The open source graph database vendor released a new version of its platform with new features to enable users to better deal with the challenges of growing data workloads.
Open source graph database vendor ArangoDB updated its namesake database with new capabilities that improve both scalability and search.
ArangoDB 3.9 became generally available on Feb. 15 for open source and enterprise users, as well as on the ArangoDB Oasis database-as-a-service cloud platform.
Graph database technology is an active market with a bright future, according to IDC analyst Carl Olofson. He noted that the market today is dominated by purpose-built graph database platforms such as Neo4j, TigerGraph and Amazon Neptune.
ArangoDB has promoted its product first as a document database, then as a multimodel database, but it seems to me that their strongest play is as a graph database. ArangoDB has made significant advances in the graph space.
Carl OlofsonAnalyst, IDC
"ArangoDB has promoted its product first as a document database, then as a multimodel database, but it seems to me that their strongest play is as a graph database," Olofson said. "ArangoDB has made significant advances in the graph space, and I think that is their strong suit."
ArangoDB 3.9 scaling graph database operations
ArangoDB 3.9 includes a number of updates to improve scalability.
"We're seeing our customers going to larger and larger clusters and use cases," said Jörg Schad, CTO at ArangoDB. "So in terms of data and in terms of deployment, we built in a lot of things to actually enable that at scale."
Among the scalability improvements is the ability to execute queries faster across a large number of database nodes. ArangoDB 3.9 also contains optimizations for database cluster balancing as workloads grow.
Hybrid smart graphs land in ArangoDB 3.9
One of the scalability capabilities in ArangoDB 3.9 comes in the form of a feature the vendor refers to as a hybrid smart graph.
ArangoDB already had a capability known as smart graphs, which enables graph data to be sharded across multiple database nodes. Schad noted that ArangoDB users have also simply replicated data, typically for small data sets, to scale out graph data sets. Replicated data provides a full copy of data, while sharding provides slices of data spread across nodes.
Hybrid smart graphs enable users of the ArangoDB query optimizer to employ a combination of both smart sharding and data replication to searched-for data, Schad said.
ArangoDB 3.9 includes updates to help optimize graph queries as workloads scale.
ArangoSearch gets a boost for graph queries
As part of the ArangoDB 3.9 release, the company has also updated its ArangoSearch functionality.
ArangoSearch is a full-text search capability that can search across different data types including graph, document data in JSON and geospatial data. Schad said he has seen applications in which organizations start with a full-text search to identify potential starting nodes for graph pattern matching for analytics as well as fraud detection.
The update for ArangoSearch provides what the company calls a segmentation analyzer for language searching.
It's not uncommon for a single document or data graph to contain content in multiple languages, and Schad noted that ArangoSearch supports more than 40 human languages. With the update, it can segment queries and results by language.
Looking forward, Schad said upcoming versions of ArangoDB will likely get improved storage performance to help make graph queries faster. He also said the vendor will keep working on enhancing the overall user experience of interacting with the ArangoDB database.
"We have a lot of really interesting features that are coming," Schad said.
