An introduction to voice telecommunications for IT professionals

This content is 19 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

I’ve invested a lot of time recently into learning about two of Microsoft’s collaboration products – Live Communications Server 2005 (real-time presence information, instant messaging and VoIP) and the forthcoming version of Exchange Server (e-mail, scheduling and unified messaging). Both of these products include telephony integration, and for many IT architects, designers and administrators that represents a whole new field of technology.

One of the things I found particularly useful at the Exchange Server “12” Ignite Tour (and which wasn’t covered by the NDA), was when Microsoft UK’s Stuart Clark talked about foundation concepts for unified messaging – i.e. an introduction to voice telecommunications for IT professionals, which I’ve reproduced here for anyone who may find it useful.

There are three main types of business phone system – Centrex, key telephone systems and private branch exchange (PBX) – of these, the PBX is by far and away the most common solution employed today.

A PBX can be thought of as a self-contained telephone system for one or more offices. It allows “internal” calls between telephone extensions (without the need for an operator) as well as allowing many internal extension numbers to share common trunk lines to the telephone company’s network, representing cost savings on outbound calls and making it less likely for an inbound call to be greeted with a busy tone. Typically each office (or campus) would have it’s own PBX, and these can be connected using tie lines to allow internal, cross-organisational calls, without using public telephone networks (i.e. just using internal extension numbers).

There are various types of PBX available:

  • Analogue PBXs retain all voice and signalling traffic as an analogue signal. Even touchtones (the “beeps” when a telephone key is pressed) are sent as analogue signals. These touchtones are technically known as dual-tone multi-frequency (DTMF) signals because they actually consist of two separate tones (one high, one low) played simultaneously – something that is highly unlikely to occur with a human voice – avoiding the possibility of confusion between control signals and conversation.
  • Digital PBXs first started to appear in the 1980s and encode analogue signals in a digital format (typically ITU G.711). Digital PBXs can also support the use of analogue trunk lines.
  • IP PBXs include network interface cards and use an organisation’s data network for voice traffic in the same manner as computers use the network for data transfer. The voice traffic is digitised and packet-switching technologies used to route calls across the LAN (and WAN).

Analogue and digital PBXs employ circuit switching technology and require the use of a gateway to connect with IP-based telephony systems. Hybrid PBXs combine digital and IP PBX technologies, allowing a more gradual migration to IP-based telephony. Many modern digital PBXs can be upgraded to hybrid PBXs; however some hybrid PBXs may still require a VoIP gateway (see below) to connect with other IP-based telephone systems due to protocol incompatibilities.

PBXs have a number of associated concepts:

  • Direct inward dialling (known as direct dial-in or DDI in Europe) is a technology used to assign external numbers to internal extensions, allowing internal extensions to be called from an external line without any operator intervention.
  • A dial plan is a set of rules employed by a PBX in order to determine the action to be taken with a call. For example, the dial plan would determine how many digits an internal extension number includes as well as settings such as the number required to dial an outside line and whether or not to allow international calls to be made. Internal extensions covered by the dial plan can call other internal extensions directly and, where a company has PBXs spanned across multiple geographic sites, the dial plan can be used to allow internal calls to be made between PBXs.
  • A hunt group is a group of extensions defined as a group around which the PBX hunts to find an available extension to which to route the call. For example, a hunt group may be used for an IT support desk and the call will be routed to the first available channel (extension).
  • A pilot number is used to identify a hunt group. It is a dummy extension without an associated person or phone, which acts as a number to dial when calling into a hunt group. In the example above, the number used to contact the IT support desk is the pilot number for the IT support hunt group. Another example pilot number would be a number used to access voicemail services. More than one pilot number might target the same group of extensions.
  • A coverage path determines how to route unanswered calls (e.g. divert to voicemail if not answered within 5 rings or if the line is busy).
  • Unsupervised transfer is the process of transferring a call to another extension without using the services of an operator, also known as a blind transfer.
  • Supervised transfer is the process of controlling a call detecting error conditions, detecting answer conditions, allowing reconnection to the original call and/or combining calls.

Having examined PBX technology, there are a number of other concepts to get to grips with in the world of IT and telephony integration.

Firstly, not all telephone lines are equal. Trunk and tie lines use high-bandwidth links that can simultaneously carry multiple channels (the term port is often used rather than channel, since it is at an entrance or exit to another system and/or network):

  • T1 lines have 24 channels with a data rate of 1.544Mbps, with either 23 or 24 channels available for voice transmission, depending on the protocol being used for signaling. T1 lines are found primarily in North America and Hong Kong.
  • E1 lines have 32 channels with a data rate of 2.048Mbps, with 30 channels available for voice transmission. Of the additional two channels, one is dedicated to timing information. The other channel is used for signaling information. E1-lines are used in Europe, Latin America and elsewhere.
  • J1 lines are found in Japan, including both 1.544 and 2.048Mbps technologies.

The channels provided by T1/E1/J1 lines are virtual channels – not separate physical wires but virtually created when the voice and/or data is put onto the line using a technology known as time-division multiplexing (TDM). Occasionally, analogue or ISDN lines may be used in place of T1/E1/J1 links. Fractional T1 and E1 lines, which allow customers to lease a portion of the line may be available when less capacity than a full T1 or E1 line is needed.

It’s also important to understand the differences between circuit-switched telephone networks and packet-switched data networks:

  • Historically, circuit switching is how phone calls have worked over traditional phone lines. In its simplest form, a circuit is started on one end, when a caller picks up the phone handset to dial a number (when a dial tone is presented). The circuit is completed when the recipient answers the call at the other end and the two parties have exclusive use of the circuit whilst the phone call takes place. Once the call is completed and both phones are hung up, the call is terminated and the circuit becomes available for use.
  • Packet-switching protocols are used by most data networks. With packet switching, the voice data is subdivided into small packets. Each packet of data is given its own identifying information and carries its own destination address (cf. addressing a letter for and sending it via a postal service). Packets can be transferred by different routes to reach the same destination.

One disadvantage of circuit-switching is that peak load capacity has to be provisioned but it cannot be repurposed at off-peak times. This means that excess capacity often sits idle. Packet-switching has the advantage that it can be repurposed at off-peak hours. The exclusive use of a given circuit is one of the big differences between circuit-switched telephony networks and packet-switched data networks. Packet-switching does not preallocate channels for each conversation and instead, multiple data streams are dynamically carried over the same physical network, making packet-switching more demand responsive.

Voice over IP (VoIP) is a term that describes the transmission of voice traffic over a data network using the Internet Protocol (IP). VoIP real-time protocols help to protect against packet loss and delay (otherwise known as jitter) attempting to achieving the levels of reliability and voice-quality previously experienced with traditional circuit-switched telephone calls. IP networks support the transport of multiple simultaneous calls.

As described earlier, some PBXs will require a gateway to provide protocol conversion between two incompatible networks (e.g. a VoIP gateway between a circuit-switched telephone network and a packet-switched data network). When connected to a digital PBX, the VoIP gateway functions as a multiplexer/de-multiplexer, converting TDM signals into a format suitable for packet-switching. When connected to an analogue PBX, the VoIP gateway has to convert the analogue voice signal into a digital voice signal between converting this into data packets. The IP packets are generally transmitted between the gateway and the PBX using a protocol called real time transport protocol (RTP), which uses small packet sizes to facilitate voice streaming at playback. IP PBXs use RTP for end-to-end communications. RTP is defined in IETF RFC 3550. Secure deployments can be enabled using secure RTP (SRTP), defined in IETF RFC 3711.

Session initiation protocol (SIP) is a real-time signalling protocol used to create, manipulate, and tear down interactive communication sessions on an IP network. The VoIP industry has adopted SIP as the signaling protocol of choice. SIP can be secured using transport layer security (TLS).

Although SIP is detailed in IETF RFC 3261 this remains a proposed standard and its implementation varies between vendors. One such variations is how SIP is mapped onto other protocols, in particular, transport layer protocols. Some map it to TCP, whilst others use UDP. These differences can make communications between VoIP applications challenging, possibly requiring a VoIP gateway to facilitate communications between two VoIP systems.

Finally, real-time facsimile (ITU T.38/IETF RFC 3362) is a fax transport protocol for the Internet. T.38 defines procedures for facsimile transmission, when a portion of the path includes an IP network. It is used when relaying a fax that originated on a voice line across an IP network in real time.

With the convergence of voice and data technologies, these terms are becoming ever more important for IT professionals to understand. Hopefully this post provides a starting point for anyone who’s struggling to get to grips with the technology.

3 thoughts on “An introduction to voice telecommunications for IT professionals

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.