Built by Trading Technology Specialists
Lunyn was founded to solve a specific problem: existing market data infrastructure couldn't deliver the performance characteristics modern trading strategies demand.
Our work focuses on practical market data processing, clear benchmark methods, and usable systems architecture.
Areas of Expertise
- Low-latency systems architecture
- Financial market data protocols (ITCH, FIX, OUCH)
- High-performance computing infrastructure
- Production trading system operations
- Real-time order book management
Technical Excellence, Practical Results
We believe institutional trading infrastructure should be:
Fast
Measured benchmark costs include 5.2 ns/message for framing and 9.9 ns/message for full decode
Reliable
Validated on a complete historical NASDAQ ITCH 5.0 session
Transparent
Reproducible performance, not marketing claims
Practical
Solves real problems, not demo scenarios
Architectural decisions are documented against defined workloads. Performance numbers are reported with the dataset, hardware, and processing mode used for measurement.
Technical Contributions
We publish research on high-performance parsing, systems architecture, and market microstructure.
Deep dives into parsing optimization, SIMD techniques, and lock-free data structures
Visit BlogBenchmark suite and evaluation tools available for community validation
View GitHubComplete methodology and reproduction instructions for all performance claims
View BenchmarksOur work advances the field of market data processing infrastructure.
Building the Future of Market Data Processing
Trading strategies are only as good as the data that powers them. If market data processing is slow or unreliable, research capacity and execution quality are held back.
Lunyn was created to remove these barriers. We approached parsing architecture from first principles, using zero copy operations, lock free concurrency, and SIMD vectorization. This led to performance that was once considered unattainable on standard hardware.
213.7 M messages per second for single-thread framing. 5.2 ns/message framing average and 9.9 ns/message full decode average on the documented benchmark session.
These results are not theoretical. They can be reproduced and are trusted by institutional trading firms for essential operations.
Let's Discuss Your Infrastructure
Schedule a technical consultation to explore how Lunyn can solve your market data processing challenges.