Rating
1500
Battle Count: 0
Relevance
10/10
Directly addresses the core infrastructure challenge of HFT: minimizing latency tails in market data ingestion. Provides actionable design rules for single-thread vs. multi-thread architectures based on empirical feed characteristics.
Implementation Complexity
7/10
Requires deep understanding of queueing theory, point processes (Hawkes), and low-latency C++ system design. Implementing the simulator is moderate, but applying the design rules requires precise measurement of feed-specific parameters (publisher period).
Reproducibility
4/5
Code is open source in the Kaspar-HFT repository. However, the raw packet captures (pcap) are not redistributable, limiting full external replication of the specific dataset, though the methodology and simulator are provided.
About this paper
Methodology: Measurement Study and Simulation. Problem types: System Design, Latency Optimization, Time Series Analysis, Queueing Theory.
The interactive Everscope explorer (charts, battles, favorites) loads below.