Warakorn - Fotolia
Performance and scalability improvements are on tap at graph database provider Neo Technology Inc. in its latest software release. That means updates to both sides of the read and write architecture that underlays the company's NoSQL database engine, Neo4j.
Neo said the Neo4j 2.2 platform speeds write throughput via enhancements such as faster buffering of updates. Meanwhile, an in-memory graph cache boosts the database's read performance.
As a graph database, Neo4j is built for fast performance in applications such as social networks, recommendation engines and master data management -- all cases in which quickly finding relationships in data is a favorable trait. But there was room for improvement in areas such as reads and writes, as well as initial imports of very large graphs, according to Philip Rathle, vice president of products at Neo, based in San Mateo, Calif.
"We're turning attention to the engine room," Rathle said, adding that the company "looked at the product from the inside out" in order to improve read and write scalability in Neo4j 2.2.
Neo also added a cost-based query optimizer for its Cypher query language in the new release. That tool is designed to look at statistical information on graph size and shape, and create query execution plans that spend less time executing. In addition, query plan visualization software lets developers see upfront how the optimizer will actually execute the queries.
"Speed hasn't been a challenge for users. But sometimes obtaining speed meant going beyond the Cypher language and writing Java code," Rathle said.
Notable customers that Neo can claim include Wal-Mart and eBay. The former uses the NoSQL database to help divine real-time product recommendations for online shoppers, while the latter employs Neo4j to optimize package routing.
Performance optimization technology for reads, writes and queries are long-established elements in the SQL-based relational databases that are the market mainstays. Newer NoSQL databases of all ilks will have to add those and other elements as they look to move up the ladder of product maturity, and take a larger place in enterprise IT architectures.
Learn how a graph database is powering a room booking system
Discover traits to look for when weighing NoSQL DBs
Read about Facebook's jump into graph databases
How a fintech firm uses graph processing for credit evaluations