Hornet's Astra DB Migration Paves the Way for AI and Vector Search Advancements in Empowering LGBTQ+ Communities
Hornet is a social networking community and app for the gay community. It has more than 35 million members worldwide, and provides a community home base that is available anytime and anywhere. Hornet has become the most popular gay app in countries such as France, Brazil, Turkey, and Taiwan, and is rapidly expanding its sizable user base across Europe and the United States. Formed in 2011, the team at Hornet is based globally across South Africa, the United States, The Czech Republic, and Hong Kong. Hornet has to support a rapidly growing community and ensure that services are available, responsive, and secure.
- Reduced cost for running Cassandra up to 20% with pay-as-you-use service
- Freed IT staff to concentrate on other issues around data infrastructure
- Implemented scalable platform for future data growth
- Real-time matching has significantly improved user experiences
The team at Hornet wanted to support the launch of a new social networking application that would complement the company’s existing dating app. The new app, SPACES, provides a safe environment for members of the LGBTQ+ community to meet each other around shared interests and locations.
“When we asked our community and other groups, we found there was interest in getting a new kind of experience that reflected their interests, that was accessible and safe, and that was not a dating app,” explained Matthew Hirst, Head of Server Side Engineering at Hornet. “We had to design our back-end infrastructure to support how this new community app would grow over time, as well as providing that real-time experience.”
Hornet has run on Apache Cassandra® for years in order to support its growth, but the team wanted to consider its approach for the future. “We have used Cassandra for our primary messaging and feed clusters, and they work really well for the Hornet app. However, our internal IT operations team had to manage those clusters and keep them up to date. With these clusters providing essential real-time functionality, each update was a major project for us that took up a lot of time and resources. We wanted to free up our team so they could deal with other problems in our data infrastructure,” explained Hirst.
In addition to this, Hornet is integrating generative AI to enhance user experiences. This involved transitioning from a universal model to a personalized recommendation grid and developing an AI-driven "wingman" feature to facilitate engaging conversations by offering tailored conversation starters based on user preferences and potential matches.
Hirst collaborated with Nate Mitchell, Lead DevOps Engineer at Hornet, on how the team could use Cassandra but take the management headache away. They decided to look at DataStax Astra DB, a cloud-native database as a service built on Cassandra and designed to simplify development and deployment.
Astra DB helps organizations to deploy their infrastructure to the cloud, with management tasks handled by DataStax. This includes automatic scaling and pay-as-you-use pricing, which was particularly attractive to the team at Hornet.
“The approach to pricing was useful to us around our new project deployment, as we could start out small with a pilot and then scale up our deployment as we added more users,” said Hirst.
Alongside the new cluster to support the SPACES launch, the team decided to migrate over the existing clusters to Astra as well. “We looked at the costs for us to continue running our clusters compared to running on Astra, and there was a distinct saving in moving to the cloud. It would also let our team work on other problems, rather than have to concentrate on Cassandra updates and management,” explained Mitchell.
The Hornet team started with the new cluster to support the development of SPACES in order to become familiar with Astra, and found that it was easy to build on.
Following this, the team looked at how to manage the migration of its existing clusters to Astra using the DataStax Zero Downtime Migration Tool. This involved implementing a cloud proxy for all application data coming in that sent transactions to both the existing Cassandra cluster and to the new clusters on Astra. Alongside this, the existing data in the cluster was migrated over in the background to bring the new cluster on Astra up to date.
This strategic migration to Astra DB laid the foundation for Hornet’s AI and vector search initiatives. By simplifying database management and improving platform performance, Hornet was better equipped to harness AI and advanced databases to empower LGBTQ+ communities worldwide.
“With millions of users, delivering relevant and unique recommendations was a significant challenge. Astra DB’s vector search enabled us to identify user preferences and make real-time matches, enhancing user engagement,” said Hirst.
Once the new clusters had the historical data in place, the Hornet team cut over to the Astra deployment that supported the company’s messaging and feed services as well as the new application.
“What was the big win for us? Nothing changed for our users. We delivered the same great response times during and after the migration, and we achieved this with no downtime. Following the move over, we have been able to improve our feed time, while using Time To Live and our new compaction algorithm means that we have the same or better performance for users,” said Mitchell.
Hornet now has its Cassandra clusters running in the cloud with operational, performance, and maintenance tasks handled by DataStax. “Our move was completed without downtime and we can take advantage of the best-in-class expertise that DataStax has around Cassandra management and operations. We also looked at our costs for managing our internal clusters compared to moving to Astra, and we estimate that we will see savings of between ten and twenty percent on our costs,” said Hirst.
Real-time performance is crucial for Hornet to meet user demands. Astra DB ensures that AI-driven features, such as personalized recommendations and safety measures, work seamlessly, providing users with an engaging platform.