← Back
Creator Hardware, Storage, NAS Archiving & Technical Workflows

A 10GbE link negotiates correctly but the NAS disks cannot sustain enough aggregate throughput for four simultaneous 4K streams, causing cache starvation.

Problem

A 10GbE link negotiates correctly but the NAS disks cannot sustain enough aggregate throughput for four simultaneous 4K streams, causing cache starvation.

Solution

Root Cause / Diagnostic:
10GbE networking provides up to 1100MB/s of theoretical bandwidth, but mechanical hard drive arrays have strict sequential read limits. A 4-drive or 6-drive RAID array of HDDs provides ~600–800MB/s aggregate throughput; when four editors stream 4K uncompressed/ProRes media simultaneously, disk head seeking destroys throughput, starving NLE playback buffers.

Actionable Fix:
1. All-Flash NVMe / SSD Cache Tiering: Implement an all-flash NVMe/SSD pool or high-speed NVMe read cache on the NAS dedicated exclusively to active editing projects.
2. Multi-Tier Media Architecture: Keep active multi-stream timelines on local workstation scratch drives or flash NAS pools, archiving dormant media to mechanical hard drive pools.
3. Aggregate Drive IOPS Benchmark: Benchmark aggregate NAS pool disk throughput using FIO (`fio --rw=randread --bs=1M --numjobs=4`) to determine true multi-client throughput limits.

Pro Tip:
A 10GbE network connection cannot make slow mechanical hard drives faster; if four editors are streaming 4K footage at once, you need an all-flash SSD array on your NAS to prevent buffer starvation.