What's new and what's next for Google Cloud databases

Generative AI’s capabilities have grown tremendously over the past year, and the pledge we made to infuse generative AI across Google Cloud to make our products and services helpful for all users is now a reality. Today, we’re announcing a new era for database professionals with Gemini in Databases, part of our work to deliver Gemini for Google Cloud. With this, we’re delivering AI-powered assistance to help simplify all aspects of the database journey including developing, monitoring, optimizing, securing and migrating database-driven applications.

We’re also continuing to help you unlock the full potential of generative AI by providing a unified data and AI platform that allows you to connect all your data to AI. We’re announcing vector support across even more of our databases to help enterprise developers quickly build gen AI apps that are accurate, relevant and grounded in enterprise truth. We’re also helping you future-proof your architecture with industry-leading database services that help you drive innovation and transform customer experiences. Let’s deep dive into the announcements.

Supercharge database development and management with AI

Databases are a core component of every company's innovation strategy. However, customers often tell us that databases are also the trickiest part of the stack to design, code, operate, and manage. In addition, migrations are complex, time consuming and resource intensive. Companies often end up investing in specialized roles and tools to help them, but still often find themselves blindsided by operational issues and inefficiencies. What if we could use AI to help? 

Our journey to bring AI-assisted capabilities to our databases started last year with Duet AI in Spanner and Duet AI in Database Migration Service. Since then, we've expanded AI assistance across all aspects of the database journey to supercharge database development and management. Available today, Gemini in Databases helps developers, operators, and database administrators get their jobs done better, faster, and easier. Gemini in Databases enables them to easily generate SQL; additionally, they can now manage, optimize and govern entire fleets of databases from a single pane of glass; and finally, accelerate database migrations with AI-assisted code conversions. Imagine being able to ask questions like "which of my production databases in east Asia had missing backups in the last 24 hours?" or “how many PostgreSQL resources have a version higher than 11?” and getting instant insights about your entire database fleet. 

Specifically, we’re delivering three AI-assisted experiences to help you be more productive:

1. Code development with Database Studio: Database Studio, our rich SQL editor in the Google Cloud console, can allow developers to easily generate, summarize, and fix SQL code with intelligent code assistance, code completion, and guidance directly in the editor — improving application development productivity. Developers can also leverage a context-aware chat interface that uses natural language to help build database applications faster and with higher-quality SQL suggestions. Database Studio supports popular SQL dialects such as MySQL and PostgreSQL.

2. Fleet management with Database Center: Operators can now manage an entire fleet of databases from a single pane of glass. Intelligent dashboards proactively assess your availability, data protection, security, and compliance posture. Using natural language, database teams can interact with their databases to quickly and easily find the information they need. In addition, an LLM-powered conversational agent generates troubleshooting tips tailored to your specific problems. 

3. Assisted migrations with Database Migration Service: With the help of Gemini, you can examine and convert database-resident code such as stored procedures, triggers, and functions, to the PostgreSQL dialect. Additionally, to help upskill and retrain SQL developers, Gemini-powered database migration plans focus on explainability, with side-by-side comparison of dialects, along with detailed explanations of the code and recommendations.
   
   “With Gemini in Databases, we can get answers on fleet health in seconds and proactively mitigate potential risks to our applications more swiftly than ever before.” - Bogdan Capatina, Technical Expert in Database Technologies, Ford Motor Company

Gemini in Databases is your ultimate database companion, helping you keep everything running smoothly and making sure you're always in the know. Gemini in Databases is available today in public preview. To learn more, visit https://cloud.google.com/products/gemini/databases

Build generative AI apps grounded in operational data with AlloyDB

In February, we announced the general availability (GA) of AlloyDB AI, an integrated set of capabilities in AlloyDB to more easily build enterprise generative AI applications. Our goal with AlloyDB AI is to make it a great home for building generative AI apps with PostgreSQL. We have continued to innovate on AlloyDB and today we are announcing the next-generation of AlloyDB AI, which includes new vector capabilities, easier access to remote models, and secure and flexible natural language support.

At Google, we have more than 12 years of experience innovating on real-world vector algorithms to support some of our most popular services, including Google Search, and YouTube. We had to invent new ways of indexing and searching vectors to meet the most demanding use-cases. Last year, we announced support for open-source pgvector for our PostgreSQL databases, and today, we're delivering the next generation of tree-based vector capabilities to relational databases. We're excited to provide PostgreSQL developers another vector option in AlloyDB AI with a new pgvector-compatible index which is based on Google’s state-of-the-art approximate nearest neighbor algorithms, the ScaNN index. In our performance tests, AlloyDB AI offers up to four times faster vector querying than the popular `hnsw` index in standard PostgreSQL, up to 8 times faster index creation, and typically uses 3-4 times less memory than the HNSW index in standard PostgreSQL. With this announcement, AlloyDB AI now addresses some of the memory, indexing speed, and querying performance requirements displayed in demanding real-world use cases and applications - and here’s a whitepaper that goes into detail about how we accomplished this. This capability is available as a technology preview in AlloyDB Omni today, and coming soon to AlloyDB on Google Cloud.

To facilitate easier management of inferencing endpoints, we’re announcing the AlloyDB model endpoint management, which makes it even easier to call remote Vertex AI, third-party, and custom models. In addition to Vertex AI, model catalog can also be easily configured for third-party services such as Anthropic and Hugging Face. This is available in AlloyDB Omni today, and coming soon to AlloyDB on Google Cloud.

"Nuro is an autonomous driving company that uses vector similarity to help classify objects that autonomous vehicles encounter while driving on the road, to ultimately trigger the right action. Nuro currently has hundreds of millions of vectors that they are moving to AlloyDB AI in order to simplify their application architecture." - Fei Meng, Head of Data Platform, Nuro

Finally, we're announcing two new features in AlloyDB AI to support flexible, accurate, and secure natural language experiences. First, we’re enabling gen AI developers to build applications that accurately query data with natural language — just like they do with SQL— for maximum flexibility and expressiveness. That means generative AI apps can respond to a much broader and more unpredictable set of questions. Second, we’re adding a new type of database view called “parameterized secure view” that makes it easy to secure your data based on the end-users’ context enabling you to deliver richer and more flexible natural language experiences. Together, these advances available in AlloyDB Omni today, present a new paradigm for integrating operational data into generative AI apps. Learn more here. 

Additionally, in February, we announced vector capabilities across even more of our databases including Spanner, Memorystore for Redis, and Cloud SQL for MySQL, joining our PostgreSQL engines AlloyDB and Cloud SQL for PostgreSQL. Today, we are adding vector support to Firestore. We’ve also made it easy to connect to AI inferencing with Vertex AI, and integrated with the open-source orchestration frameworks like LangChain. All these capabilities help you to connect your data to AI, ensuring its outputs are relevant and grounded in your operational data, for enterprise truth.

Deliver industry-leading database services

To take advantage of industry trends like generative AI, you need best-in-class data infrastructure. Our cloud-first Spanner, Bigtable, and Firestore databases are the backbone for Google services such as Ads, Gmail and YouTube, and provide some of the highest levels of availability and reliability in the industry. With an up to 99.999% SLA, Spanner and Bigtable process billions of transactions per second (4 billion+ and 7 billion+ at peak, respectively), while Firestore has more than 500,000 monthly active developers using the service today!

Today we’re announcing Bigtable Data Boost — a breakthrough technology that delivers high-performance, workload isolated, on-demand processing of transactional data. Similar to the Spanner Data Boost capability we announced last year, Bigtable Data Boost can let you execute analytical queries, ETL jobs and train machine learning models directly and frequently on your transactional data, without disrupting operational workloads. We’re also announcing authorized views which allow multiple teams to leverage the same tables securely and share data like sales and inventory with partners, suppliers or sellers directly from your database. And lastly, with the new Bigtable distributed counters you will be able to process high-frequency event data like clickstreams directly in your database to deliver real-time operational metrics and machine learning features at scale.

We’ve also been focused on our open-source MySQL, PostgreSQL and Redis engines. Last year, we launched the Cloud SQL Enterprise Plus edition for MySQL and PostgreSQL, which delivers up to three times higher read throughput and up to two times improvement in write latency compared to the Enterprise edition. It also supports a 99.99% availability SLA inclusive of maintenance. Today, we’re adding more value to the Enterprise Plus offering with advanced failover capabilities such as orchestrated switchover and switchback — a welcome addition to enterprises’ business continuity plans. Now, Day 2 operations like maintenance and instance scale-up can happen with subsecond downtime, with no proxy required.

Finally, for Redis workloads, Memorystore for Redis Cluster offers zero downtime scaling and supports up to 14.5TB of memory per cluster. Today, we're announcing new capabilities for Memorystore for Redis Cluster that provide higher resiliency and cost-optimal options to support various workloads. With the support for both AOF (Append Only File) and RDB (Redis Database)-based persistence, you can now take periodic snapshots or log every write, achieving a near-zero recovery point objective (RPO), and delivering additional resiliency in the event of a zonal failure—all at no extra cost. We are also providing customers with new node shapes for maximum flexibility and better cost management.

Embrace an AI-driven future

There’s a wealth of data in operational databases just waiting to power the next transformative generative AI applications. All these capabilities we're announcing today help you connect your data to AI to boost productivity and to accelerate your generative AI journey with the databases you already use and love.

To learn how to get started, visit cloud.google.com/databases or watch What's next for Google Cloud Databases Spotlight.