#3 Hardware Validation Programme — Community NVMe Test Reports

[Anonymous]

Originally created by: yonasabeselom

Purpose

This issue tracks the AAD-50 empirical hardware validation programme.

The USENIX FAST '27 paper submission requires hardware test results across multiple NVMe drive manufacturers, controller generations, and NAND geometries. This issue is the central collection point for all community-contributed test reports.

---

How to Submit Your Result

Run AAD-50 on a drive you are retiring and post your result as a comment on this issue using the template below.

Both dry-run and live results are welcome. Dry-run results confirm the framework works correctly on your system. Live results are the priority for the empirical validation programme.

---

Report Template

Platform:         [Windows 11 / Ubuntu 24 / etc.]
Drive Model:      [full model string]
Drive Capacity:   [e.g. 1 TB]
NAND Type:        [MLC / TLC / QLC / unknown]
Controller:       [if known]
Mode:             [--dry-run / live]
AAD-50 Version:   [1.1 / etc.]
All 50 Cycles:    [yes / no - if no, which cycle failed?]
SHA-256 Hash:     [from PDF Certificate or JSON audit log]
Outcome:          [SUCCESS / FAULT]
Notes:            [anything unusual - slow cycles, SSTAT delays, anomalies]

---

What We Need Most

• Different manufacturers - Samsung, WD, Seagate, Micron, SK Hynix, Kioxia, Kingston, Crucial
• Different NAND types - MLC, TLC, QLC
• Different capacities - 256 GB through 4 TB+
• Both Linux and Windows
• Any anomalies - slow SSTAT reporting, inconsistent completion, unexpected behaviour

Even a single result from a manufacturer not yet in the table is a meaningful contribution.

---

Current Status

• [x] Dry-run confirmed on Windows 11 — WD PC SN730 SDBQNTY-256G-1001
• [x] Dry-run confirmed on Linux — GitHub Codespaces kernel 5.15
• [ ] Live hardware results — open for community contribution

---

Repository

github.com/yonasabeselom/aad50