Neo4j makes AuraDB High Memory instances up to 2TB generally available
Neo4j has made AuraDB High Memory instances generally available on AWS and Google Cloud, offering up to 2TB of RAM and raising the service’s maximum storage to 5TB across every cloud. The company says the managed database is designed for large, memory-intensive graph workloads that previously required self-managed Neo4j Enterprise Edition deployments.
- AuraDB High Memory instances are generally available on AWS and Google Cloud.
- The largest configuration provides 2,048 GB of RAM on AWS and 1,952 GB on Google Cloud.
- AuraDB maximum storage has increased to 5TB across AWS, Google Cloud and Azure.
- The instances target large knowledge graphs, fraud analysis, identity resolution and other demanding workloads.
- Availability depends on regional cloud capacity and prepaid customer terms.
More memory for large graph workloads
AuraDB Business Critical and Virtual Dedicated Cloud customers can now request four High Memory instance sizes on each cloud. AWS options range from 768 GB to 2,048 GB of RAM, while Google Cloud options range from 744 GB to 1,952 GB.
The available configurations pair those memory sizes with storage ranges from 1.5 to 5 TB. Each size includes a matching vCPU allocation, according to Neo4j, although the company says customers should ask their account team about options in their region.
Neo4j says memory is particularly important for graphs containing frequently accessed nodes, relationships and indexes. Graph queries can traverse relationships, evaluate paths and retrieve context across connected entities. When a large graph spills to disk, the company says multi-hop traversals and deep path queries are among the first operations to feel the effect.
Storage increases across AWS, Google Cloud and Azure
Alongside the High Memory launch, Neo4j has increased the maximum storage for AuraDB Business Critical and Virtual Dedicated Cloud from 2TB to 5TB on AWS, Google Cloud and Azure. The increase applies to instances with 192GB of RAM and up, within a 1:16 RAM-to-storage ratio.
The company positions the larger configurations for knowledge graphs and GraphRAG, fraud and financial crime analysis, identity resolution, customer 360 applications and digital twins. These workloads can involve documents, embeddings, transaction histories, customer interactions or streams of telemetry that continue to expand.
Neo4j also describes the instances as an option for migrations from self-managed deployments. Customers can move large existing graphs to AuraDB without splitting them into smaller pieces to fit infrastructure constraints.
Managed through existing AuraDB tools
High Memory instances use the same Aura workflows as other AuraDB instances. Customers can create, resize, pause, clone and snapshot them through the Aura API or Console. Neo4j says they also include the operational, scale and security features of the customer’s tier, including secondaries, Multiple databases, SSO and Private Link.
Memory and storage can be set separately during creation or resizing. RAM is billed per GB-hour and includes disk at a 1:2 RAM-to-storage ratio, while additional storage is billed per GB-hour at the standard rate. High Memory instances are available to AuraDB Business Critical and Virtual Dedicated Cloud customers on prepaid terms.
Availability depends on regional cloud capacity and must be confirmed before Neo4j enables the instances for a project. Customers can request a larger instance from the Aura Console’s sizing menu or contact their Neo4j account team.
