Netflix is built on it. Home Depot's online curbside pickup service is built on it. Both business juggernauts are using Apache Cassandra—the open-source NoSQL distributed database platform.
There is a tendency to think of Apache Cassandra as a platform for huge companies. However, there are use cases for midsize companies as well.
Apache Cassandra tends to get "pigeonholed [as] this very, very large-scale type of workload. It's the platform that Netflix is built on. It's the platform that Uber is built on," said Bill McLane, CTO of cloud at DataStax.
McLane said DataStax—with its AstraDB Apache Cassandra platform for building applications, including GenAI applications—delivers consistent throughput and latency when working with data.
In addition, it is effective "because the midmarket, especially on the enterprise architecture side of things, what a lot of folks are looking for is a highly distributed approach," he said.
Alhough midmarket organizations may be much smaller than an Uber, they may have customer data and application needs that span the globe and require massive scalability.
Midsize retailers, for example, can experience "microbursts" of retail traffic during events like Black Friday, McLane said.
"You need to be able to have that ability to understand … what my back-end infrastructure is going to provide [and] is going to scale and manage without having to even think about it," he added.
McLane also spoke about GenAI applications, which are intriguing to businesses of all sizes.
"Everybody's trying to figure out how do they bring generative AI into their enterprise architecture. I think the trend that we're going to see out of that is … basic functionality for things like natural language processing chatbots makes perfect sense for generative AI to come into play," he said.
Some organizations, no matter their size, may be wondering how to proceed with building their own scalable and secure GenAI applications and integrate them into their environments.
"What I'm hearing a lot of is … how do we take what the existing architectures that we know are proven [and] how do we take things like an Apache Cassandra deployment or an Astra deployment that's built on 10-plus years of technology that's proven in the marketplace, proven to scale, proven to perform, proven to provide security, and how do we extend that to these generative AI applications?" McLane said.
"And I think the trend is you're going to see a lot of these upcoming vector database companies that are specifically designing for that generative AI space. They're either going to get subsumed by other companies or they're going to fade into the background for very niche types of applications," he said.
