A Policy-Aware Cross-Layer Auditing Service for Tiering and Throttling in Starlink

TL;DR

A policy-aware, cross-layer methodology for edge-side auditing of service tiering and quota-based throttling in Starlink is presented and it is shown that a lightweight rule on windowed medians separates high-speed from low-rate operation without operator visibility.

Abstract

We present a policy-aware, cross-layer methodology for edge-side auditing of service tiering and quota-based throttling in Starlink. Using a multi-week plan-hopping campaign (232.8 h) on a UK residential terminal, we align 1 Hz terminal telemetry with host-side probes to obtain portal-labeled traces spanning priority (pre-quota), post-quota throttling, stay-active operation, and residential service. Using portal status only as ground truth (independent of throughput), we show these policy regimes manifest as distinct signatures in goodput, PoP RTT, and an internal-to-user ratio $R=C_{\mathrm{int}}/T_{\mathrm{user}}$. A lightweight rule on windowed medians separates high-speed from low-rate operation without operator visibility.

Figures (5)

• Figure 1: Starlink deployment and measurement setup (WL: Wireless Link, EC: Ethernet Connection).
• Figure 2: Download throughput and RTT distributions across representative tariff states.
• Figure 3: Distribution of the internal-to-user ratio $R=C_{\mathrm{int}}/T_{\mathrm{user}}$ across tariff states, where $C_{\mathrm{int}}$ is a terminal-reported internal downlink indicator and $T_{\mathrm{user}}$ is host-side TCP goodput. Ratios are clipped for visualization; clipped counts are annotated.
• Figure 4: Quota depletion, enforcement-delay window $G$, and throttling onset for the priority plan.
• Figure 5: Service-tier fingerprints in the (throughput, $R$) feature space. High-speed and low-rate regimes form disjoint clusters on our dataset; the horizontal and vertical thresholds ($T_{\mathrm{d}}, T_{\mathrm{r}}$) illustrate a conservative decision rule.