RAID 0 (striping): data is cut into chunks, these chunks are spread out over the disks in the array. retains full capacity of all disks, very fast, very insecure (1 failed disk generally means all data is lost)
RAID 1 (mirroring): data is mirrored over all disks (usually 2). write performance equals a single disk, read performance can be better because data can be read from multiple disks, actual read performance depends on implementation. same capacity as a single disk, little performance gains, secure (if one disk fails the data is still on the other disk)
RAID 5: data is split into chunks, which are spread over (n-1) disks (i.e. data is spread over all disks but 1). a checksum is generated which is stored on the final disk. this checksum data is spread over all disks in the array. if any one disk fails all data remains intact, missing chunks from the failed disk can be calculated using the remaining data and the checksum. capacity of n-1 disks (you loose the capacity of 1 disk). speed is usually good, but depends on the cpu doing the checksum calulation (dedicated RAID controllers are very fast, affordable NAS boxes usually do this in software on a normal CPU). secure (if 1 disk fails data remains intact. 2 simultaneus failures kill all data)
RAID 6: same as RAID 5, but 2 checksums are generated using 2 different algorithms. you loose 2 disks worth of capacity for this parity data, but data remains intact if 2 disks fail. really only interesting for 8 or more disks.
Facebook
Twitter
Instagram