Exact Extract — Huawei Training Material, Page 283 (Slide 15):
Up to 72:1 data reduction ratio
Test case: 200 VMs in a virtualization scenario, daily full backup for 28 days, 4% modifications per day, 0.125% new data per day.
Huawei OceanProtect Backup Storage uses a multi-stage data-reduction architecture to optimize backup storage efficiency.
The system combines backup-data preprocessing, multi-layer inline variable-length deduplication, feature-based compression, and byte-level compaction.
During preprocessing, the system identifies data structures and separates information that may interfere with effective duplicate detection.
Variable-length deduplication divides backup streams into content-defined chunks and identifies previously stored segments.
Repeated backup data can reference existing physical information instead of creating redundant copies.
After deduplication, feature-based compression further reduces the physical capacity required for unique data.
Byte-level compaction then improves how the resulting compressed information is organized and stored.
Under the training benchmark, the system achieves a maximum data-reduction ratio of 72:1.
This means that, under the specified conditions, 72 units of logical backup data may be represented using approximately one unit of physical storage.
The actual reduction ratio depends on application characteristics, data similarity, change rates, backup frequency, and retention policies.
A highly repetitive virtual-machine backup environment may achieve greater reduction than encrypted, compressed, or largely unique datasets.
The V3.0 training answer is C — 72:1.