Discover How Neptune Analytics Empowers AWS with Vector Search and Graph Data Capabilities

A lively debate in the AI community revolves around which type of database plays a crucial role in unearthing truthful information within generative AI applications: graph or vector databases. AWS has taken a step to bridge this gap by introducing a new service called Amazon Neptune Analytics, unveiled today at AWS re:Invent.

During the AI keynote at re:Invent, Swami Sivasubramanian, AWS's vice president of data and machine learning, showcased this innovative tool. He emphasized that AWS aimed to synthesize the strengths of both database types. “Both graph analytics and vectors focus on revealing hidden relationships within data, so we pondered: ‘What if we merged vector search with the capability to analyze vast graph datasets in mere seconds?’ And today, we are proud to announce we’ve done just that,” he stated.

Sivasubramanian elaborated on how this new service enables customers to analyze existing Neptune graph data or data lakes supported by S3 storage. By leveraging vector search, users can gain essential insights. “Neptune Analytics simplifies the discovery of relationships in your graph through vector search by integrating graph and vector data storage,” he noted.

Ray Wang, founder and principal analyst at Constellation Research, described this tool as a smart integration of the two technologies. “Vector databases excel in managing and querying high-dimensional data for machine learning, while graph databases are fantastic for relationship mapping,” he explained. “Vector databases can struggle with dimensionality, and graph databases often face scalability issues. Traditionally, these systems have operated separately. Merging them enhances each other's capabilities.”

As a fully managed service, Neptune Analytics means AWS handles the complex infrastructure management. This allows users to focus on solving challenges through queries and workflows. According to AWS, “Neptune Analytics dynamically allocates compute resources based on graph size and swiftly loads all relevant data into memory, enabling users to execute queries in seconds,” as outlined in a recent blog post.

Beginning today, Neptune Analytics is accessible as a pay-as-you-go service across seven AWS regions, including US East (Ohio, N. Virginia), US West (Oregon), Asia Pacific (Singapore, Tokyo), and Europe (Frankfurt, Ireland).

AI, AWS re:Invent 2023, cloud, enterprise, graph databases, Neptune Analytics, vector databases.