The Nuts and Bolts

10 min read

The Nuts and Bolts

First, a history lesson. Where did VoIP come from? Many people will cite Israeli company Vocaltec as one of the pioneers of VoIP. Formed in 1994, they developed the first PC-based Internet telephony application. Think Skype, but 10 years earlier. This proved the concept of carrying voice in IP packets over the Internet but like many first-movers, Vocaltec was overtaken by those that followed. American start-ups NBX and Selsius took the idea and turned it into what we now call IP telephony. Big companies saw them doing so and said “I want a piece of that action” so 3Com bought NBX and Cisco bought Selsius. The rest, as they say, is history.

But aren’t VoIP and IP telephony one and the same thing? If you were to type the letters ‘VoIP’ into an Internet search engine there would be 150 million returns to choose from. Do the same with ‘IP Telephony’ and you will get a mere 21 million returns. Whether the purists like it or not, the term VoIP has come to mean any form of Internet Protocol-based packet voice service, but it is important to understand the differences between the various forms of VoIP.

The most widely adopted form of VoIP is Internet Telephony (see diagram below) of which by far the most popular service is Skype with an estimated 150 million software downloads and almost 8 million users online at any one time, as of November 2006 – which is probably why E-bay paid $2.4 billion to buy the company last year. But because there is no guaranteed quality of service across the Internet at this time, any form of Internet Telephony is best thought of as ‘consumer class’ rather than ‘business class’ – though it shouldn’t be entirely discarded for business communications: after all, we use cell phones for business calls, despite the occasional poor reception.

When the voice traffic originates on a PABX (as shown in the next diagram) and is then carried over a private WAN or IP VPN then the term VoIP is more properly used. Many circuit switched PABX and key system manufacturers have ‘IP enabled’ their products to work over IP networks. Their offerings would best be described as hybrid IP Telephony. Such products have their place in the market, especially if a customer already has an existing system and only wants to upgrade it to take advantage of low cost IP networking to carry their telephone calls. But some observers suggest that buying a brand new hybrid IP telephony system would possibly not be the best way forward as the user won’t be able to take full advantage of the benefits offered by a ‘pure’ IP telephony system.

So what is pure IP telephony? As the diagram below shows, IP Telephony uses IP phones, which convert speech into digital data and put it into IP packets, which are carried over the Local Area Network and out over an IP Wide Area Network or (for calls to phones not on the private network) through a gateway to the PSTN. The set-up and management of the calls is handled by a ‘call controller’ which could be a software application running on a server or a dedicated appliance. All the major PABX and networking companies now offer IP telephony systems ranging from SMB products supporting less than 20 extensions to high-end multi-site systems supporting ten’s of thousands of users.

Hosted IP telephony is similar to basic IP telephony, in as much as all the elements use IP as the transmission protocol, but instead of the call controller residing in the user’s premises, a service provider ‘hosts’ it in their network cloud. This packet telephony service is proving popular with small and medium-sized businesses that want to take advantage of the new technology but don’t want all the hassle of running their own IP telephony system.

Quality of service has been mentioned: what is it and why is it so important? Quality of Service, or QoS (pronounced ‘kwoz’) is the term used to describe the way the IP WAN and LANs deal with the packets carried across them. Without it, the packets may be delayed, may arrive in a different order than originally sent, or may not arrive at all. There are specialist terms – latency, jitter and packet loss – to describe each of these phenomena, but it is not the purpose of this article to delve into the technical depths of networking. Rather, we shall quote ABC of XYZ who says “VoIP without QoS is like the post at Christmas: you know it will probably arrive sometime, but I wouldn’t bet on when!” To be sure of delivering an acceptable quality of service, the IP network, be it a WAN, LAN or both, has to be designed to overcome these inherent problems.

Is Class of Service the same as QoS? Yes and no. Class of Service (CoS) is a way of managing traffic in a network by grouping similar types of traffic (for example, email, streaming video, voice, large document file transfer) together and treating each type as a class with its own level of service priority. Unlike QoS traffic management, CoS does not guarantee a level of service in terms of bandwidth and delivery time; they offer a ‘best-effort’. On the other hand, CoS technology is simpler to manage and more scaleable. One can think of CoS as ‘coarsely-grained’ traffic control and QoS as ‘finelygrained’ traffic control.

But what about the voice quality? In the early days of VoIP, justifiable accusations of poor voice quality were laid against the technology. The reasons for the lack of quality were manifold: not enough bandwidth to carry the voice traffic; poor compression methods and no quality of service being the main three. This meant that users were somewhat wary of adopting the new technology. But the drawbacks were overcome one by one and now VoIP voice quality can (when the network is designed properly) be as good as PSTN voice or even better. The diagram below, reproduced with permission of BT, shows how the various VoIP services compare to the PSTN, in their opinion. The most commonly used measure of voice quality is the Mean Opinion Score, which is an indication of what users would think about the quality of the call. In the early days of MOS surveys of users were used to determine the score, between 1 (bad) and 5 (excellent) but nowadays engineering formulae are used to determine the score.


What standards should be applied to the design of a VoIP network? There are literally hundreds of technical standards that could be applied to the design of a network suitable to carry data, voice (and video) with a sufficient quality to satisfy the majority of business and public sector endusers. These standards are set by many different (and often competing) standards bodies. However, there are just two organisations which are considered to be the most influential in the area of VoIP: the International Telecommunications Union, being the world’s foremost voice and video standards body; and the Internet Engineering Task Force considered by many to hold a similar position in the IP world. Of the many VoIP standards the ITU and IETF have developed, the most important are those for Call Control because these standards include call set-up, call management, call tear down and the definition of supplementary call control features. There are two main call control standards: H.323 (from the ITU) and SIP (from the IETF).Each is described below:


H.323. ITU standard H.323 (Version 5 as of July 2003) is an older and more mature VoIP standard that defines audio, video, and data communication between IP network terminals. H.323 defines terminals, gateways, gatekeepers, and MCU devices. Terminals are basically the LAN endpoints—these may be PCs or IP telephone units. Gateways support communication among IP terminals and other terminals on a switch-based network. Gateways may also pass communications to another H.323 gateway in the network. Gate-keepers provide services to terminals. Multipoint Control Units (MCUs) allow conferencing by allowing multiple terminals and gateways to participate in a conference. H.323 assumes that the network does not guarantee QoS. Because H.323 supports video, as well as audio, there is substantial overhead—a disadvantage for IP applications.

SIP. SIP (Session Initiation Protocol) is a scaleable platform defined by RFC 2543. While SIP handles many of the same tasks that H.323 does, SIP is also intended to offer more features. Like H.323, SIP can support Internet-based calls and conferences through the use of clients, gateways, and other communication devices. SIP also handles conference calls through the use of MCUs. SIP supports protocols such as RSVP (Resource Reservation Protocol), RTP (Real-Time Transport Protocol), and SDP (Session Description Protocol), among others. Unlike H.323, SIP has very little overhead and uses many of the mechanisms found in HTTP.


Of the two, SIP has more or less become the standard of choice in the industry, as it offers more major features than H.323, such as Instant Messaging and Presence, two components of Unified Communications, the catch-all term for the coming together of voice, video, data on both wired and wireless networks, together the desktop and server business applications and devices that we all use. But even though SIP is a standard, different manufacturers may develop their own ‘extensions’ to enable differentiating features unique to their offerings. To paraphrase Mr. Spock: “it’s SIP, Jim, but not as we know it”.

There are other standards which affect VoIP but that are more specific to the underlying IP network design. These include those which pertain to the aforementioned Quality of Service, and others such as Power over Ethernet. These and many more will be covered in greater detail in next month’s article on LAN Infrastructure. The SIP standard merits its very own primer which will be appearing later this year.

This technical stuff is all very well, but the question I want answered is this: how do we make money out of VoIP? In 2003, Ron Elliott, managing director of Nte, a reseller of telecommunications equipment based in Peterlee, County Durham, wrote a dissertation entitled ‘Value Innovation and the Value-Added Reseller’ for his MBA. In it he wrote:

‘To achieve sustained profitable growth, companies must break out of the competitive and imitative trap and must cultivate value innovation. The emphasis on value places the buyer and not the competition at the centre of strategic thinking. Value-added resellers in the telecoms domain must think in terms of a total customer solution rather than supplying hardware and software, even if this pushes beyond the industry’s traditional offerings.’ He used the diagram below to illustrate his point; it was one that the author of this primer had used at a Comms Business Convergence Summit in Coventry in late 2002.

The £ Value in blue on the left represents the actual revenue associated with each element of a VoIP sale. The greatest amount of money comes from the sale of the IP infrastructure products such as routers and LAN switches. Smaller proportions of the deal value are associated with the call control software, the servers and IP handsets that are needed in all IP telephony systems. If only one IPT application such as voice mail, is added to the package, again not much extra money will be made. Of course, if a larger number of applications such as contact centres, IVR, etc.) can be attached to the sale, then the pale green slivers show that the £ Value starts to increase. This should be the goal of all resellers: to add as many applications as possible to the deal. Not only will the deal value increase, but the profits will rocket because, as illustrated by the % margin block on the right, the margins to be made from software applications and the professional services wrap are much, much higher than from the tin alone.

© CQC Consult.


This paradigm is also applicable in the new age of Unified Communications that is dawning. If the reseller is able to supply not only the tin and the IP telephony applications, but also the business applications and the extras that go to make up a complete Communications Enabled Business Solution (CEBS) then not only does the overall deal size shoot up; not only is the total deal margin much higher, but the customer would find it hard to move their business elsewhere.

So there you have it. VoIP is the way forward and has been since the early days of this century. Customers are now buying more IP telephony systems than TDM PABXs, according to all the analysts. Companies like Microsoft are moving into the market with Unified Communications, which will offer even more opportunity for the channel to make money. What we would recommend is that the channel either ‘grows its own’ IP and software skills, both in sales and support; or if that is a step too far, partner with resellers that have those skills. Whatever you do, don’t ignore VoIP or UC any longer: it isn’t going to go away.



Some commonly used VoIP terms.

IP: Internet Protocol. IP is the lingua franca of data communications; the network layer protocol in the TCP/IP communications protocol suite. IP contains a network address and allows messages to be routed to a different network but it does not ensure delivery of a complete message. TCP provides that guarantee.

IP PABX . An IP PABX, also known as an IP Telephony system, provides similar functionality to legacy PABXs but employ a different switching technology.

IM: Instant Messaging. This is a relatively new communications medium, particularly for the corporate environment. Its origins are that of the Internet.

LAN: Local Area Network. A communications networks that links users within a confined geographic network such as a building or a campus. Please also see WAN.

Latency: this is the average time it takes for a packet to travel from the source to the destination.

PABX: Private Automatic Branch Exchange. An in-house telephony system that connects, disconnects, and transfers calls.

PSTN: Public Switched Telephony Network. The global communications infrastructure that handles telephone calls.

QoS: Quality of Service. QoS is the key VoIP issue. This is a broad term used to describe mechanisms that (a) detect that the data packets are those of a real-time medium (voice or video) and (b) that route the packets according to their priority.

Softphone: PC or PDA that runs IP telephony software. Can be used as a phone or in combination with an IP phone, e.g. computing resources and graphical interfaces are used to facilitate advanced telephony applications such as conferencing.

SIP: Session Initiation Protocol. This lightweight, flexible protocol is used to set up communication sessions (communication links) and it is widely employed in converged voice-data applications.

UC: Unified Communications. The addition of real-time communications (telephony) functionality to unified messaging, e.g. the ability to listen to emails from a cellular phone.

UM: Unified Messaging. An inbox such as Microsoft Exchange or Lotus Notes that contains all messaging media types: email, facsimile and voice mail. May also include SMS.

VoIP: Voice over IP. This indicates that the voice signal has been digitised and converted into the packet format used by IP.

VPN: Virtual Private Network. A private network created within a public network to enable secure communications, often with Quality of Service, by using Tunneling.

WAN Wide Area Network. A communications network connecting multiple sites separated by distances greater than would be served by Local Area Network technologies.