Graph database vendor Neo4j Inc. today introduced a fully managed version of its product with what it calls enterprise-grade security and massive scalability.

The new Neo4j Aura Enterprise isn’t the company’s first foray into managed services. It launched Aura Professional, a service aimed at developers, in 2019. Although the new offering is functionally the same, it’s delivered in a single-tenant environment for greater security and includes such corporate-oriented features as end-to-end encryption, built-in virtual private cloud isolation and role-based and node-level access controls.

“Aura Enterprise is a bit of a child of COVID,” said Kurt Freytag, director of product management for the platform. “Our original plan wasn’t to launch it this year but we saw a massive spend moving to the cloud and the right thing to do strategically was to rapidly accelerate deployment.” The “vast majority” of the company’s fourth-quarter business was from customers managing their own cloud instances, he said.

Neo4j’s business historically has been based on on-premises deployments of an enterprise version of its open-source graph database, which is a type of data manager that represents complex relationships and enables rapid navigation between elements to discover correlations. The company also has a version available on the Google LLC Cloud Platform, but it is not built on cloud-native principles.

Aura “was built it from ground up as cloud-native,” Freytag said. “It’s orchestrated by Kubernetes, which means the system is inherently self-healing. Kubernetes can automatically take care of a lot of the lifecycle and management challenges. It solves a lot of resilience problems that would otherwise be cumbersome to build.” Kubernetes is an orchestrator for the self-contained operating environments called software containers.

Containers are designed for stateless applications, which live in isolation with no stored knowledge of or reference to past transactions. That makes them poorly suited for stateful applications like database managers, which require continuous relationships between tables and elements.

However, “Kubernetes has been maturing such that we can be on the cutting edge and use its inherent capabilities to take advantage of its reliability,” Freytag said.

Billions of nodes

Aura Enterprise provides for fast query performance across data sets with billions of nodes and relationships, which are the graph the equivalent of relational rows and columns. Capacity-based pricing is on a consumption basis with pre-purchased credits applied against usage and discounts for long-term contracts. Consumption is measured in gigabytes of memory used.

The service meets atomicity, consistency, isolation and durability, or ACID, standards. It also delivers a guaranteed 99.95% service uptime with upgrades and patches applied without interruption.

Fine-grained security is a particularly important feature, Freytag said. “If you have a graph of sales information history you can control access for so that marketing and sales can only see a portion of it,” he said. Such capabilities weren’t included in the earlier Developer version.

Aura Enterprise is already in production with 15 customers, Freytag said. One of them is Lucinity Group ehf, a Reykjavik-based startup that uses knowledge graphs to correlate patterns and behaviors that help financial institutions identify money laundering, less than 1% of which is ever detected.

“Aura has allowed us to be the [software as a service] SaaS company that we wanted to be from the beginning and to consume Neo4j the way we wanted to,” said Chief Executive Gudmundur Kristjansson. “Aura is consumption-based, we get the full power or Neo4j, we can launch a cluster, segregate clients like we want and have full security in the cloud. We don’t need to worry about the management.” Kristjansson called the process of moving from the on-premises version to the managed cloud service “simple.”

Image: Neo4j