ZK-Storage

Reproducible third‑party benchmarks for storage in AI training

Published 2026-07-14 · ZK-Storage Insights

Storage is the hidden ceiling for many AI training clusters: top‑tier GPUs can sit idle while jobs wait on data. Reproducible third‑party benchmarks for storage aim to expose that ceiling in a way that enterprise buyers can trust, compare, and act on.

Why reproducible storage benchmarks matter for AI training

AI training is a complex I/O workload that couples dataset access patterns, framework behavior (prefetching, shuffling), and multi‑GPU orchestration. A storage benchmark that reports raw MB/s or IOPS without context is not actionable. Reproducibility requires a documented test harness, stable testbed configuration, workload traces that resemble real training jobs, and statistical reporting practices that make results comparable across vendors and labs.

Core evaluation criteria (what to measure)

Testbed design for reproducibility

  1. Hardware and software inventory: list exact GPU models, CPU, NICs, kernel, drivers, storage firmware, and firmware versions. Use the same OS kernel and framework versions across runs.
  2. Fabric and topology: document switch, cable types, and whether NVMe‑oF/RDMA is used. Disaggregated all‑flash platforms (NVMe‑oF) will differ from SAN/NAS behavior.
  3. Workloads: include at least one real training job (e.g., a transformer or convnet run with defined batch size and dataset sampling), plus microbenchmarks (fio profiles for streaming and random I/O). Archive and publish the workload scripts and data access patterns where licensing allows.
  4. Repetition and seeds: run each scenario multiple times (≥5) with cold and warm caches, capture medians and tails.
  5. Telemetry: collect GPU metrics (utilization, PCIe throughput), storage telemetry (latency percentiles, queue depth), and host metrics (CPU, NIC). Tools: nvidia‑smi/DCGM, iostat/blktrace, perf, and fio for storage probing.

Scenario matrix: what to test (four recommended scenarios)

Each scenario requires a tailored mix of batch streaming and small random workloads. For example, training clusters emphasize sustained reads and prefetch behavior, while inference focuses on p99/p99.9 latency isolation.

Reproducibility checklist (must‑have items to publish)

Comparison table (high‑level tradeoffs)

Criterion Local DAS (per‑node NVMe) Shared SAN/NAS Disaggregated NVMe‑oF (e.g., WS5000) Cloud Block Storage
Typical latency (small I/O) Low Moderate‑high Low‑moderate Varies, often moderate
Sustained bandwidth per GPU Limited by local media Can be high if provisioned Scales with fabric and pooling Elastic but variable
Multi‑tenant QoS Weak Often better with controller Stronger QoS designs possible Provider controls QoS
Scalability for many GPUs Shard per node, scaling by adding SSDs Scale with arrays, more complex Designed to scale independently of compute Easy to scale but opaque
Reproducibility for third‑party tests Easier to reproduce per‑node Needs array config details Requires fabric/topology disclosure (but reproducible) Hard to reproduce exact conditions
Typical use case Single‑node high performance Enterprise file shares Training farms / retrofit Variable, hybrid workloads

Note: the WS5000 is an example of a disaggregated all‑flash NVMe‑oF appliance designed to address GPU starvation patterns; independent validation is ideal for trust-building. See https://goni.top for vendor details and published validation artifacts where available.

Interpreting results and trade‑offs

Practical recommendations for buyers and lab analysts

Key takeaways

Resources and next steps: collect a reproducible test harness (fio job files + ML training script), standardize your topology descriptions, and request raw telemetry from vendors or third‑party labs. For vendor specifics and published validation artifacts you can review vendor documentation such as ZK‑Storage’s WS5000 materials at https://goni.top.