Neo4j Now ‘Multi-Data Center Aware’
Customers can now scale their Neo4j graph databases horizontally across multiple data centers around the globe, Neo Technology announced today. Neo4j version 3.2 also brings new governance and security capabilities and a focus on performance, among other enhancements.
Today’s launch brings the first major enhancements to the Causal Clustering architecture that Neo Technology introduced last December as a way to construct large graph database clusters that span multiple data centers and support ACID transaction guarantees without introducing additional complexity.
With Neo4j 3.2, the database is now “multi-data center aware,” Neo Technology says. Clusters running this database management system now sport “intelligent routing” capabilities designed to squeeze the most scale and performance out of the horizontally deployed database, it says. The Causal Clustering architecture sports a third generation of the Raft protocol that delivers robustness, reliability, scale, and ACID transaction integrity, the company adds.
This release brings a focus on improving write speeds into the database, including a 30% to 250% increase in write speeds by rewriting label indexes as “graph native,” the company says. Other performance gains come from the inclusion of new composite indexes that can have more than one property; a new compiled runtime for Cypher (the company’s graph development language); and replacing the rules-based query optimizer with Cypher’s cost-based optimizer.
Security and governance were also points of emphasis for this release. Neo4j now sports a new Kerberos module to bolster security and authentication. The database is also now able to enforce data quality rules through a new schema constraint, called Node Keys. Version 3.2 also gains expanded monitoring metrics that should give admins more control over, and insight into, the queries running on the system.
Neo unveiled Neo4j today during GraphConnect Europe, a Neo-sponsored event taking place this week in London.
“I’m so excited to introduce Neo4j 3.2 on the same stage at GraphConnect as customers like Telia, eBay, Airbnb and UBS,” Neo Technology’s Vice President of Products, Philip Rathl, tells Datanami. “This release is customer-driven from the need for global scale, governance and security in the enterprise, to the ingenious ways we have improved performance in the native graph stack. Every Neo4j developer will enjoy 3.2 because it is faster, more scalable and easier to use.”
Neo Technology is a leader in the graph database market, with more than 200 customers, including Wal-mart, UBS, Cisco, HP, adidas Group and Lufthansa. The San Mateo, California company released the first version of Neo4j back in 2002, and since then the open source product has been downloaded more than 3 million times.