June 21, 2006
Voice over IP (VoIP)--The basics
you've ever wondered anything about VoIP at the basic level, here's an
excerpt from Chapter 2 of Internet Phone Services Simplified. The
chapter, Voice Over IP, presents the very basics of what VoIP is and
how it works.
Voice over IP (VoIP)
The term Voice over IP (VoIP) does not refer to a single service but
encompasses an entire collection of services that can fill the phone
service needs of many different residential and business customers.
VoIP can be used by a service provider to optimize its capability to
carry many calls. VoIP can be used by small and large businesses for
their office phone systems. VoIP can also be used as a good alternative
(or supplement) to the public phone system for residential phone
service, which is the focus of this book.
You might already be using VoIP and not even know or realize it.
Many telephone service providers are starting to use some form of VoIP
(transparently to you) inside their networks because of the cost
efficiencies it can afford them. Many online voice chat services, such
as Xbox Live voice chat, Skype, and so on, rely on VoIP. You might find
it worthwhile to spend a few minutes to understand how VoIP works.
Don't worry, you don't need to know how it works to use it, but it
might help to understand the advantages and limitations we discuss
Circuit Switching and Packet Switching
The main difference between traditional phone systems and VoIP
systems is circuit switching versus packet switching. The public
switched telephone network (PSTN) uses circuit switching to carry your
voice from your phone to the person you are calling (See Figure 2-1).
This means that while you are on the phone, a connection is made
end-to-end through the phone system. This requires resources (in this
case, a series of wires, switches, and connections) in the phone
network that are dedicated for the duration of your call. While you are
using them, no one else can use them. The end-to-end circuit is
reserved for your conversation.
This approach works well, but imagine the resources that are
required to carry millions of calls each day coast to coast. At first,
each call required a separate set of copper wires. Technology got
better, and now millions of calls can be carried over fiber-optic
cables (and still circuits get overloaded on Mother's Day). But even
though density improved, the basic principles of circuit switching
still apply today—each call consumes a channel on the wire end to end
for the duration of the call.
Figure 2-1. Circuit Switching versus Packet Switching
Transoceanic fiber cables can carry more than 100 million phone calls each.
Even the more ordinary fiber cables have thousands of strands but can carry
1 million+ calls.
Packet switching works differently (See Figure 2-1).
Instead of having a dedicated connection end-to-end, packet switching
breaks the voice conversation into pieces, transmits the pieces, and
then reassembles the pieces at the other side back into the voice
conversation. You might be asking yourself: How does that save
anything? Well, if you remember in circuit switching, you are consuming
a dedicated resource end-to-end. But in packet switching, many people
can share that same resource at the same time.
In the example shown in Figure 2-1, the word Hello spoken by
the caller is broken into five packets, one per letter sound, and
transmitted across the network with millions of other packets from
other phone conversations. The receiving switch or phone knows how to
reassemble these five packets into the sounds spoken by the caller, and
the word Hello is played out the handset speaker.
Note Alas, we are intentionally oversimplifying again. In
reality, it takes about 50 packets to transmit each second of speech.
But we had a hard time finding a word in the dictionary with 50 letters
that could be spoken within one second for use in our example.
The important difference to understand is that during a
traditional phone call, you are using a dedicated circuit for the
duration of your call. Transmission is constant. In packet switching,
the pieces of the conversation find their own way through the network
and are re-assembled on the other end, which allows many more
conversation to take place than in curcuit switching. So, lots of other
folks can use the same circuit at the same time you are.
How VoIP Works
Now that you understand a fundamental
difference in the way VoIP compares to the PSTN, we look in more detail
at how VoIP works. Any phone service has the following four primary
- Signaling--Refers to the communications between your
handset and the phone service, for example, how the system recognizes
you want to make a call, how it receives the number you want to call,
and so on.
- Conversation--Sometimes referred to as the
"bearer" component. This is the actual voice conversation being
transmitted and received across the network.
- Features--Phone services offer many features including call waiting, call forwarding, voice mail, and so on.
- Power--How the handset in your home receives electric power for it to operate.
Signaling refers to how a central office switch in the phone network
communicates between itself and your phone, or to other switches in the
network. You need to understand a few important signals.
First, how does the phone network know you want to place a call? As
Figure 2-2 illustrates, when you lift the handset in your house, this
signals the phone network that you want to make a call. You might not
realize it, but when you lift the handset, the first thing the central
office does is send you a dial tone sound. Then you happily dial your
numbers, which is the next step in signaling. After you have dialed,
the phone network sends and receives a flurry of digital messages
across the rest of the phone system to determine how to route your
call. The destination central office then notifies the person you are
calling by ringing his handset.
Figure 2-2. PSTN-to-PSTN Call (No VoIP)
If the person you called answers, a conversation path is set up
between you and the person you called for the duration of the call.
When one of you hangs up, the central office is signaled to disconnect
the call. Again, after some magic digital messages inside the phone
network, the call is disconnected. An important thing to understand is
that in the case of a call between two traditional phones on the PSTN,
the part inside your house works pretty much the same way (analog) that
it worked 20 or 30 years ago.
We now look at how this changes if the caller is using a
broadband phone service (that is, VoIP) and calls the same person with
a PSTN phone service, as shown (See Figure 2-3).
Figure 2-3. VoIP-to-PSTN Call
In this case, a terminal adapter now connects the handset in your house
to your broadband Internet connection. The terminal adapter acts as a
translator, converting the handset signals into VoIP signals (in other
words, it takes the analog voice and converts it to a digital signal).
For example, when you lift the handset, instead of the central office
recognizing that your phone is off hook, the terminal adapter
translates it to a message sent to the broadband phone provider that
you want to place a call.
From that point, the signaling is similar to the PSTN example
previously described, except at each step, the terminal adapter is
translating your phone handset actions into digital messages that are
being sent over your broadband Internet connection to the broadband
phone service provider's softswitch. The softswitch takes care of
routing your call, just like the central office would. Notice that in
this case, the call is still routed through the PSTN to the person you
are calling. This is done by a gateway between the broadband phone
service and the PSTN. To you, it's totally transparent.
Note Softswitch refers to a central office switch, except
instead of having a bunch of phone lines connected to it, it only
receives digital messages. The "soft" part refers to the fact that it's
a telephone switch without hard wires connected to it. While a central
office needs to be in fairly close proximity to you, a softswitch could
be thousands of miles away--anywhere the messages can reach it. Think
of a softswitch as a fast computer that understands how to route
Finally, we take a look at how this changes if both the caller
and the called party are using broadband phone services (not
necessarily even the same one), as shown in Figure 2-4.
Figure 2-4. VoIP-to-VoIP Call (No PSTN)
In this case, a terminal adapter now connects both handsets to
their broadband Internet connections. The caller goes off hook and
dials the destination number. The softswitch that serves the caller
routes the call to the softswitch that serves the called number. The
destination terminal adapter gets a digital message for the incoming
call and converts it to ring the handset.
When the called person picks up the phone, again a flurry of
digital messages are exchanged between terminal adapters and
softswitches on both ends, and voil', the call is connected.
Another important function to understand about the terminal
adapter is that it converts your voice conversation into packets that
can be sent over the Internet. The next section discusses how this is
How VoIP Carries a Conversation
Human speech is made up of
analog sound waves, which can be transmitted using straightforward
techniques. A phone on the PSTN can represent your voice as a
continuous stream of voltage changes on a copper wire (this is referred
to as a carrier signal). When the carrier signal reaches the other end
(the receiving phone), the electric signals excite a diaphragm (more
commonly known as a speaker), which produces a good approximation, or
"analogy," of your voice.
For digital telephony (including VoIP), a dedicated circuit does not
transmit the voice, so human speech must be converted to a digital
stream (or a series of 1s and 0s) by the transmitter and then
re-created on the receiving end. Analog-to-digital conversion is
accomplished by sampling, which is the process of taking many
instantaneous measurements of an analog signal.
If you were to look at human speech on a meter, it would look something like what Figure 2-5 illustrates.
Figure 2-5. Analog Waveform of Human Speech
To convert this waveform into a digital signal, the waveform is
measured thousands of times per second. For every voltage level (which
is what you are measuring), a corresponding combination of 1s and 0s
exists, and that combination is sent across the digital network. This
process of measuring and converting is called sampling. On the
receiving end, the combination of 1s and 0s is read and the
corresponding voltage is re-created.
If enough samples are taken, the original analog signal can be
nearly exactly replicated by "connecting the dots" of the instantaneous
measurements re-created on the receiving end (see Figure 2-6).
Figure 2-6. Packetizing Voice
The trick for "near exact" replication of the original signal is
to take the right amount of samples, because too few samples can result
in multiple waveforms that could possibly connect the dots (and
remember, we are trying to exactly match the waveform). Too many
samples can provide fantastic sound quality, but it can also require
too much data transmission to be cost effective. The right amount turns
out to be twice the rate of the highest frequency in the waveform. For
the sake of simplicity, consider a pure tone of 1000 Hz. Figure 2-7
illustrates what the tone would look like on a meter.
Figure 2-7. Waveform of a Pure, 1000-Hz Tone
Hertz (Hz) is also referred to as cycles per second, and in this
case, the pattern or cycle on the scope would repeat 1000 times per
second. This signal can be nearly perfectly replicated going from
analog (the tone) to digital and then back to analog by measuring the
signal 2000 times per second. This rate (the rate required for good
replication) is called the Nyquist rate, named after the clever fellow
who figured it out.
Human speech is a jumble of many different tones, ranging from
very low (bass) sounds of 100 Hz up to treble sounds of about 1700 Hz
and higher. With harmonics and other "noises," speech includes tones of
up to about 4000 Hz. Therefore, to replicate all the sounds, speech is
usually sampled at about 8000 times per second.
What did they say? OK, sampling is a bit complicated. Think of
it like this. If you wanted to paint a picture of a flower, you would
probably look at the flower, paint a little, look at the flower again,
paint a little, and so on until you completed the painting. Speech
sampling is similar. The VoIP phone "looks at" your voice about 8000
times per second while it is preparing the information to transmit to
the other phone about what sounds you are making.
After enough samples are taken, the data (in the form of bytes)
is shoved into a packet and sent on its way to the other phone (see
Figure 2-6). When the packet reaches the other phone, the sampled data
re-creates the original waveform, which excites a diaphragm, which
moves a speaker, producing your grandma's voice from the old country
telling you about this year's bumper beet crop.
With your traditional phone service, you have
no doubt become accustomed to a bunch of features such as call waiting,
call forwarding, caller ID, and so on. How do these change with a
broadband phone service?
Simple—they don't. You still get the features you are used to
as well as a few new cool ones that we talk about in the next chapter.
The features work similarly to those you have become used to with the
Power in VoIP Networks
You might be asking yourself why you
need to know where a broadband phone gets its electric power. It's an
important issue to understand.
With the public telephone system, your phone receives its power from the central office switch (See Figure 2-8).
This means that even when your home has no power, your phone still
does, as long as the central office itself has power. The same set of
wires that carry your voice conversation also send power to your
Figure 2-8. Where Does My Phone Get Its Power From?
In the case of a broadband phone service, electric powering is
different. The broadband connection between your house and your
broadband provider does not provide power, only a data communication
path. In this case, the terminal adapter itself provides the power to
your handset. Terminal adapters are plugged into the electric outlet in
your house; therefore, if your house loses electricity, so does the
terminal adapter, and more importantly, so does the handset.
Chapter 4, "Knowing Your Limits," covers some of the
limitations with broadband phone services, such as loss of electricity,
in more depth.
Many people routinely ditch their trusty old telephone handsets
for new-fangled cordless phones. Interestingly enough, these phones
require electricity in your house to power the base. This is why when
you lose electricity, your cordless phone typically does not work. Most
people haven't really thought about it, but if you are concerned, you
should always keep one traditional handset plugged into a telephone use
Putting It All Together
just looked at how broadband phone service works in comparison to
traditional telephone service. As you can see, a lot of similarities
and a few important distinctions exist. In general, how you use your
handset to place and receive calls does not change. Even the handsets
you now have in your home can still be used.
What does change is how your call is handled inside the network
and how your voice is carried. Other than that, the providers have done
a great job of integrating into the public telephone system to make it
nearly painless to consumers.
So that is what you need to know about the network and the technology
for both the traditional phone network and the VoIP network. With this
knowledge, we can now discuss the advantages and limitations of
broadband phone services.
About the Authors
Jim Doherty is the director of marketing and programs with
Symbol Technologies' industry solutions group. Before Symbol, Jim
worked at Cisco Systems, where he led various marketing campaigns for
IP telephony and routing switching solutions. Jim holds a B.S. degree
in electrical engineering from N.C. State University and an M.B.A. from
Neil Anderson is a senior manager in enterprise systems
engineering at Cisco Systems and is currently responsible for
enterprise wide-area networking, branch-office network, and teleworking
systems architectures. Neil holds a bachelor's degree in computer
To contact either author, please email: firstname.lastname@example.org and use
Internet Phone Services Simplified/post question as the subject line.
Title: Internet Phone Services Simplified
ISBN: 1-58720-162-3 Authors: Jim Doherty, Neil Anderson
Chapter 2: Voice Over IP (VoIP)
Published by Cisco Press
Reproduced from the book Internet Phone Service Simplified.
CopyrightęŁ , Cisco Systems, Inc. Reproduced by permission of
Pearson Education, Inc., 800 East 96th Street, Indianapolis, IN 46240.
Written permission from Pearson Education, Inc. is required for all
*Visit Cisco Press for a detailed description and to learn how to purchase this title.
Come back soon for other recently published book excerpts.