Never forget, always have a backup!
The below paragraphs contain links to outside websites and sources of data; we maintain that the individual must verify the links therein.
If you have already acquired a NAS, as described in DHitMA: Hardware II, you should consider next the topic of RAID: redundant array of independent disks. Utilizing RAID allows you to increase the reliability and/or speed of your data. We will not focus upon those methods of RAID that increase speed alone (like RAID-0, or striping), for the reason that we care principally about data reliability. We maintain that the reader must research their NAS OS' documentation for a full understanding of the implementation of each RAID method before attempting to implement one; additionally, the reader should refer to outside sources if the Internet is available. For most NAS OSes, the WebGUI (which can be accessed without an Internet connection: just turn your NAS on and connect to it with a computer) provides decent documentation for RAID operations. There are several implementations of RAID that can be used to make your stored data more reliable.
- RAID-1: RAID-1, also known as mirroring, requires, for one array, two disks of the exact same size. The make and model should be the same as well. RAID-1 copies data evenly across both disks, so that in the event of a drive failure, the other disk is operable. This allows one to copy the data onto a new disk to rebuild the array. Whatever the size of the disks, that much space is lost in redundant data. It is a simple strategy and performs well, however, the reader may protest the loss of an entire disk's worth of usable space. For the intermediate or starting data hoarder, we recommend RAID-1 fully.
- RAID-5: At least three drives are required for RAID-5. Data is not mirrored like in RAID-1. RAID-5 stripes data across all drives, including a parity checksum. This parity checksum does take up a portion of each drive's space. In the event of a drive failure, this parity checksum allows all data on the failed drive to be accessed. However, if another disk dies in this scenario, all data on the first disk is lost.
- RAID-6: At least four drives are required for RAID-6. RAID-6 is the same as RAID-5, except with double parity checksums. This allows two simultaneous drive failures to be withstood. However, a large portion of the drive spaces will be lost to the parity blocks.
- RAID-10: At least four drives are required for RAID-10. Each drive must be identical. RAID-10 combines RAID-0 and RAID-1, striping and mirroring, respectively. Half of the total drive space is lost to redundancy. Only one disk failure may be tolerated.
- Remember, using RAID does not mean you have a backup!