Build an I/O Benchmarking Harness for an Edge Storage Appliance
Overview
What this challenge is about.
Receive the appliance specs (4x 7.68TB Gen4 NVMe, ZFS, Linux kernel 5.15), the 3 target workload profiles (4KB random read at QD32, 1MB sequential write at QD8, mixed 70/30 read/write 16KB at QD16), and the existing vendor datasheet (with no methodology). Build a benchmarking harness wrapping fio, run each workload for 30 minutes after a 5-minute warmup, capture latency percentiles (p50/p99/p99.9), throughput, IOPS, queue depth via iostat. Cross-validate fio's reported latency against blktrace for the small-block workload. Identify the appliance's bottleneck per workload (NVMe queue saturation, ZFS ARC pressure, CPU on checksum). Deliver the harness scripts (reproducible), the raw + analyzed dataset, a 6-page engineering-defensible datasheet, and a 2-page sales-readable summary.
The Brief
What you'll do, and what you'll demonstrate.
Build a reproducible fio + iostat + blktrace benchmarking harness for an NVMe appliance and produce an engineering-defensible performance datasheet across 3 workloads.
Earning criteria — what you'll demonstrate
- Design fio workloads that match real customer patterns
- Cross-validate user-space latency with kernel-level blktrace
- Identify per-workload bottlenecks across NVMe, filesystem, and CPU
- Communicate benchmark methodology that survives technical scrutiny
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Career mappings coming soon.