In Statistical Multiplexing the communication link is shared among different units which are pretty similar Dynamic Bandwidth Allocation. In this communication channel is divided into variable bit rate digital channel or data streams. The sharing is adapted as per the traffic requirement on the channel; it is an excellent way of dividing the channel dynamically over the various traffic demands which are not fixed hence it works best for the Packet Switching whose bandwidth demands vary over the period of time rather than fixed as it is being done in TDM and FDM. It provides better channel utilization which is known as the statistical multiplexing gain.
It is Statistical Multiplexing in which packet mode or packet-oriented communication it is this feature which is used in packet switching of the computer networks. In this each stream is divided into several packets which are delivered on the basis of first come first serve.
In Time Division Multiplexing (TDM) the channel is divided for each user with the set period of time, even if the user is not going to use the channel the channel slot is left vacant as nothing is sent over the channel, hence wastage of resources/channel bandwidth. Moreover it is difficult to accommodate newer node as the channel TDM is fixed hence every time newer configuration is needed to be done, whereas in Statistical Multiplexing the division of channel depends upon the number of users, channel bandwidth, traffic etc. and is lot more acceptable where newer nodes are there in the channel and even leaves the network randomly. This helps in maintaining the better throughput and overall network performance.