On the Low-latency Region of Best-effort Links for Delay-Sensitive Streaming Traffic
Boris Bellalta

TL;DR
This paper investigates the low-latency region of best-effort links supporting delay-sensitive streaming and background traffic, identifying an optimal fair traffic allocation that balances throughput and delay, validated with real-world traffic traces.
Contribution
It introduces a proportional fair allocation framework within the low-latency region for mixed traffic, based on a throughput-delay trade-off, applicable to realistic streaming scenarios.
Findings
Existence of a proportional fair traffic allocation in the LLR.
The throughput-delay trade-off holds across various traffic distributions.
Validation with Google Stadia traces confirms practical applicability.
Abstract
This letter analyzes the Low-latency Region (LLR) of a best-effort link (i.e., no traffic differentiation, and first come first serve scheduling) carrying both delay-sensitive (DS) streaming and non-delay-sensitive (NDS) background traffic. Moreover, inside the LLR, we show it exists a proportional fair arrival rate allocation for both the DS and NDS traffic streams. This optimal operating point results from maximizing a simple throughput-delay trade-off that considers the NDS traffic load, and the mean delay of the DS packets. To show how the presented trade-off could be used to allocate NDS traffic in a realistic scenario, we use Google Stadia traffic traces to generate the DS flow. Results from this use-case confirm that the throughput-delay trade-off also works regardless the distribution of the packet arrival and packet service times.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
