MTRlens

MTR link quality · Mathis formula

# MTRlens **Link quality from MTR output · Mathis formula** A Chrome extension that turns MTR (My Traceroute) report output into a clear quality score and comparison. Paste your MTR results, get a 0–100 score based on the Mathis TCP throughput model, compare multiple paths side by side, and export or print reports. --- ## What it does - **Score a single path** — Paste MTR report output, name the path (e.g. “LA CN2”, “SJC 163”), and get a 0–100 score, RTT/loss/jitter, Mathis Q, estimated throughput, and short diagnostics. - **Compare multiple paths** — Add several MTR runs; the extension ranks them and shows a comparison chart. When scores are close, it warns about Braess-style congestion risk and suggests splitting traffic by score ratio. - **Recognize backbones** — Detects common cross-border backbones in the hop list (e.g. CN2, 163, 9929, CMI) and labels the line type. - **History & export** — Recent analyses are saved (configurable limit). Export a text report to the clipboard or open the report in a new tab for printing/PDF. Scoring is tuned for **cross-border links** (e.g. from China to overseas). Endpoints with RTT < 100 ms are treated as non–cross-border and not scored. The last hop is always used for metrics; intermediate hops with very high loss are treated as ICMP noise and skipped when choosing the endpoint. --- ## How the score works The main score is based on the **Mathis approximation** for TCP throughput: - **Q ∝ 1 / (RTT × √loss)** Q is mapped to 0–100 using reference bounds (Q ≈ 1 for barely usable, Q ≈ 40 for premium cross-border quality). Higher score means better throughput potential. - **Jitter** is not folded into this Q; it is applied as a separate penalty so that high jitter reduces the final score without distorting the RTT/loss relationship. - **Single path:** one absolute score. **Multiple paths:** a mixed display score (absolute × 0.7 + relative × 0.3) so you see both “how good this path is” and “how it ranks in this set.” Ranking is by the relative score; when all paths are very close (Q spread < 3), the UI falls back to absolute score and notes that differences are within measurement noise. --- ## How to use 1. Run MTR in **report mode** (e.g. `mtr -r -c 100 <target>` on Linux/macOS, or WinMTR in report mode on Windows). 2. Copy the full output and paste it into the extension’s **Input** tab. 3. Enter a line name and click **Analyze**. 4. View the result on **Analyze** (single-path details and timeline), then use **Compare** if you have multiple paths, and **Report** (in a new tab) for export/print. The in-extension **Help** page has step-by-step instructions and algorithm notes. --- ## Tech - **Chrome extension**, Manifest V3, no external dependencies. - **Languages:** English, 简体中文, 繁體中文 (UI and reports follow browser language). - **Privacy:** All processing is local; no data is sent to any server. --- ## Permissions - **clipboardRead** — Used to read pasted MTR text and to offer “paste from clipboard” when MTR-like content is detected. Optional; you can type or drag a file instead. ---