Aerospike, Inc. is debuting their new graph database, built to be developer-ready, real-time, and highly scalable. The launch of Aerospike Graph responds to the lack of graph database solutions that can support large graph datasets with a sustainable cost model, often pushing companies to drain their own resources in an effort to create ad-hoc, inflexible solutions.
Aerospike Graph is an innovation in both performance and cost, delivering millisecond multi-hop graph queries at extreme throughput across trillions of vertices and edges, according to the company. The development of this solution directly addresses the phenomenon of organizations creating their own solutions that pose a massive drain on resources, including time and money.
“Many people have either had to experiment with the existing technologies that are out there for graph, but when they put them into production, sometimes they have to recode everything in a bespoke manner to get the scale they have,” said Lenley Hensarling, chief product officer at Aerospike. “And this [Aerospike Graph] will just make that a lot easier, and they'll be able to apply graph technology for production use at scale for more and more things without having to hand-code solutions.”
The database allows data teams to scale compute and storage independently, ensuring that users only pay for what they need. It is further proven to offer radically low TCO, where even at the highest levels of scale, Aerospike Graph requires up to 80% less infrastructure compared to its competitors, according to the company.
“Everything we do has to one, scale linearly, two, have low latency, three, have high availability and four, do this in a cost-effective manner,” explained Hensarling. “We don't just say, ‘Hey, hardware is getting better, so this'll work, don't worry about how efficient it is.’ We actually are massively parallel in the way we code. We actually count instructions, you know, in key areas of the code to make sure that this makes maximal use of the hardware available.”
“We don't go through the operating system, we go directly to the NVMe driver for the device, and we treat it as a block memory device. And that allows us to use the storage devices and gain near in-memory access speeds,” he continued. “That's sort of our secret sauce, that we have engineers who really dig in and do that kind of work, year after year.”
Aerospike Graph also features native support for the Gremlin query language, enabling developers to write applications with new and existing Gremlin queries within Aerospike Graph. Heavy transactional (OLTP) applications can now run in real-time, supporting use cases such as AdTech, Customer 360, fraud detection and prevention, and more.
On top of its highly performant, highly scalable, and low TCO benefits, Aerospike is also multi-model, supporting diverse workloads across key value, document, and graph NoSQL data models. By centralizing these models within a single, real-time data platform, Aerospike eases the burden of data management while increasing efficiency at scale, in the core and at the edge.
Looking to the future, Aerospike will continue to look at other models to include within its database, particularly noting the popularity in LLMs and vector databases.
“We continue to look at new spaces. With LLMs being a big deal…we think we have a place there much like with graph,” said Hensarling. “When we do it, it will be something that supports scale and can apply this in a cost-effective manner…with these key design centers that we have to bring efficiency, low latency, and high throughput.”
To learn more about Aerospike Graph, please visit https://aerospike.com/.