• Thu. May 30th, 2024

Aerospike invades the graph database space with a little help from Apache TinkerPop

Bynewsmagzines

Jun 20, 2023
Early adopters of data streaming see up to 10x returns, survey finds

[ad_1]

Join top executives in San Francisco on July 11-12, to hear how leaders are integrating and optimizing AI investments for success. Learn More


Real-time database vendor Aerospike is expanding its multi-model capabilities with the launch of the Aerospike Graph database.

Aerospike got its start back in 2009, providing a NoSQL database that in its early years focused on advertising applications. Over the past decade Aerospike has evolved to become a real-time database platform, useful for adtech, financial services and customer data platforms among other use cases.

In 2022, the company began its shift to offering what is known as a multi-model database, providing support for the JSON document model, which has become increasingly popular in recent years in part due to the success of document database vendor MongoDB.

Now Aerospike is expanding further with the general availability of Aerospike Graph, which brings graph data model capabilities.

Event

Transform 2023

Join us in San Francisco on July 11-12, where top executives will share how they have integrated and optimized AI investments for success and avoided common pitfalls.

 

Register Now

>>Don’t miss our special issue: Building the foundation for customer data quality.<<

A graph database is a type of data model structured to help users better understand relationships between different data points and content. There are multiple graph databases in the market today, including Neo4J and Amazon Neptune. Even Oracle has a graph database.

Graph databases are helpful for many different use cases, including fraud detection, an area where Aerospike customers have increasingly been headed and have needed a solution.

“What we decided last year was setting the course and a strategy to build a multi-model, multicloud data platform really focused on real-time workloads,” Subbu Iyer, CEO of Aerospike, told VentureBeat. “Our platform is really suited well for high performance at scale, low latency and high availability, so that’s what pushed us into looking at graph to really go after this space.”

Aerospike Graph: Built with open-source Apache TinkerPop technology

Aerospike didn’t build its graph database entirely from scratch. 

Rather what it did was find a suitable open-source base in the Apache TinkerPop project to build upon. Apache TinkerPop is a graph computing framework that includes its own query language known as Gremlin.

“When we found Apache TinkerPop, we realized it is a great solution, and we actually work with some of the original authors of TinkerPop,” Iyer said.

In effect what Aerospike has done with its graph database is develop a commercially supported version of TinkerPop. Iyer explained that Aerospike Graph handles the separation of compute and storage, enabling either type of resource to scale independently as needed. The database is available both as an on-premises technology and in a database-as-a-service (DBaaS) model in the cloud.

Aerospike Graph architecture
Image credit: Aerospike

With the initial release of Aerospike Graph, the company will be supporting the Apache TinkerPop project’s Gremlin query language. In the future, Iyer said that Aerospike could support other query languages for graph databases. Today there are multiple approaches for graph queries, including the cypher query language backed by graph database vendor Neo4j and the Property Graph Query Language (PGQL) backed by Oracle.

Is the next stop for Aerospike more AI?

As Aerospike continues to grow its platform, artificial intelligence (AI) capabilities are high on Iyer’s agenda.

The core Aerospike database platform is already being used by organizations as a feature store for AI pipelines, according to Iyer. There has also been a lot of effort in the industry overall in recent months to use existing data sources to help augment large language model (LLM) data for generative AI. That’s an area where vector databases are playing a role and it’s a space that Iyer is tracking closely.

“We’re looking at it very carefully,” Iyer said about vector databases. “It actually fits in very well with our multi-model strategy.”

>>Follow VentureBeat’s ongoing generative AI coverage<<

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.

Leave a Reply

Your email address will not be published. Required fields are marked *