The Primary H.323 VoIP-Related Protocols

H.323 is an umbrella-like specification that encompasses a large number of state machines that interact in different ways depending upon the presence, absence, and topological relationship of participating entities and the type of session (for example, audio or video). There are many subprotocols within the H.323 specification. In order to understand the overall message flows within an H.323 VoIP transaction, we will concern ourselves with the most common ones that relate to VoIP. Figure 1 shows the relevant protocols and their relationships.

 

Figure 1: VoIP-Related H.323 Protocol Stack

H.323 defines a general set of call setup and negotiating procedures—the most important in VoIP applications being H.225, H.235, H.245, and members of the Q.900 signaling series. Basic data-transport methods are defined by the real-time protocols RTP and RTCP. H.323 also specifies a group of audio codecs for VoIP communications, the G.700 series:

• H.225/Q.931 Defines signaling for call setup and teardown, including source and destination IP addresses, ports, country code, and H.245 port information.
• H.225.0/RAS Specifies messages that describe signaling, Registration Admission and Status (RAS), and media stream information.
• H.245 Specifies messages that negotiate the terminal capabilities set, the master/slave relationship, and logical channel information for the media streams.
• Real Time Protocol (RTP) Describes the end-to-end transport of real-time data.
• Real Time Control Protocol (RTCP) Describes the end-to-end monitoring of data delivery and QoS by providing information such as jitter and average packet loss.
• Codecs The G.700 series of codecs used for VoIP includes:

1. G.711 One of the oldest codecs, G.711 does not use compression, so voice quality is excellent. This codec consumes the most bandwidth. This is the same codec used by PSTN and ISDN.
2. G.723.1 This codec was designed for videoconferencing/telephony over standard phone lines and is optimized for fast encode and decode. It has medium voice quality.
3. G.729 This codec is used primarily in VoIP applications because of its low bandwidth requirements.

H.323 signaling exchanges typically are routed via gatekeeper or directly between the participants as chosen by the gatekeeper. Media exchanges normally are routed directly between the participants of a call. H.323 data communications utilizes both TCP and UDP. TCP ensures reliable transport for control signals and data, because these signals must be received in proper order and cannot be lost. UDP is used for audio and video streams, which are time-sensitive but are not as sensitive to an occasional dropped packet. Consequently, the H.225 call signaling channel and the H.245 call control channel typically run over TCP, whereas audio, video, and RAS channel exchanges rely on UDP for transport. Table 1 shows H.323 VoIP ports and protocols.

Table 1: H.323 VoIP Ports and Protocols 

Protocol	Function	Port(s)	Layer 4
H.225	(Q.931) Call Setup	1720	TCP
H.225	RAS	1719	UDP
H.245	Call Capabilities Negotiation	DYNAMIC	TCP
RTP/RTCP	Media Transport	DYNAMIC	UDP

In addition, H.235 recommends an assortment of messages, procedures, structures, and algorithms for securing signaling, control, and multimedia communications under the H.323 architecture. We will now look at each of these major VoIP-related protocols in more detail. Figure 2 shows the major signaling paths in an H.323 VoIP environment, and illustrates the several paths that signaling can take. In order to simplify the messaging sequence discussion we will ignore Fast Connect and Extended Fast Connect. There are two types of gatekeeper call signaling methods: Direct Endpoint signaling, where the terminating gateways or endpoints transfer call signaling information directly between themselves; and Gatekeeper-Routed call signaling, where setup signaling information is mediated by a gatekeeper.

 

Figure 2: Typical H.323 Channels