Couchbase intros generative AI feature for its Capella DBaaS

Adding generative AI

Couchbase developed Capella as an alternative to its Couchbase Cloud offering first introduced in February 2020, providing customers with a fully managed DBaaS platform as opposed to a cloud-based platform that required more hands-on management.

The DBaaS platform eased some of the burdens placed on users, specifically targeting developers by making its tools more flexible and accessible, Redmonk analyst Stephen O'Grady told TechTarget at the time Capella was launched.

Recently, in June 2023, Couchbase further targeted developers by launching an integration with Netlify to simplify web application development.

The addition of Capella iQ is yet another move aimed at making it easier and faster for developers to build and manage applications.

Once generally available, the generative AI tool will enable developers to write SQL and other application-level code by automatically providing recommended sample code.

The feature uses foundation models to deliver recommendations in the Capella integrated development environment. In the IDE, developers can then use natural language to rapidly generate code, sample data sets and unit tests rather than go through the time-consuming process of writing the code and developing the sample data sets and unit tests themselves.

Because of the potential Capella iQ has to simplify application development and make developers more efficient, its addition to Couchbase's DBaaS is an important step, according to Carl Olofson, an analyst at IDC.

"Simplicity is always better," he said. "They've tooled the environment so it's easier to … develop large language models and foundation models. I think it's significant. They're heading in the right direction."

Olofson added that once generally available, Capella iQ could provide more comprehensive generative AI capabilities than the generative AI search capabilities so far unveiled by MongoDB, which he termed Couchbase's closest competitor.

"Couchbase is going beyond [search capabilities] to say that the generative AI work you're already doing can be extended to the data that's in Couchbase, and it provides you with easy ways to do that," he said. "What [Couchbase] is talking about is more comprehensive."

Couchbase, which used generative AI technology from OpenAI to help develop Capella iQ, began building the new tool following the launch of ChatGPT in late 2022, according to Scott Anderson, the vendor's senior vice president of product management and business operations.

Couchbase saw generative AI as a way to make its tools easier to use. It therefore developed Capella iQ to not only improve efficiency but also broaden Couchbase's use within organizations and make it more attractive to potential new customers.

"This is a [way] to jumpstart their learning on Couchbase to make it very efficient," Anderson said. "We view this as an accelerant for our existing [users] and net new customers who want to explore Couchbase."

More new features

Beyond Capella iQ, Couchbase unveiled additional capabilities also targeting improved efficiency and simplified development.

Just as the vendor's recent integration with Netlify was aimed at making application development easier, new integrations between Capella and developer tools Vercel and IntelliJ are similarly geared toward making developers and development teams more efficient.

In addition, Couchbase added more disk storage and improved the input/output per second (IOPS) performance of Capella so Capella clusters can handle the increasing size of data and analytics workloads more efficiently and with an improved user experience.

Increased IOPS allows Capella cluster to manage an increased number of concurrent read and write operations, according to Couchbase. That subsequently enables applications to remain available no matter what other usage making be taking place.

Finally, Couchbase added new security and governance capabilities that enabled Capella to achieve InfoSec Payment Card Industry Data Security Standard compliance and Cloud Security Alliance Security, Trust, Assurance and Risk certification, both of which are data security standards. Previously, the DBaaS achieved Service Organization Control 2 Type II audit and HIPAA compliance, which are enterprise requirements for cloud applications.

"We're focused on developer adoption," Anderson said. "That's what you're seeing with the integrations with IntelliJ and Vercel. It's making sure that we have the right ecosystem to allow developers to use the tools of their choice along with Capella to build their applications."

The added disk storage and improved IOPS similarly address users' experiences, he continued. In particular, they aim to help customers keep down the cost of using Couchbase's cloud-based tools by enabling them to do more with less compute power.

The vendor offers three pricing tiers: Basic, from 28 cents per hour per node; Developer Pro, from 35 cents per hour per node; and Enterprise, from 56 cents per hour per node. The more work customers can get done per node, the less they spend.

The increased security and governance, meanwhile, address customer confidence to ensure the safety and security of their data, Anderson said.

Perhaps more significant than the new Capella features is the launch of the Couchbase AI Accelerate partner program, according to Olofson.

The program is designed to create an ecosystem in which the vendor's customers can develop AI-powered applications using both Couchbase's capabilities as well those of partners with which the vendor has integrations.

"Naturally, what you want to do is have strong partnerships with technology providers to make sure you're providing customers with the tools to serve their own customers because that helps you sell more databases," Olofson said. "Also, it improves your system overall. It ensures you have a strong direction going forward. This has been true throughout time."

Future direction

With Capella iQ now in private preview and the other new Capella capabilities generally available, one focus of Couchbase's product development plans is to enable vector search, according to Anderson.

Vectors are a means of giving structure to unstructured data, such as text and images, so the previously unstructured data can be used in concert with structured data to inform decisions and actions. In addition, vectors help inform generative AI outputs, importantly reducing incorrect or misleading responses.

Neo4j and MongoDB are among other vendors planning to add vector search capabilities.

"We're in active development on vector search," Anderson said.

They've tooled the environment so it's easier to … develop large language models and foundation models. I think it's significant. They're heading in the right direction.

Before releasing vector search, however, he noted that Couchbase is making sure the capability meets the requirements of organizations that cannot afford mistakes, such as a failed supply chain, without suffering significant consequences.

"It has to be highly available, highly performant, highly scalable and highly reliable," Anderson said.

Beyond vector search, he added that Couchbase's roadmap includes improving its analytics service and taking further steps to make its use more cost-efficient.

Olofson, meanwhile, said that Couchbase and all database vendors should be focused on creating conversational interfaces so customers can essentially have a natural language conversation with their system to discover data and retrieve it.

"That goes for structured databases as well as unstructured databases," he said. "And I don't feel that's an unrealistic thing. When I talk to vendors, they say they're working on [conversational interfaces], but they say they don't have a timeline for them yet."

Databases are generally set up to serve specific needs, Olofson continued. Users know what data is in there so they can easily retrieve it for their needs.

But data volumes are growing. So is data complexity.

At a certain point, there will be too much disparate data within databases for data engineers and application developers to track. Conversational interfaces have the potential to make it easy for them to discover data that might otherwise get overlooked.

Beyond generative AI, Olofson noted that while Couchbase's strengths are its ease of use and compatibility with various other systems, the vendor needs to improve its visibility.

"It's business execution where I have my concerns," he said. "They need to be more aggressive in their marketing, and they need to work harder to recruit the developer community. For Couchbase, it's critical to go [after developers] and to back it up with large investments in advertising and marketing."

Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.