The HPE Ninja Online Block Storage Sizer scenario prescribes a default of 20 percent non-reducible data when the customer ' s actual data composition is unknown. Non-reducible information includes encrypted files, previously compressed archives, certain multimedia formats and random-looking application datasets that are unlikely to benefit substantially from another compression or deduplication pass. Assuming excessively good reduction can cause an Alletra Storage MP B10000 proposal to be undersized. An estimated 20 percent non-reducible fraction provides a planning assumption for a mixed workload until representative evidence is obtained. To understand the effect, consider 80 percent of data reducing at 4:1 and the remaining 20 percent not reducing at all: the combined reduction is approximately 2.5:1, not 4:1. Different data mixes yield different effective capacities even on identical SSD hardware. The architect should document this assumption and later refine it using sampled datasets, existing storage analytics or customer application profiles. Procurement should consider growth margin and operational free-space headroom as well as the assumed reduction ratio.

Study Guide Reference: B10000 capacity sizing, non-reducible data defaults and compaction assumptions.

Official reference(s): B10000 QuickSpecs