Digital Audio exists in different forms within equipment and on the cables used to interconnect it. Internally it exists as a sequence of parallel digital numbers representing the audio waveform and in dedicated hardware often as one of a number of serial formats. All data transfers are synchronised to a Word Clock at the sampling frequency, but may occur at various times within that period. Usually different pieces of equipment are synchronised to a single master Word Clock, a higher (e.g. x256) multiple of the word clock or a video signal ("Black and Burst"). This may be a separate signal or may be derived from the digital audio signal.
Externally a serial data protocol is used to transport data via cables, this contains other information as well as the audio. The most common is AES/EBU and S/PDIF for stereo signals. The original AES digital audio proposal was RS422 using existing XLR audio cables and this was modified by the EBU to include isolating transformers because the AES forgot about grounding problems with analogue audio [sic.]. A cheaper coaxial system was derived by Sony and Philips for consumer applications and was known as S/PDIF. Although the data structures are similar there are differences for professional and consumer applications and before long professional format was being transmitted on S/PDIF and consumer format on AES/EBU. This is now ratified as IEC60958 and either format may be transmitted on either medium and most equipment will recognise the audio in either format.
Other formats are TDIF and ADAT Lightpipe for 8 channels and MADI for higher channel counts on professional recording devices. There are also several varieties of multiple AES/EBU pairs on D-type connectors--all mutually incompatible!
In general, the more that is squeezed onto one cable the higher the signal frequencies used have to be and these will always be above 4MHz. This means that considerations that were unnecessary for analogue cabling can no longer be ignored. The cabling has to be regarded as a Transmission Line.