@nihal, so we are only talking about basic RAID levels with parity e.g RAID-5 and RAID-6
In a 3 disk RAID-5 set, we have three disks D1, D2 & D3 comprising LUN 1 which is mapped to say the R: drive on your system. for the sake of this explanation we will only write 100 bytes to each disk
When a write is made to the R: drive the data is split up into 100 Byte chunks, so the first 100 bytes (C1) are written to D1, the second 100 bytes (C2) to D2, and those two chunks(C1, C2) are then XOR’d together and the resultant Parity (P1) is written to D3, so thats stripe 1, on subsequent stripes the chunks are rotated. e.g.
D1 D2 D3
S1 C1 C2 P1
S2 P2 C3 C4
S3 C5 P3 C6
On normal reads the Parity chunk is ignored, but if say D2 fails then when reading stripe S1, Chunk C1 is XOR’d with Parity P1 and this gives the data that was stored in C2, thereby providing the redundancy that RAID-5 is known for