Google Cloud Broadens Database Offerings Through Integration of Advanced AI Features

This week, Tokyo hosts Google’s Cloud Next conference, with an agenda centered around enhancing database solutions for artificial intelligence (AI) applications, a focal point for major technology corporations in 2024. The enhancements encompass the Spanner SQL database, now equipped with graph and vector search functionalities, alongside improved full-text search features.

Integral to the announcement are the Gemini-powered capabilities, extending to BigQuery and Looker, which aim to streamline data engineering, analysis, governance, and security operations for its users.

Google maintains that although a large portion of businesses believe generative AI will be crucial to their success, there’s a recognition that a significant fraction of their data is unorganized. This disarray excludes crucial datasets from their AI and analytics efforts.

“To leverage generative AI effectively, companies must break free from their data silos and consolidate their data ecosystem. This involves handling both structured and unstructured data, as generative AI excels in analyzing the latter. Integrating static data with dynamic data streams is essential,” shared Gerrit Kazmeier, Google’s VP and GM for database, Data Analytics, and Looker. He emphasized that these innovations are designed to enhance enterprise data utilization.

Enhancements to Spanner: Introducing Graph and Vector Search

The backbone of Google services such as Search, Gmail, and YouTube, Spanner services top-tier clients including Home Depot and Uber. Spanner’s ability to manage vast datasets makes it ideal for integrating into GenAI applications, necessitating the addition of vector and graph database capabilities to support advanced foundational models.

“Our goal is to expand Spanner’s capabilities, including its reliability and scalability, to become the premier database solution for operational GenAI applications,” stated Andi Gutmans, Google’s VP and GM for databases. Adding graph features through the GraphQL standard marks Google’s initial step towards enhancing Spanner’s functionality, facilitating its use in GenAI applications with Retrieval Augmented Generation (RAG).

Spanner now also offers vector and full-text search, with vector search powered by Google’s ScaNN algorithm. “The addition of Spanner Graph, full-text, and vector search evolves Spanner into a multi-model database with advanced functionalities, making it capable of supporting a new breed of AI-powered applications,” Google explains.

Moreover, Spanner introduces “Spanner editions,” a flexible, optional pricing strategy catered to provide users with more adaptable options, moving away from the fixed single or multi-region offerings.

Bigtable’s Evolution to Support SQL

In another strategic move, Google announced significant updates to Bigtable, its NoSQL database designed for unstructured data and demanding latency conditions, which now includes support for SQL queries (specifically, GoogleSQL). This development vastly simplifies access for developers.

Earlier, the Bigtable API was the only way to query data, but now it supports approximately 100 SQL functions.

Oracle Databases Welcome on Google Cloud

Oracle database enthusiasts can now deploy their Oracle Exadata and Autonomous database services directly within Google Cloud data centers, allowing for integration between Google Cloud and Oracle Cloud applications. This partnership brings additional workloads to Google Cloud while Oracle maintains its licensing revenue, regardless of the customers’ cloud preference.

Google Cloud also broadens its offerings with support for open-source Apache Spark and Kafka for data streaming and processing, in addition to real-time data streaming via the Analytics Hub, facilitating secure inter-organizational data sharing.