With the general release of Aura Enterprise on Wednesday, Neo4j is offering its first fully-managed cloud graph database designed specifically for large organizations.
The new database is a follow-up for Neo4j, a graph database vendor founded in 2007 and based in San Mateo, Calif. It previously released Aura Professional, a fully managed cloud graph database designed for use by small and medium-sized businesses, in November 2019.
Graph databases enable users to access and use their data in different ways than traditional relational databases. They simplify the connection of data points, allowing them to simultaneously connect with more than one other data point in order to more quickly pull data from disparate sources and speed up the process of turning data into insights.
Common use cases include fraud detection, social networking and generating recommendations.
The new technology
Aura Enterprise is designed to enable organizations to build graph applications at enterprise scale, advancing the capabilities of Aura Professional, which empowered small- and medium-sized business to build their applications. It facilitates the management of large data sets with billions of nodes and relationships for fast data queries, and provides enterprise-grade security with end-to-end encryption and node-level access controls.
As with Aura Professional, pricing is consumption-based. Customers pay for only the compute power they use in the Neo4j cloud graph database. They are able to purchase capacity in advance — and get discounts by doing so — and then flexibly choose when and how to use that capacity.
Mike LeoneSenior analyst, Enterprise Strategy Group
With that pricing model for its fully managed service, Neo4j is making graph database technology more available to potential customers, according to Mike Leone, senior analyst at Enterprise Strategy Group, a division of TechTarget.
“Graph databases are gaining traction, but hesitancy remains due to the perceived effort it may take to ramp up deployments or ensure performance at scale for distributed data,” he said. “Organizations view the public cloud to be a big part, if not the basis for their ongoing data analytics strategy, and Neo4j is focused on making graph in the cloud more accessible by lowering the barrier to adoption.”
In particular, he added, the burdens of infrastructure management and upkeep can be a hindrance to migration to the cloud and adoption of graph databases. Neo4j has taken steps to eliminate such concerns, Leone continued, which enables data scientists to spend their time developing models, building applications and discovering relationships in their data rather than doing mundane maintenance.
The Orchard, a music distribution vendor, is one organization that has adopted Neo4j’s technology and was part of the beta testing for Aura Enterprise. Given the complexities of music distribution rights, The Orchard was attracted to graph technology and began with Neo4j in October 2019. What it wanted, however, was a managed service.
“Neo4j has a plethora of customers, and one of the benefits of their managed solution compared to us hosting on premises is that they’re working with tons of clients,” said Jacob Fowler, CTO at The Orchard. “They’re seeing other clients and how other clients are handling scale, so we get the benefits of being part of a broader ecosystem.”
Meanwhile, by not having to maintain its graph database, The Orchard’s technology officers are able to pursue other projects.
“We have a pretty aggressive roadmap,” Fowler said. “Our DevOps team is constantly striving to not only build scalable solutions but also be forward-thinking about how to set up multi-region [operations]. It frees them up to focus on other roadmap items that they would have gotten to a quarter or two down the road.”
Aura Enterprise’s evolution
Neo4j introduced Aura Professional before developing Aura Enterprise because small- and medium-sized businesses were a low-risk place for Neo4j to start with a fully managed service in the cloud, according to Kurt Freytag, director of product management at Neo4j.
The tool evolved from the applications users were building on Neo4j’s open source platform, and the people building the applications were not users the vendor expected to monetize. Those users, however, wanted a service to which they could outsource operations, and so Neo4j developed Aura Professional.
“They wanted high-quality, high-availability hosting, and they didn’t have DevOps people,” Freytag said. “They are smaller shops that want to focus on building their applications and building their business value, and outsource everything else. It was a low-risk place for us to start … and we had a lot of success.”
After learning how to become a service business with its first SaaS offering, the plan was to build a feature product for larger organizations, but one with more advanced capabilities.
“We built a foundation for the service on the premise that it was highly available, always-on and had a quality of service that was ultimately targeted to the enterprise,” Freytag said. “But we held back on some of the security aspects that would ultimately differentiate our higher tier offering for our enterprise customers.”
Before developing Aura Enterprise, however, Neo4j’s plan was to target lower-end developers to help them get started with the vendor’s platform. When COVID-19 struck, Neo4j pivoted and instead focused on readying Aura Enterprise for release. Since Aura Professional had been built with the idea that it would serve as the basis for an enterprise version, the vendor was able to turn around development of Aura Enterprise in less than a year.
Essentially, according to Freytag, Neo4j’s managed services are quite similar, other than added security measures for Aura Enterprise and the size of the graphs and databases that Aura Enterprise can handle. Aura Professional tops out at 64 GB of RAM.
That scale, meanwhile, is part of what attracted The Orchard.
“Their motto and ethos has been low latency, sub-second performance, and we’re seeing it,” Fowler said. “That was the whole reason we built our metadata permissions model into Neo4j. We could have done it in a relational database, but the amount of performance that would have been sacrificed wouldn’t have worked for us.”
Future plans for Aura Enterprise
With Aura Enterprise now generally available and, according to Freytag, 15 customers on board, Neo4j is beginning an early program for Aura Enterprise on Amazon Web Services (AWS) and hopes to start a similar program for Microsoft Azure.
In addition, Neo4j plans to continue developing security features for its cloud graph database service.
“The feedback we’ve gotten is that it’s a strong 1.0 start, but there are certainly additional features in the security access that customers want,” Freytag said, identifying third-party single sign-on (SSO) integration as a requested feature.
Meanwhile, Freytag said there will also be that focus on lower-end developer adoption that was temporarily pushed aside by the pandemic. Neo4j plans to broaden the accessibility of Aura by enabling developers to get started at low costs and even for free.
“Organizations are still learning how graph databases can truly transform their organization,” Leone said. “[Aura] will enable the next generation of graph-database users to gain value faster than they likely thought was possible, and further cements Neo4j as a leader in this space.”