Couchbase’s Capella Columnar paves the way for developers to build more adaptive applications – SiliconANGLE

Couchbase Inc. is adding some powerful new capabilities to its flagship database-as-a-service offering, paving the way for developers to build more intelligent and adaptive applications that leverage real-time insights.

The company is also enhancing the mobile version of its database with new functions including vector search, making it more suitable for generative AI applications that employ retrieval-augmented generation techniques to fine-tune their outputs.

The company is the creator of the popular Couchbase NoSQL database service that allows enterprises to store large volumes of semistructured data. Its flagship product is the cloud-based Couchbase Capella database-as-a-service product, and it also sells a traditional on-premises version called Couchbase Server, plus a mobile version.

Couchbase Capella’s main advantage over traditional databases such as Oracle Database is that it can process both structured and unstructured information at the same time. It’s a fairly unusual capability that makes Couchbase a better choice for certain kinds of applications that need to access both types of data, as it means they can use just one product. The platform can also act as a data cache, meaning users actually get three systems for the price of one.

Adaptive, real-time applications

With the launch of Capella Columnar on Amazon Web Services, Couchbase says, it’s giving developers a simple way to perform traditional analytics on data formatted using the flexible JavaScript Object Notation or JSON data format, so they can benefit from real-time insights into their applications.

JSON is an open standard file format and data interchange format that uses human-readable text to store and transmit data objects. As Couchbase explains, it has become popular with developers when building business-critical applications because it enables much greater flexibility in terms of the types of data those apps can leverage.

But the challenge with JSON data is that it’s difficult to apply traditional analytics systems that are built for information that adheres to more rigid structures. Without those formal structures, teams are required to spend hours configuring their data so they can perform analytics on it.

Capella Columnar helps to address the challenges of parsing and transforming JSON data into an analytics-ready columnar format. The secret sauce is Capella IQ, a generative artificial intelligence coding assistant, which can transform JSON data into SQL++ so it can be analyzed instantly. It can ingest data from multiple sources, including other databases besides Couchbase, and do this in real-time, before “writing back” the results to the operational side of Capella immediately.

According to Matt McDonough, senior vice president of product and partners at Couchbase, this “writeback problem” has never been addressed by traditional analytics systems before, because it was always too difficult to anticipate what a developer might want to do with the transformed data. “Capella Columnar implements the solution, and the needs of AI-powered applications provide the motive,” he explained.

The result is that developers can build more adaptive, real-time applications that deliver personalized experiences to users. For instance, an e-commerce retailer might use Capella Columnar to analyze a site visitor quickly in order to recommend suitable products, while a video game developer can use it to provide in-game metrics that boost engagement.

International Data Corp. analyst Carl Olofson said Capella Columnar is a priming new capability that builds on the already unique capabilities of Couchbase’s database-as-a-service offering.

“[It delivers] blended analytic-transaction processing support that leverages the performance advantage of columnar data management together with vector search in support of applications demanding intelligent data access at the speed of business,” he explained. “These are capabilities that the market has been looking for but are hard to find contained in a single product.”

Vector search comes to Couchbase Mobile

The addition of vector search capabilities in Couchbase’s embedded database offering, Couchbase Mobile is another key development, paving the way for offline, edge-based AI applications, the company said. Couchbase Mobile is built atop of Couchbase Lite, a lightweight version of its flagship database that can be embedded in mobile and internet of things applications.

Those devices will now be able to support vector search at the edge of the network, meaning they can perform reliable and fast semantic search and even power RAG-based generative AI when there’s no internet connection.

Vector search is key to enhancing generative AI applications. Vectors are used to store unstructured information as “embeddings,” or mathematical representations of data. This is necessary because unstructured information cannot be stored in traditional rows and columns. Therefore, generative AI can only access that data if it’s stored as vectors in a database that also supports vector search.



“By enabling on-device vector search, we’re not just speeding up applications and eliminating downtime due to internet outages; we’re opening up new possibilities for secure AI development,” said McDonough. “Our solution combines the cloud’s scalability to handle massive AI datasets with edge capabilities for immediate, on-device processing.”

Developers get a freebie

In a final update, Couchbase is creating a new database tier called the Capella Free Tier, which, as the name suggests, enables developers to get started using its database with limited capacity for free.

Launching later this month, it’s really a teaser, giving developers the chance to experiment with Couchbase’s database and see what kind of difference it can make to their applications. The free tier includes access to Capella IQ, the generative AI assistant, and Capella Workbench, its coding tool.

RedMonk analyst Kate Holterhoff said the free tier should encourage more developers to try Couchbase’s platform, because most want to see what it can do before making any financial commitment. “By offering a free tier to Capella, Couchbase acknowledges the tinkering spirit that motivates these important builders and creators,” she said.

Image: SiliconANGLE/Microsoft Designer