News

Neo4j 3.1 Delivers New Clustering Architecture

• By John K. Waters
• October 26, 2016

Neo Technology, the creator and chief commercial supporter of the Neo4j open source NoSQL graph database implemented in Java, announced the public beta of a major update of its namesake product. Neo4j 3.1 is an enterprise-focused release with new operational security and features designed to make it easier to deploy it across the entire organization.

"Because this release had some pretty heavy-duty and complex enterprise features, we really didn't want to just call it GA once it was feature complete," Philip Rathle, VP of products at Neo Technology, told ADTmag. "We decided that it made sense to go through an extended beta to allow all the customers who have been asking for this stuff to tell us if it came out the way they were hoping it would."

The company expects this release to become generally available in December, Rathle said.

Core to Neo4j 3.1 release is a new clustering and security architecture based on the Raft protocol, which the company calls Casual Clustering. The Raft consensus algorithm supports ultra-large clusters and a range of cluster topologies for the datacenter and the cloud. Load balancing is handled in this release by Neo4j drivers and the Neo4j Bolt protocol, which operate in concert with the Neo4j cluster to provide "tunable consistency models" built on top of Neo4j's ACID (Atomicity, Consistency, Isolation, Durability) foundation.

This release's new support for "causal consistency" provides applications with an always-consistent view of their data at scale, Rathle explained, for the many applications where "eventual consistency," which is the NoSQL norm, is not sufficient.

This release also comes with new security features, including role-based authorization with native user management and federated authentication via LDAP/Active Directory among others. Collectively these features will help meet various enterprise security compliance requirements, such as HIPAA and PCI-DSS, Rathle said.

Also new in this release is what Rathle called "long-requested" improvements in the database kernel, including more efficient space management, a graph Schema View included the Neo4j Browser developer tool, and a new natively encrypted command-line interface. Also in this release new administrative features, including commands to view and kill running queries.

This release also makes official Neo4j support for IBM POWER8 with CAPI Flash, which enables in-memory and near-memory graph processing on an ultra-large scale, Rathle said, "setting a new and higher bar for the scale of graph problems that can be addressed with other technologies."

Neo4j continues to be something of a harbinger of the growing need for advanced graph database technology. Analysts at Forrester Research estimate that more than 25 percent of enterprises will be using a graph database by 2017. That growth will span virtually all industries, the analysts predict, in areas ranging from digital content management to bioinformatics, ID management to the Internet of Things.

"In today's ever-connected world, Big Data continues to get bigger, and so do the challenges that come with it," said Neo Technology CEO Emil Eifrem, in a statement. "As a result, graphs are no longer on the cusp of going mainstream; they have been established as the power and basis for the transformation of how organizations do business…"