Need for DSP Resources

DSPs provide four major functions in a voice gateway:

• Voice termination
• Transcoding
• Conferencing
• Media termination point (MTP)

Voice termination is the process of digitizing and packetizing the audio stream on a TDM interface. A DSP is required to convert the traditional audio stream to VoIP. DSPs can handle multiple TDM calls. The quantity of DSPs required depends on the type of DSP and the complexity of the coder/decoder (codec) in use. While the DSP is performing voice termination, it provides echo cancellation, voice activity detection, and jitter management.

Transcoding is the process of matching two disparate VoIP streams. This is required when the codecs or sampling rate of two VoIP streams do not match. Figure 14-1 illustrates a common requirement for a transcoder.

Figure 14-1. Transcoder Operation

The DSPs in a Cisco voice gateway can be registered to a CallManager to provide hardware-based conferencing. The number of supported conferences and conference participants is detailed in the next section, "Determining the Resources Required."

You insert MTPs into a call to provide supplementary services if H323v2 is not supported end to end. Supplementary services include hold, conference, or other signaling instructions that are invoked during a call. MTPs might also be required with Session Initiation Protocol (SIP) gateways or endpoints. As standards evolve, supplementary services should have more interoperability, thus reducing the requirements for MTPs. Cisco voice gateways support both software-based MTPs, which do not require DSPs, and hardware-based MTPs, which do require them. A software-based MTP supports G.711 to G.711 or G.729 to G.729 calls. A hardware-based MTP supports only G.711 calls.