SAN FRANCISCO – October 4, 2022 – ArangoDB, the company behind the most complete graph data and analytics platform, today announced the GA release of ArangoDB 3.10 to provide graph usability at scale. ArangoDB 3.10 includes new features, such as EnterpriseGraph, ArangoSearch 2.0, and locality-sensitive hashing, that not only make it simpler for organizations to implement and use graph technology at scale, but also combine it with full-text search.
“At ArangoDB, we believe in the power of graph technology to solve today’s biggest business problems, and it is our mission to make it easier to drive value from connected data, faster,” said Claudius Weinberger, ArangoDB co-founder and chief product officer. “ArangoDB 3.10 makes it even easier to implement and use graph across large datasets, as well as easily combine graph traversals and search queries for deeper insights.”
“As we counter intellectual property infringement across hundreds of millions of products listings, we need to process graphs with billions of relationships,” said Michael Haapaoja, Software Engineering Manager at OpSec Security. “ArangoDB has the capabilities to operate at this scale and generate robust data grouping.”
Key features in ArangoDB 3.10 include:
EnterpriseGraph
EnterpriseGraph distributes large graph datasets in such a manner that minimizes the number of network hops between servers. It does so without requiring database administrators to define the graph’s sharding attributes – making it simple to accommodate large volumes of connected data without sacrificing performance.
ArangoSearch 2.0
ArangoSearch, ArangoDB’s natively-integrated full-text search and ranking engine, now allows users to create inverted indices at the collection level. This not only makes it easier to implement ArangoSearch, as it removes the need for user-defined search views, but also more seamless to integrate search and graph queries together. This frees engineering teams from time-consuming management of separate graph and search server infrastructure.
Locality-sensitive hashing
Locality-sensitive hashing reduces complexity around entity resolution by reducing the number of documents that need to be compared when looking for duplicates. This provides performance improvements and reduces complexity around knowledge graph use cases, such as fraud detection, that need to find connections in large volumes of disparate data.
ArangoDB 3.10 also includes:
- Computed values: Allows users to create new values that are computed based on other, already existing values. This increases performance for queries that have predefined parameters.
- Read from Followers: In order to increase query throughput, ArangoDB now offers the option to read from Follower shards in addition to Leader shards.
- Parallelism for Sharded Graphs: Graph traversals with different start vertices can now run in parallel across distributed data. This provides an almost linear performance improvement.
- Native M1/ARM support: ArangoDB is now available for the ARM architecture, in addition to the Intel x86-64 architecture. It can natively run on Apple silicon (e.g. M1 chips), as well as 64-bit ARM (AArch64) chips under Linux.
Furthermore, ArangoDB 3.10 improves shard rebalancing, optimizes rule management, creates indexes in parallel, and projects traversals, among further functionality.
Packaging and availability
ArangoDB 3.10 is available immediately. In both the open source ArangoDB Community Edition and the commercial ArangoDB Enterprise Edition are: ArangoSearch 2.0, computed values, native M1/ARM support, improved shard rebalancing, and optimizer rule management.
Available only in ArangoDB Enterprise Edition are: EnterpriseGraph, locality-sensitive hashing, Read from Followers, Parallelism for Sharded Graphs, parallel index creation, and traversal projections.
For a self-managed deployment, ArangoDB Community and Enterprise Editions can be downloaded here. ArangoDB can also be deployed via ArangoGraph Insights Platform, ArangoDB’s fully-managed, next-generation graph data and analytics platform.
