96-day study measures SRT streaming behavior and UDP flood effects
Original title: Temporal Dynamics and Security Assessment of SRT Protocol: A 96-Day Longitudinal Analysis of Production Streaming Infrastructure
Controlled 500-pps UDP floods raised CPU 176–806% while SRT metrics moved only about 13%.
A paper reports a 96-day study of Secure Reliable Transport on a production stream between Iraq and Turkey, using 262,672 transport samples and 5,033,104 system measurements from October 2025 to January 2026. NAK rates varied by hour and day, but the authors say time period explained under 4% of variance, and CPU utilization was the strongest correlate of transport variation. In controlled tests, UDP floods at 500 packets per second sharply raised CPU while SRT metrics changed little, and a Z-score detector caught all seven security events.
- NAK showed 7.6-fold hourly and 11.5-fold daily variation (p < 0.001).
- Effect sizes were small, with eta^2 under 0.04.
- CPU utilization correlated with NAK rate at r = 0.617.
- An anomaly reached 97.7% CPU and helped distinguish attack types.
- Authors call this the first controlled security assessment of SRT infrastructure.
Impact: UDP floods drove large CPU spikes on the streaming host
| Product | Affected |
|---|---|
| Secure Reliable Transport | — |
- 1.Use multi-metric monitoring and Z-score anomaly detection on production SRT deployments.
- Products
- Secure Reliable Transport
From the original article
1 min read at arxiv.orgThis study presents a 96-day longitudinal analysis of Secure Reliable Transport (SRT) protocol behavior in a production streaming environment between Iraq and Turkey, examining 262,672 transport metric samples and 5,033,104 system resource measurements collected between October 2025 and January 2026. The research addresses temporal patterns, system-resource relationships, attack vulnerability, and anomaly detection.
By Ahmed F. K. Koysha · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.