Dragonfly

Dragonfly vs. Redis

A Drop-In Redis Replacement That’s 25x More Performant

Redis built the foundation for in-memory data. Dragonfly rebuilt it for the cloud era — delivering up to 25x higher throughput, 80% lower cost, and zero operational complexity.

Fully Redis-compatible. Instantly faster. Backed by open benchmarks you can verify yourself.


Throughput in QPS
125k
3900k
SET
130k
3800k
GET
115k
4300k
SETEX
Redis
Dragonfly
QPS benchmark on AWS c6gn.16xlarge. Snapshot benchmark on AWS c6gn.4xlarge. Source.
Dragonfly triangle background

Redis Struggles to Scale. Dragonfly Doesnt.

Redis’s single-threaded core was revolutionary in 2009 — but in 2025, it’s not able to scale to handle demanding modern workloads. Redis scales via an inefficient clutter architecture that wastes resources and is overly complex, introducing network bandwidth issues, performance headaches, and ballooning cloud bills.

Dragonfly takes a different approach: a multi-threaded, non-locking architecture that uses every CPU core in parallel. The result is simple: vertical and horizontal scalability at the same time that results in much better performance on a smaller infrastructure footprint.

DragonflyDB logo

Dragonfly

  • Redis API Compatibility
  • Memcached API Compatibility
  • QPS per Instance
    3.9M
  • Snapshotting Speed
    1260MB/s
  • Lua
    5.4.4
  • Async Core
  • LRFU Eviction
  • Native OpenTelemetry
  • Threading Model
    Multi-Threaded & Shared Nothing
  • Memory Efficiency
    40% Less Memory with Dashtable
  • Deployment & Operations
    Simplified Operations
  • Scalability
    Vertical & Horizontal Scaling

Redis

  • Redis API Compatibility
  • Memcached API Compatibility
  • QPS per Instance
    150K
  • Snapshotting Speed
    107MB/s
  • Lua
    5.1
  • Async Core
  • LRFU Eviction
  • Native OpenTelemetry
  • Threading Model
    Single-Threaded with Enhanced I/O
  • Memory Efficiency
    Traditional Hash Table
  • Deployment & Operations
    Complex Cluster Management
  • Scalability
    Premature Horizontal Scaling

Real-World Benchmarks

25x More Throughput. 17x Faster Reads. Same Hardware.

In standardized AWS benchmarks (c6gn.16xlarge instances, YCSB workloads), Dragonfly consistently outperformed Redis across read-heavy and write-heavy operations — with no tuning or clustering required.

WorkloadWrite-heavy (SET)
Redis QPS125k
Dragonfly QPS3.1M
Improvement25x throughput
WorkloadRead-heavy (GET)
Redis QPS240k
Dragonfly QPS4.2M
Improvement17x throughput
WorkloadMixed (80/20)**
Redis QPS185k
Dragonfly QPS3.7M
Improvement20x throughput

Test details: AWS c6gn.16xlarge instances (64 vCPUs, 128 GB RAM) running YCSB default workloads.

Why Teams Switch to Dragonfly

Scaling icon

Simpler scaling

One node replaces many

Lighting Squircle

Predictable performance

Linear throughput with modern hardware

Play Squircle

Instant migration

Drop-in compatible with Redis clients.

Dragonfly logo

Try Dragonfly in Minutes

Run Dragonfly locally, in your cloud, or side-by-side with your current Redis environment.