There are several fault-tolerance options to a System Administrator depending upon the integrity requirements, the availability requirements, and the recovery issues of the application being deployed. The Compaq SMART Array Controller RAID levels relevant to this discussion include 0, 1, & 5. RAID-0 simply refers to data striping multiple disks into a single logical volume and has no fault-tolerance. RAID-1 and RAID-5 are both fault-tolerant, but require a different amount of drives to accomplish the fault-tolerance. RAID-1 or drive mirroring uses 50% of the drives in a volume for the fault-tolerance. RAID-5 requires only a single drive worth of the volume's total capacity for the fault-tolerance. For example, if a RAID-5 volume is configured with 6 1GB drives, the usable capacity of the volume is 5GB. With RAID-1 this same 5GB's of usable space requires 10 1GB drives. Which level of RAID to use is completely dependent on the individual situation. No one alternative is best for every application. Some applications may need complete redundancy and the System Administrator is willing to pay for it; while others may be willing to risk some down time for a more cost-effective implementation. In addition, the I/O loads generated by a particular application may fit very well in the performance constraints of RAID-5; while other applications will suffer serious performance penalties with RAID-5. This section explains some of the trade-offs associated with each alternative and possible performance implications, but it is up to the implementor to evaluate each alternative's applicability to their environment.