Media Server - Media Servers in Telephony

Media Servers in Telephony

In the world of telephony, a media server is the computing component that processes the audio and/or video streams associated with telephone calls or connections. Conference services are a particular example of how media servers can be used, as a special 'engine' is needed to mix audio streams together so that conference participants can hear all of the other participants. Conferencing servers may also need other specialized functions like "loudest talker" detection, or transcoding of audio streams, and also interpreting DTMF tones used to navigate menus. For video processing, it may be needed to change video streams, for example transcode from one video codec to another or rescale (transrate) a picture from one size to another. This media processing functions are the core responsibility of a media server.

With telephony networks moving more towards VoIP technology, and using Session Initiation Protocol (SIP), the idea of media servers has started to gain some traction. Typically, an application (the 'application server') has the controlling logic, and controls a remote media server (or multiple servers) over an IP connection, possibly using SIP. Protocols such as Netann, MSCML and MSML have been created for this way of working, and a new protocol, MediaCTRL is under development at the IETF.

The IP Multimedia Subsystem (IMS), the blueprint for next generation networks, defines a component called the MRF (Media Resource Function), which is a kind of media server. In the case of IMS, the 'controlling logic' is provided by the MRFC (MRF controller), which, along with layers above constitutes an 'application server'. Although the MRF has been associated largely with the legacy telecom H.248 protocol (see Gateway Control Protocol), it is claimed that SIP-based protocols like MediaCTRL will ultimately prevail.