Why OFDM?

Problem : How to distribute the bandwidth with the multiple user?

Solution : divide entire bandwidth into multiple small bandwidths (sub-carrier) and distribute sub-carrier to every user ->  FDM (Frequency Division Multiplexing)

Problem : What happen if one user has more data than other user?

Solution : one user can divide its data  into multiple streams (serial to parallel) and send over multiple sub-carrier -> known as MC (Multi-carrier Communication)

Problem : How to avoid interference with subsequent sub-carrier?

Solution-1 : put a guard band in between sub-carrier. But, guard band cannot transmit anything then it introduce frequency inefficiency.

Solution-2: make sub-carrier orthogonal. So there is no interference b/w sub-carriers and they can be overlap to each other and save bandwidth -> known as OFDM

Problem : one sub-carrier requires one pair of transmitter and receiver. Therefore,  we require multiple transmitter and receiver to support multiple sub-carriers. How to solve this issue?

Solution : If we put an algorithm to integrate all parallel signal at transmitter and distribute back to parallel signal at receiver. Then we need only one pair of transmitter and receiver. Fourier Transformation fulfills that requirement. IFT(inverse FT) converse parallel signals and FT distribute back to parallel signals.

Problem : Does FT is capable to handle continuous signals efficiently? No… How to Solve this?

Solution : Use DFT(discrete Fourier Transformation), sample the original continuous signals and get a series of discrete sample values. And optimize further for fast processing using FFT(Fast Fourier transformation).