Graph Query Engine Claims Record

(Bakhtiar Zein/Shutterstock)

Cambridge Semantics, the graph-based provider of analytics and data management services, claimed a benchmark record this week for database query performance running on a public cloud platform.

The Boston-based analytics vendor said Wednesday (Dec. 7) its Anzo Graph Query Engine completed a load and query of 1 trillion “triples” on the Google Cloud Platform in just less than two hours. That result, the company claimed, was 100 times faster than other query engines running on a metric called the Lehigh University Benchmark operating at the same data scale.

The company said examples of 1 trillion triples include six months worth of global Google (NASDAQ: GOOGL) searches or 100 million facts describing the details of 10,000 clinical trial studies—in other words, a lot of data.

Barry Zane, vice president of engineering at Cambridge Analytics, noted that a major challenge for semantic-based analytics “has been enabling a load and query performance on very large data sets from a data lake” that are being demanded by enterprise analytics users. Zane claimed the benchmark test validated that the loading and querying steps, a process that previously took businesses more than a month, could now be performed in less than two hours.

Cambridge Semantics also is touting the benchmark results to promote its graph-based online analytical processing approach as data diversity and volumes grow. The company further argues that current relational database management systems are failing to keep pace with soaring data volumes and the onslaught of unstructured data and streaming video from social media, Internet of Things and other sources.

The Anzo Graph Query Engine is a clustered, in-memory graph analytics engine based on open semantic standards that intended for ad hoc and interactive queries and analytics across large and varied data sets. The query engine runs on dedicated servers or, in the case of the benchmark test, can be provisioned on cloud infrastructure such as Google Cloud Platform.

The Lehigh benchmark is designed to evaluate the performance of semantic web repositories based on queries of large datasets.

Earlier this year, Cambridge Semantics acquired the intellectual assets of graph database specialist SPARQL City, which was co-founded Zane. At the time, the company said the addition of SPARQL City’s in-memory graph query engine would expand its Anzo query engine based on semantic web technology. The platform is intended to help customers develop interactive, real-time data analytics capabilities.

Recent items:

Cambridge Semantics Buys Graph Database Specialist

A Semantic Approach to Big Data Governance