Telecommunication Protocols Overview: VoIP

Revolutionary Transition from TDM to IP Networks
Voice over IP (VoIP) technology
VoIP and Triple Play: Key Protocols for Multimedia Transmission in IP Networks

Voice-over-IP (VoIP), also known as IP telephony, connects TDM networks with channel switching to IP networks with packet switching. It also facilitates the gradual transition from TDM to IP networks. Introduced in the late **1990s**, VoIP is one of the earliest telecommunication technologies to enable the use of IP phones, IP PBXs, and similar equipment; the suite of VoIP protocols is **crucial** among other telecom protocols.

According to the common definition, IP telephony is real-time voice signal transmission over a packet-switched network. It converts a phone number into an IP address and the analog voice signal into a digital one.

The birth year of Internet telephony is considered to be 1995, when Vocaltec released the Internet Phone software for telephone transmission using the IP protocol. **Until** the mid-1990s, Internet phone network implementation was possible only via telephone modems, resulting in significantly lower voice quality compared to traditional phones. Nevertheless, this laid the foundation for VoIP.

Since then, the development has been so rapid that VoIP's capabilities now far exceed its formal name. Essentially, this technology allows for the transmission of not only voice but any type of information using the IP protocol, so the term shifted to a broader one, "multimedia." Corresponding data structures can include voice, images, and data in any combination, commonly referred to as Triple Play.

The VoIP network structure can be viewed as two planes. The lower one represents the transport mechanism for non-guaranteed delivery of multimedia traffic as a protocol hierarchy (RTP, UDP/IP), and the upper one is the call service management mechanism. The key protocols here are H.323 ITU-T, SIP, MGCP, and MEGACO, each offering different implementations for call **services** in IP telephony networks.

Real-time Transport Protocol (RTP) provides transport services to multimedia applications, but it doesn't guarantee delivery or packet order. RTP helps **apps** detect packet loss or order issues by assigning a number to each packet. RTP works in point-to-point or point-to-multipoint modes with no regard to the transport mechanism, but usually it is UDP.

RTP works with the Real-Time Control Protocol (RTCP), which manages data flow and checks for channel overload. RTP session participants periodically exchange RTCP packets with statistical data (number of packets sent, lost, etc.), which senders can use to adjust transmission speed and load type dynamically.

Recommendation H.323

The recommendation H.323, historically the first method for making calls in an IP network, involves the following types of data exchange:
- Digital audio
- Digital video
- Data (file/image sharing)
- Connection management (communicating function support, logical channel management)
- Session and connection establishment and termination

Key H.323 network elements include terminals, gateways, gatekeepers, and Multipoint Control Units (MCU).

A terminal provides real-time communication with another H.323 terminal, gateway, or MCU.

Gateways connect H.323 terminals with different network protocol terminals by converting the information back and forth between networks.

Gatekeepers participate in managing connections by converting phone numbers to IP addresses and **vice versa**.

Another element of the H.323 network, the proxy server, operates at the application level to identify application types and establish necessary connections.

The H.323 call service plane includes three main protocols (see picture): RAS (Registration Admission and Status) for terminal registration and resource access control, H.225 for connection management, and H.245 for logical channel management. RAS uses UDP, while H.225 and H.245 use TCP for guaranteed information delivery. UDP's delivery is not guaranteed, so if confirmation isn't received in time, UDP retransmits the message.

The process of establishing a connection involves three stages. The first one is to detect the gatekeeper, register terminals with the gatekeeper, and control **terminals'** access to network resources using the RAS protocol. The next two stages involve H.225 signaling and H.245 control message exchanges.

Recommendation H.225 outlines the procedures for connection management in H.323 networks using a set of signal messages from the ITU-T's Q.931 recommendation.

Recommendation H.245 describes the procedures for managing information channels: determining the master and slave devices, communicating terminal capabilities, and opening and closing unidirectional and bidirectional channels. It also covers delays, information processing modes, and the state of information channels by organizing loops.

The exchange of signaling messages between interacting H.323 network devices happens over H.245 logical channels. The zero logical channel, which carries control messages, must remain open for the entire duration of the connection.

SIP (Session Initiation Protocol)

The second method for handling calls in VoIP networks involves using the Session Initiation Protocol (SIP), specified in RFC 2543 by the IETF. As an application-level protocol, it is designed for organizing multimedia conferences, distributing multimedia information, and setting up phone connections. SIP is less suited for interaction with PSTN but is easier to implement. It's more suitable for ISPs offering IP telephony services as part of their package.

Key features of SIP include user mobility support, network scalability, the ability to add new functions, integration with the existing Internet protocol stack, interaction with other signaling protocols (e.g., H.323), enabling VoIP users to access intelligent network services, and independence from transport technologies.

It's worth noting that user mobility support is no longer exclusively a SIP feature. H.323 now also supports it (see ITU-T H.510, "Mobility for H.323 Multimedia Systems and Services").

A SIP network contains user agents, or SIP clients, proxy servers, and redirect servers.

User agents are terminal equipment applications that include a client (User Agent Client, UAC) and a server (User Agent Server, UAS). The UAC initiates the service request, while the UAS acts as the calling party.

The proxy server combines UAC and UAS functions. It interprets and, if necessary, rewrites request headers before sending them to other servers.

The redirect server determines the location of the called subscriber and informs the calling user.

MGCP (Media Gateway Control Protocol)

The third method for building an IP telephony network relies on the Media Gateway Control Protocol (MGCP), proposed by the IETF's MEGACO workgroup. The architecture of this protocol is probably the simplest of all three in terms of functionality. An MGCP network contains a media gateway (MG) for converting voice data between PSTN and IP telephony networks, a signaling gateway (SG) for processing signaling information, and a call agent (similar to an H.323 gatekeeper) for managing gateways.

Like H.323, MGCP is convenient for organizing PSTN-compatible IP telephony networks. However, in terms of functionality, MGCP surpasses H.323. For example, an MGCP call agent supports SS7 signaling and transparent transmission of signaling information over the IP telephony network. In contrast, H.323 networks require any signaling information to be converted by a gateway into H.225 (Q.931) messages. MGCP messages are transmitted in plain text format.

The fourth method for building an IP network, an improvement over MGCP, was developed by the IETF's MEGACO group together with ITU-T's SG 16, hence the name MEGACO/H.248. It mainly differs from its older sibling in connection organization. Thanks to this, the MEGACO/H.248 controller can change the port connection topology, allowing for flexible conference management. The MEGACO protocol supports two methods of binary encoding.

MGCP has also evolved in the field of the Internet of Things (IoT), where it is used to manage media gateways and transmit voice information in various scenarios, such as smart homes or intelligent buildings.