Statistical multiplexing is a type of communication link sharing, very similar to dynamic bandwidth allocation (DBA). In statistical multiplexing, a communication channel is divided into an arbitrary number of variable bitrate digital channels or data streams. The link sharing is adapted to the instantaneous traffic demands of the data streams that are transferred over each channel. This is an alternative to creating a fixed sharing of a link, such as in general time-division multiplexing (TDM) and frequency division multiplexing (FDM). When performed correctly, statistical multiplexing can provide a link utilization improvement, called the statistical multiplexing gain.
Statistical multiplexing is facilitated through packet mode or packet-oriented communication, which among others is utilized in packet-switched computer networks. Each stream is divided into packets that normally are delivered asynchronously in a first-come-first-served fashion. In alternative fashion, the packets may be delivered according to some scheduling discipline for fair queuing or differentiated and/or guaranteed quality of service.
Statistical multiplexing of an analog channel, for example, a wireless channel, is also facilitated through the following schemes:
- Random frequency-hopping orthogonal frequency division multiple access (RFH-OFDMA)
- Code-division multiple access (CDMA), where a different amount of spreading codes or spreading factors can be assigned to different users.
- Statistical multiplexing normally implies an "on-demand" service rather than one that preallocates resources for each data stream. Statistical multiplexing schemes do not control user data transmissions.
Examples of statistical multiplexing are:
- The MPEG transport stream for digital TV transmission. Statistical multiplexing is used to allow several video, audio and data streams of different data rates to be transmitted over a bandwidth-limited channel (see Statistical multiplexer). The packets have constant lengths. The channel number is denoted Program ID (PID).
- The UDP and TCP protocols, where data streams from several application processes are multiplexed together. The packets may have varying lengths. The port numbers constitute channel identification numbers (and also address information).
- The X.25 and Frame relay packet-switching protocols, where the packets have varying lengths, and the channel number is denoted virtual connection identifier (VCI). The international collection of X.25 providers, using the X.25 protocol suite was colloquially known as "the Packet-switched network" in the 1980s and into the beginning of the 1990s.
- The Asynchronous Transfer Mode packet-switched protocol, where the packets have fixed length. The channel identification number consists of a virtual connection identifier (VCI) and a Virtual Path Identifier (VPI).