VoIP

The emergence of VoIP brings with it a wide range of possibilities. By virtue of transporting voice over a data stream, VoIP frees the voice stream from the confines of a voice-specific network and its associated platforms. VoIP can be received and transmitted via PCs, laptops, personal digital assistants (PDAs), and IP handsets. Where IP exists, there can be VoIP.

Origins of VoIP

In November 1988, Republic Telcom (yes, one e) of Boulder, Colorado, received patent number 4,782,485 for a "Multiplexed Digital Packet Telephone System." The plaque from the Patent and Trademark Office describes it as follows: "A method for communicating speech signals from a first location to a second location over a digital communication medium comprising the steps of: providing a speech signal of predetermined bandwidth in analog signal format at said first location; periodically sampling said speech signal at a predetermined sampling rate to provide a succession of analog signal samples; representing said analog signal samples in a digital format thereby providing a succession of binary digital samples; dividing said succession of binary digital samples into groups of binary digital samples arranged in a temporal sequence; transforming at least two of said groups of binary digital samples into corresponding frames of digital compression."

Republic and its acquiring company, Netrix Corporation, applied this voice over data technology to the data technologies of the time (X.25 and frame relay) until 1998 when Netrix and other competitors introduced VoIP onto their existing voice over data gateways. Although attempts had been made at Internet telephony from a software-only perspective, commercial applications were limited to using voice over data gateways that could interface the PSTN with data networks. Voice over data applications were popular in enterprise networks with offices spread across the globe (eliminating international interoffice long-distance bills), offices where no PSTN existed (installations for mining and oil companies), and for long-distance bypass (legitimate and illegitimate).

The popularity and applications of VoIP continued to grow. VoIP accounted for 6 percent of all international long-distance traffic in 2001.^[2] Six percent may not seem like an exciting sum, but given a mere three years from the introduction of a technology to capturing 6 percent of a trillion-dollar, 100-year-old industry, it is clear that VoIP will continue to capture more market share.

How Does VoIP Work?

The first process in an IP voice system is the digitization of the speaker's voice. The next step (and the first step when the user is on a handset connected to a gateway using a digital PSTN connection) is typically the suppression of unwanted signals and compression of the voice signal. This has two stages. First, the system examines the recently digitized information to determine if it contains a voice signal or only ambient noise, and it discards any packets that do not contain speech. Secondly, complex algorithms are employed to reduce the amount of information that must be sent to the other party. Sophisticated coders /decoders (codecs) enable noise suppression and the compression of voice streams. Compression algorithms (codecs) include G.723, G.728, and G.729. G.711 is the codec for uncompressed voice at 64 Kbps.

Following compression, voice must be packetized and VoIP signaling protocols added. Some storage of data occurs during the process of collecting voice data, since the transmitter must wait for a certain amount of voice data to be collected before it is combined to form a packet and be transmitted via the network. Protocols are added to the packet to facilitate its transmission across the network. For example, each packet will need to contain the address of its destination, a sequencing number in case the packets do not arrive in the proper order, and additional data for error checking. Because IP is a protocol designed to interconnect networks of varying kinds, substantially more processing is required than in smaller networks. The network addressing system can often be very complex, requiring a process of encapsulating one packet inside another and, as data moves along, repackaging, readdressing, and reassembling the data.

When each packet arrives at the destination computer, its sequencing is checked to place the packets in the proper order. A decompression algorithm is used to restore the data to its original form, and clock-synchronization and delay-handling techniques are used to ensure proper spacing. Because data packets are transported via the network by a variety of routes, they do not arrive at their destination in order. To correct this, incoming packets are stored for a time in a jitter buffer to wait for late-arriving packets. The length of time in which data are held in the jitter buffer varies depending on the characteristics of the network.

VoIP Signaling Protocols

VoIP signaling protocols, such as H.323 and the Session Initiation Protocol (SIP), set up the route for the media stream or conversation over an IP network. Gateway control protocols such as the Media Gateway Control Protocol (MGCP) and MEGACO (also signaling protocols) establish control and status in media and signaling gateways.

Routing (User Datagram Protocol [UDP] and Transmission Control Protocol [TCP]) and transporting (Real-Time Transport Protocol [RTP]) the media stream (conversation) once the route of the media stream has been established are the function of routing and transport protocols. Routing protocols such as UDP and TCP could be compared to the switching function described in Chapters 2 and 3.

RTP would be analogous to the transport function in the PSTN. The signaling functions establish which route the media stream will take over the network delivering the bits, that is, the conversation. This is illustrated in Figure 6-3.

Figure 6-3: Signaling and transport protocols used in VoIP

The process of setting up a VoIP call is roughly similar to that of a circuit-switched call made on the PSTN. A media gateway or IP phone must be loaded with the parameters to allow proper media encoding and the use of telephony features. Inside the media gateway is an intelligent entity known as an endpoint. When the calling and called parties agree on how to communicate and the signaling criteria is established, the media stream over which the packetized voice conversation will flow is established. Signaling establishes the virtual circuit over the network for that media stream. Signaling is independent of the media flow. It determines the type of media to be used in a call. Signaling is concurrent throughout the call. Two types of signaling are currently popular in VoIP: H.323 and SIP.^[3]

Figure 6-3 details the relationship between signaling and media flow. This relationship between transport and signaling is very similar to the PSTN in that Signaling System 7 (SS7) is out-of-channel signaling, such as that used in VoIP.

H.323 H.323 is the International Telecommunications Union-Telecommunications Standardization Sector (ITU-T) recommendation for packet-based multimedia communication. H.323 was developed before the emergence of VoIP for video over a LAN. As it was not specifically designed for VoIP, it has faced a good deal of competition from a another protocol, SIP, which was designed specifically for VoIP over any size network. H.323 has enjoyed a first-mover advantage and a considerable installed base of H.323 VoIP networks now exists.

H.323 is comprised of a number of subprotocols. It uses protocol H.225.0 for registration, admission, status, call signaling, and control. It also uses protocol H.245 for media description and control, terminal capability exchange, and general control of the logical channel carrying the media stream(s). Other protocols make up the complete H.323 specification, which presents a protocol stack for H.323 signaling and media transport. H.323 also defines a set of call control, channel setup, and codec specifications for transmitting real-time video and voice over networks that don't offer guaranteed service or QoS. As a transport, H.323 uses RTP, an Internet Engineering Task Force (IETF) standard designed to handle the requirements of streaming real-time audio and video via the Internet.^[4] H.323 was the first VoIP protocol for interoperability among the early VoIP gateway/gatekeeper vendors. Unfortunately, the promise of interoperability between diverse vendors platforms did not materialize with the adoption of H.323. Given the gravity of this protocol, it will be covered in a separate chapter.

The H.323 standard is a cornerstone technology for the transmission of real-time audio, video, and data communications over packet-based networks. It specifies the components, protocols, and procedures providing multimedia communication over packet-based networks. Packet-based networks include IP-based (including the Internet) or Internet packet exchange (IPX)-based LANs, enterprise networks (ENs), metropolitan area networks (MANs), and WANs. H.323 can be applied in a variety of mechanisms: audio only (IP telephony); audio and video (videotelephony); audio and data; and audio, video, and data. H.323 can also be applied to multipoint-multimedia communications. H.323 provides myriad services and therefore can be applied in a wide variety of areas: consumer, business, and entertainment applications.

SIP: Alternative Softswitch Architecture? If the worldwide PSTN could be replaced overnight, the best candidate architecture, at the time of this writing, would be based on VoIP and SIP. Much of the VoIP industry has been based on offering solutions that leverage existing circuit-switched infrastructure (such as VoIP gateways that interface a PBX and an IP network). At best, these solutions offer a compromise between circuit- and packet-switching architectures with the resulting liabilities of limited features, expensive-to-maintain circuit-switched gear, and questionable QoS and reliability as a call is routed between networks based on those technologies. SIP is an architecture that potentially offers more features than a circuit-switched network.

SIP is a signaling protocol. It uses a text-based syntax similar to the Hypertext Transfer Protocol (HTTP) used in web addresses. Programs that are designed for parsing HTTP can be adapted easily for use with SIP. SIP addresses, known as SIP uniform resource locators (URLs), take the form of web addresses. A web address can be the equivalent of a telephone number in an SIP network. In addition, PSTN phone numbers can be incorporated into an SIP address for interfacing with the PSTN. An e-mail address is portable. Using the proxy concept, one can check his or her e-mail from any Internet-connected terminal in the world. Telephone numbers, simply put, are not portable. They only ring at one physical location. SIP offers a mobility function that can follow a subscriber to whatever phone he or she is nearest to at a given time.

Like H.323, SIP handles the setup, modification, and teardown of multimedia sessions, including voice. Although it works with most transport protocols, its optimal transport protocol is RTP. Figure 6-4 shows how SIP functions as a signaling protocol while RTP acts as the transport protocol for a voice conversation. SIP was designed as a part of the IETF multimedia data and control architecture. It is designed to interwork with other IETF protocols such as the Session Description Protocol (SDP), RTP, and the Session Announcement Protocol (SAP). SIP is described in the IETF's Request for Comments (RFC) 2543. Many in the VoIP and softswitch industry believe that SIP will replace H.323 as the standard signaling protocol for VoIP.

Figure 6-4: Proxy server SIP architecture

SIP is part of the IETF standards process and is modeled upon other Internet protocols such as the Simple Mail Transfer Protocol (SMTP) and HTTP. It is used to establish, change, and tear down (end) calls between one or more users in an IP-based network. In order to provide telephony services, a number of different standards and protocols must come together, specifically to ensure transport (RTP), enable signaling with the PSTN, guarantee voice quality (Resource Reservation Setup Protocol [RSVP]), provide directories (Lightweight Directory Access Protocol [LDAP]), authenticate users (Remote Access Dial-In User Service [RADIUS]), and scale to meet anticipated growth curves.

How Does SIP Work? SIP is focused on two classes of network entities: clients, also called user agents (UAs), and servers. VoIP calls on SIP to originate at a client and terminate at a server. Types of clients in the technology currently available for SIP telephony would include a personal computer loaded with a telephony agent or an SIP telephone. Clients can also reside on the same platform as a server. For example, a PC on a corporate WAN might be the server for the SIP telephony application, but it may also be used as a user's telephone (client).

SIP Architecture SIP is a client-server architecture. The client in this architecture is the UA), which interacts with the user. It usually has an interface towards the user in the form of a PC or an IP phone (an SIP phone in this case). Four types of SIP servers exist. The type of SIP server used determines the architecture of the network. The servers are UA servers, redirect servers, proxy servers, and a registrar.

^[2]"TeleGeography 2002-Global Traffic Statistics and Commentary," TeleGeography, www.TeleGeographycom 2001.

^[3]Bill Douskalis, IP Telephony-The Integration of Robust VoIP Services (New Jersey: Prentice Hall, 2000).

^[4]Bill Douskalis, IP Telephony-The Integration of Robust VoIP Services (New Jersey: Prentice Hall, 2000), 9.