, monitoring, backup schedules and copies, restore drills, patching, topology changes, and incident response unless another managed layer is contracted separately.

Enterprise Docker images use an -enterprise suffix and require license acceptance. The course does not start one automatically because the mandatory path must remain free/local and contract-neutral. If a learner has valid Enterprise/evaluation rights, these are examples of edition-specific evidence rather than steps required to complete Chapter 01:

SHOW DATABASES YIELD name, type, access, currentStatus, default, home RETURN *;SHOW SERVERS YIELD name, address, state, health, hosting RETURN *;SHOW ROLES;SHOW USERS;

On a standalone Enterprise instance, SHOW SERVERS still describes the server topology available to the DBMS; in a cluster it becomes much more meaningful. Later chapters teach the current cluster model, where servers and databases are decoupled and databases have primary/secondary allocations. Do not back-port older “core/read replica” mental models without checking the current release.

Aura: managed infrastructure does not outsource application correctness

Aura is Neo4j’s fully managed cloud service. A driver can interact with Aura through the same broad Cypher/Bolt application model, usually using a secure neo4j+s:// URI supplied by the service. But a managed deployment removes or changes server-level controls: you do not treat Aura like a VM where you edit neo4j.conf, browse /data, or run arbitrary local neo4j-admin commands. Aura provides service-specific console/API, snapshots, metrics/logs, and tier-dependent controls.

AtlasMart still owns query correctness, graph modeling, indexes/constraints, application secrets, driver configuration, authorization choices exposed by the tier, data retention decisions, integration failure handling, and application-level recovery plans. Neo4j’s current Aura terms also make clear that customers retain responsibilities for secure configuration and appropriate backups. “Managed” should therefore be translated into a responsibility matrix, not “no operations.”

Desktop Developer use is not a shortcut around Enterprise licensing

Neo4j Desktop 2.x is a developer client/application for local DBMS instances and remote connections. Current Desktop documentation says it includes a Developer edition with Enterprise capabilities for an individual named user on a single machine for development use. That is valuable for learning Enterprise commands without building production infrastructure, but it is intentionally not equivalent to a production Enterprise license or a multi-user production deployment.

A useful course pattern is therefore: Community Docker for mandatory reproducibility; Desktop as an optional GUI/developer experience; Enterprise/evaluation only when a chapter genuinely needs the feature and the learner has appropriate rights; Aura as an optional managed comparison; deterministic simulations for topology/backup/security concepts where direct reproduction would otherwise create a paywall.

Infinigraph changes the scale architecture

Infinigraph is not a larger Aura tier. The current Operations Manual describes it as a specialized Enterprise Edition with automatic sharding for property databases and graph analytics capabilities, requiring a separate subscription. Property sharding was introduced in the 2025.12 line and is explicitly unavailable on Aura. It requires Cypher 25. That combination is a strong example of why server line, Cypher version, edition, and deployment type must be recorded together.

The course will not simulate Infinigraph by pretending several Community containers implement the same semantics. When direct reproduction is unavailable, a later lesson can use topology diagrams, deterministic partition/shard examples, query routing traces from official documentation, and explicit “cannot reproduce identically in Community” labels.

Re-run the Community lab and classify the evidence

Keep the Chapter 01 Community container running. Collect evidence, then annotate each line with “portable application behavior,” “self-managed operational evidence,” or “edition-specific behavior.”

docker exec atlasmart-neo4j cypher-shell -a neo4j://localhost:7687 -u neo4j -p atlasmart-course-2026 "CALL dbms.components() YIELD name, versions, edition RETURN name, versions, edition;"docker exec atlasmart-neo4j cypher-shell -a neo4j://localhost:7687 -u neo4j -p atlasmart-course-2026 "SHOW CURRENT USER;"docker inspect atlasmart-neo4j --format '{{.Config.Image}}'docker inspect atlasmart-neo4j --format '{{json .Mounts}}' 

CALL dbms.components() is application/admin-visible DBMS evidence. The Docker image and mounts are deployment evidence and have no Aura equivalent. SHOW CURRENT USER exists across relevant modes, but the meaning of its role/security columns depends on edition/tier. This classification habit prevents “works in my container” from becoming an undocumented production assumption.

Deliberately wrong approaches and their repairs

Production selection checklist

AtlasMart should decide from requirements outward. Start with tolerated downtime, RPO/RTO, expected graph size and growth, query concurrency, write rate, graph algorithms, security/SSO/RBAC needs, geographic placement, compliance, observability, backup/restore, team operations skill, upgrade cadence, and cost. Then map those requirements to a deployment. Include an exit path: how data can be exported, how applications isolate driver/config differences, which features create lock-in, and how rollback works during migration.

Summary and next step

Community, Enterprise, Aura, Desktop, and Infinigraph share important Neo4j concepts but differ in feature surface, license/terms, topology, and who operates what. The safe course baseline remains pinned Community; paid/managed capabilities are taught with explicit boundaries rather than silently assumed.

Next, install or start the DBMS deliberately and verify Browser, Cypher Shell, authentication, runtime versions, and driver connectivity end to end.

Authoritative references

• Current Neo4j versions — Official current-release and LTS patch snapshot.
• Neo4j Operations Manual — Authoritative self-managed operational documentation for the current release.
• Neo4j system requirements — Supported operating systems and JVM requirements.
• Cypher Manual — Current Cypher 25 reference and language semantics.
• Configure the Cypher default version — Cypher 5 versus Cypher 25 default and override behavior.
• Neo4j in Docker — Official image tags, ports, editions, and Docker starting point.
• Neo4j editions — Current Community, Enterprise, and Infinigraph feature framing.
• Neo4j Desktop — Desktop 2.x and Developer edition development-use boundary.
• Neo4j Aura documentation — Managed Aura products, tiers, console and operations.
• Cypher and Aura — Cypher and administrative differences in Aura.
• Infinigraph property sharding — Current subscription, Aura, and Cypher-version requirements.
• Neo4j Legal Terms Overview — Current entry point for Community, Desktop, self-managed and Aura terms.