An emulation-driven study finds that learning-based congestion control schemes underperform in dynamic LEO conditions while BBRv3 and some loss-based schemes handle throughput better, exposing fairness and latency issues.
BBRv3: Algorithm Bug Fixes and Public Internet Deployment,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NI 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Evaluating Learning Congestion control Schemes for LEO Constellations
An emulation-driven study finds that learning-based congestion control schemes underperform in dynamic LEO conditions while BBRv3 and some loss-based schemes handle throughput better, exposing fairness and latency issues.