FAT32 Limits in a Video-Recording Lab
A polytechnic media lab has 30 Windows PCs and three 1 TB USB hard disks rotated between benches. Students record practical demonstrations at about 25 MB per minute; a typical long recording is 3 hours, and each weekly batch produces 12 such videos plus 20,000 small text/CSV files averaging 2 KB. The disks are formatted FAT32 with 32 KB clusters. Symptoms: a 4.5 GB video refuses to copy with “not enough space” even though one disk shows 180 GB free; copying the small files takes several minutes; and the disk's used-space figure is about 600 MB larger than the sum of the file sizes. The lab can reformat one disk this week, but the same disk must be readable on Windows, Linux, and a smart TV, and no new hardware can be bought. Existing files must survive the change.
1. Explain the 4.5 GB failure and the roughly 600 MB discrepancy using the FAT32 cluster size. Show calculations.
Read the worked answer
The 4.5 GB failure is not caused by total free space. FAT32 has a maximum file size of 4 GB minus 1 byte. A 4.5 GB video exceeds that per-file limit, so the copy fails even with 180 GB free. The disk would have to split the file, which FAT32 does not do transparently.
The discrepancy is cluster slack. Each file occupies a whole number of clusters. With 32 KB clusters, a 2 KB file still occupies one 32 KB cluster:
- Allocated: 20,000 × 32 KB = 640,000 KB ≈ 625 MB
- Actual data: 20,000 × 2 KB = 40,000 KB ≈ 39 MB
- Wasted: about 586 MB, close to the observed 600 MB
So the “used space” figure counts allocated clusters, not the bytes actually written. The small-file copy is also slow because 20,000 separate directory entries and cluster allocations must be created.
2. Recommend a file system and cluster size for the reformatted disk. Justify the choice against all three readability constraints, and state one risk with a mitigation.
Read the worked answer
Reformat to exFAT with 32 KB clusters. exFAT removes the FAT32 4 GB per-file limit, so the 4.5 GB videos fit. Windows supports exFAT natively; current Linux kernels include exFAT support; and many smart TVs that reject NTFS will read exFAT, although support is model-dependent. NTFS would work well on Windows and Linux but is more likely to be read-only or unsupported on a television. FAT32 cannot meet the video requirement at all.
32 KB clusters are a reasonable compromise for this mixed workload. Larger clusters reduce file-system overhead for the big videos but increase slack for 20,000 small files. For example, 64 KB clusters would waste about 62 KB per 2 KB file, roughly doubling the slack.
Risk: the particular smart TV may not recognise exFAT. Mitigation: test one reformatted disk on the TV before converting the others, and keep a FAT32-formatted copy of any files that must be played directly on the TV.
3. If the lab refuses to reformat, what is the longest 25 MB/min recording that can be stored on FAT32? Show the calculation and give one practical workaround.
Read the worked answer
Use the FAT32 per-file ceiling of 4 GB = 4,096 MB. At 25 MB per minute:
4,096 MB ÷ 25 MB/min = 163.84 min
That is 2 hours 43 minutes 50 seconds. In practice, leave a margin for container overhead and file-system rounding, so plan for about 2 hours 40 minutes.
A workaround is to record in segments that each stop below 4 GB-for example, split every 2 hours-and join or play them as a set. This avoids the per-file limit but is less convenient and may cause a small break at each split. It does not fix the small-file slack problem, so exFAT remains the better long-term fix.