This work evaluates the oxygen reduction system (ORS) approach to fire prevention in large-scale fire tests. A two-tier fuel array of standard commodities is used in a rack-storage configuration within an enclosure. A constant nitrogen/air mixture flow is supplied to the enclosure at a desired oxygen concentration. The oxygen concentration varies from 9% to 17%. A premixed propane ignitor is used as ignition source. The tested materials include five standard commodities of Class 3, CUP, CEP, UUP and UEP. The results of fire propagation success are obtained for the five standard commodities under different oxygen concentrations with a sustained igniter (hard limits) and without a sustained igniter (soft limits). The resulting limiting oxygen concentration (LOC) values are shown to be generally lower than the oxygen design concentrations recommended by existing standards including VdS 3527 and EN 16750 due to different test conditions. The hard limits are close to the fundamental LOC values for gases and vapors and do not depend significantly on the ignition duration and array size, while the soft limits vary significantly with the size and configuration of the fuel array and ignition duration. It is concluded that the hard limits are more suitable for ORS design purposes.