Case Study: Abusix Eliminates Bottlenecks and Scales Threat Intelligence with Dragonfly

​Abusix: Leading Provider of Network Security Solutions

Abusix is a leading provider of advanced network and email security solutions. Their Guardian platform delivers real-time threat intelligence, dynamic blocklists, and advanced network abuse management tools, helping organizations stay ahead of evolving cyber threats.

Abusix Switched to Dragonfly: A 15x Performance Gain

Abusix’s threat detection algorithm relies heavily on sorted sets. However, as data volumes grew, Redis became a bottleneck. After conducting a proof of concept (POC) with Dragonfly, the team saw a remarkable 15x performance gain, thanks to Dragonfly’s advanced and efficient sorted set implementation. Dragonfly’s B+ tree-based sorted sets not only reduce memory usage by up to 40% but also deliver significantly higher throughput compared to traditional in-memory data stores such as Redis.

Successful Migration to Dragonfly Cloud

Following the successful POC, Abusix migrated their full workload to Dragonfly in just two days. They now operate Dragonfly Cloud data stores with memory capacity of hundreds of gigabytes in the AWS eu-central-1 region.

We migrated to Dragonfly Cloud about a year ago, and since then, our application, which relies heavily on sorted sets, has been performing flawlessly.
— Harsha Rathnayaka, Senior Site Reliability Engineer @ Abusix

Keep Scaling with Dragonfly

As Abusix’s threat intelligence platform grew, their in-memory workload demands doubled recently—but scaling their data store was effortless. With Dragonfly Cloud, they instantly expanded capacity without service disruptions or complex overhauls. The platform’s elastic scaling let them handle peak traffic smoothly while maintaining real-time performance, proving ideal for their rapid growth. Now, whether scaling up for surges or optimizing costs, Abusix can adapt on demand with zero operational friction.

Powering Modern Workloads with Dragonfly

Abusix’s journey highlights how Dragonfly helps businesses tackle high-throughput, real-time workloads, especially those that rely on complex data structures like sorted sets. If your applications need a high-performance data store without the scaling headaches, Dragonfly offers the speed, scalability, and reliability you need.

Editor’s Insights: What Makes These Results Possible?

The performance gains Abusix experienced stem from architectural choices in Dragonfly’s design. Understanding these technical decisions can help you better evaluate whether Dragonfly is a good fit for you.

Dragonfly Sorted Sets Under the Hood

For developers and infrastructure engineers curious about the B+ tree choice, this was a deliberate design decision rooted in performance and efficiency. Dragonfly is a ground-up redesign featuring innovations in threading, memory management, and data structures. However, we do adopt some of Redis’s battle-tested designs, such as skiplists for sorted sets in early prototypes. After extensive benchmarking, we found skiplists’ per-element memory overhead and query latency to be problematic for intensive workloads like Abusix’s.

Learn More: Dragonfly’s Sorted Set Implementation

The modified B+ tree implementation preserves all sorted set operations while delivering predictable performance and compact memory. We’ve maintained Redis compatibility while advancing the state of the art—proof that stable APIs and cutting-edge internals are not mutually exclusive. For implementation details, see the deep dive of Dragonfly’s sorted set implementation.

Fearless Migration

Thanks to Dragonfly's compatibility, switching to Dragonfly in your backend applications typically only requires updating your connection string. For data migration, we recommend RedisShake—a battle-tested tool widely used by the community. Learn more in our live migration guide. As mentioned by Abusix, we are always here to help if you have any questions.

Learn More: Live Migration to Dragonfly Cloud

Ready to supercharge your data layer? Explore Dragonfly Cloud today and experience effortless scaling, robust high availability, and unparalleled efficiency—all in a fully managed solution.