Do you ever find yourself looking at your watch
and thinking "This call's costing me a fortune." If you do,
you're still stuck in the 20th century with telephone technology
that's barely changed since the 19th! In the 21st century, there's no
reason why we should be paying through the nose, by the minute, to
use a telephone network when most of us now have access to a very
credible alternative: the Internet. After all, if the Internet (which
relies on large parts of the telephone network) can carry text,
images, and video clips, it should be able to carry people's voices
just as easily. That's the thinking behind VoIP (Voice Over
Internet Protocol) which, simply stated, means using the Internet
to make and receive telephone calls. How exactly does it work? What
are the advantages and the drawbacks? Let's take a closer look!
Artwork: VoIP means making telephone calls using your computer, with the sound of your voice converted to digital data that travels over the Internet using the Internet Protocol—in other words, in exactly the same way as Web pages, downloads, emails, or any other Internet data.
To lots of people, using the Net means looking at
YouTube videos or buying books from Amazon.com—but both of these
things are really about the World Wide Web, not the Internet. The
Internet is the worldwide network that links virtually
every modern computer on the planet, and it's made up of
telephone
lines, satellite links,
fiber-optic cables, and old-fashioned copper wires. The World Wide Web (all those
shopping sites, home videos, and so on that you browse from your
computer) is just one of the things that uses the Internet; email
is another. The Internet is designed so that it can send all kinds of
information, in all kinds of different ways, between the various
computers that it connects together, and without any kind of rewiring
or redesign. (Technically, this is called the end-to-end principle.) That's why, in the mid-1990s, some clever technical people
were able to figure out how to send telephone calls over the Net,
much like any other kind of information. This was the birth of VoIP.
What is VoIP?
All the information that travels over the
Net—from the latest music videos on YouTube to the confirmation
email from Amazon that your book is on its way—is sent by a method
called packet switching. Something like an email, which might
be pages and pages of characters, isn't actually sent as one big
chunk: when it leaves your computer, it's broken down into many small pieces called packets, each of
which travels independently across the Internet (theoretically by a
completely different route from other packets) before being
reassembled into a copy of the original email when it arrives at its
destination. It's a bit like sending a book through the post not as a big fat parcel
but by putting every single page into a separate envelope, individually addressed
and dispatched. It might sound odd to send things this way, but packet
switching is actually an extremely quick and efficient way of
handling the billions of emails, web pages, and everything else that has
to zip back and forth across the Net every single day. (You can read
more about how it works in our main article on how the Internet works.) All the
computers connected to the Internet understand how to send and
receive packets like this; thankfully, they all agree to work in
exactly the same way using exactly the same system, which is known as
the Internet Protocol or IP. (One of the key parts of the IP
that you may be familiar with is that every computer can be
"addressed" by quoting a unique number, known as its IP address,
which is a bit like the computer equivalent of a telephone number or
building address. Currently most computers have IP addresses made from
four pairs of two digits separated by points, such as 12.34.56.78.
Photo: Packet switching is how data travels over the Internet. It's a bit like moving house by breaking the building into individual bricks and putting each one in the mail! It sounds crazy, but it works extremely efficiently.
The Internet has only one job to do: to keep packets moving back and forth.
The computers, fiber-optic cables, and other systems that make up the Net don't know what packets they're
moving or why—and they don't care. A packet might be a piece of a photo you took on
holiday in Florida, or it could be part of an email telling someone
they're fired. As long as the data you want to send is in the form of
packets, and they're formatted in the correct way according to the
IP, you can send absolutely anything over the Internet.
Sponsored links
How does VoIP work?
How, then, do you send a telephone call over the
Internet? There are really three separate problems to solve before you can do it: alerting
someone that you want to call them, turning your voice into digital
sound and sending it over the Net (and receiving replies in the
opposite direction), and "interfacing with" (linking in to) the
ordinary telephone network, if your call is going to a traditional
landline telephone or cellphone (mobile phone). Let's look at each of
these in turn.
Call signalling
When you make a traditional telephone call to a
friend, you lift the receiver and listen for the dial tone before
punching in someone's number. What's happening here is that you're
opening up an electric circuit between your home phone and the
telephone exchange. When you dial the number, the exchange opens up a
second circuit to the receiver's phone, causing their handset to
ring. As soon as your friend lifts the receiver, there's a complete circuit
open between your two phones and you can start to talk
("send and receive voice data", if you prefer).
With VoIP, things are different. Internet
telephony is much more like cellphone telephony, with people having
unique telephone numbers that aren't permanently linked to one
physical location: the person you're calling could
be anywhere on the planet (and might not be in the same place two
days running). So the first part of making a VoIP call involves your
computer locating the receiver on the Internet, signalling their computer to receive a call,
and, once that's done, the two computers agreeing the technical
nitty-gritty of how they will actually exchange the data (just as fax
machines and modems "handshake" at the start of a call). For VoIP
to work effectively, every computer that uses it has to do these
things exactly the same way—and that's why VoIP systems use carefully agreed
international standards (known as protocols). The two protocols that cover
signalling are technically known as H.323 and SIP
(Session Initiation Protocol, sometimes also known as RFC
4168). Simply speaking, these protocols set up a communication
route between two IP addresses (the sender's and the receiver's)
across which the actual telephone call data can be sent and received.
Call transmission
To send a basic telephone call over the Internet,
you have to turn a speaker's voice into digital (numeric) form.
That's relatively easy and the technology has been around for many
years. (For example, when rock bands record CDs or
MP3s, the noises they
produce with their voices or instruments, which are analog sounds,
are converted into numbers, which are digital signals, that can be stored or manipulated by
computers. For more about the difference, see our main article on analog versus digital technology. This general process is called analog to digital
conversion. When you listen to a CD or MP3, those numbers are
converted back into sounds your ears can hear by digital to analog
conversion.) The piece of software responsible for this
process—converting audio sound into digital data and back again at
the other end—is known as a CODEC (Coder-Decoder). The CODECs
used for VoIP are designed to work optimally with sounds of
frequencies from a few hundred hertz (Hz) up to perhaps 5000Hz or so
(the frequency range of the human voice), although since 21st-century
phone calls are just as likely to contain video (chat) data as voice
sounds, VoIP systems often contain video CODECs as well as audio ones.
Once a spoken voice has been turned into numbers, it's
relatively easy to break it into packets and send it over the
Internet to another computer, where it can be reassembled and turned
back into the sound of a voice by exactly the reverse process. Again,
the computers involved in sending and receiving the data have to work
according to the same protocols (agreed methods). The data-sending
protocol used in VoIP is called RTP (real-time protocol), and
it's also the protocol that computers use for receiving streaming
media (videos you watch as you download them, in real time, over the
Internet).
Interfacing with the telephone network
So far, we've seen that sending and receiving
phone calls between two computers connected via the Internet is
relatively simple; it's broadly the same as chatting online or
exchanging emails, except that the data travelling back and forth is
digitally encoded sound and travels in real-time. Making a telephone
call from a computer to a traditional landline phone (or vice-versa)
is more complex because it involves making a link from the Internet
to the ordinary phone network (which is technically referred
to as the PSTN or Public Switched Telephone Network).
That complicates both aspects of VoIP that we discussed above. Call
signalling is more complex, because the phone you're calling might be
on either the PSTN or somewhere on the Internet—and it has to be
located first. (One solution to this is to assign a special,
nongeographical "area code" to VoIP numbers so they can be
instantly identified and routed to the Internet.) Sending and
receiving a phone call is also more complex because if you're calling
from a VoIP phone to an ordinary landline handset, there's nothing at
the receiving end to convert the digital data back into analog sound.
So the data has to be converted before it reaches its destination.
What makes phone calls like this work is an extra piece of equipment known as a gateway,
which acts as a bridge between the Internet (on one hand) and the
PSTN (on the other). You can think of a gateway as a kind of
translator that converts telephone calls in IP-format into
traditional signals that ordinary phones can understand (and vice
versa). It's also involved in call signalling, so when you dial a
landline from a VoIP phone, the gateway converts the call-signalling
data into a format that the PSTN can understand (and rings the
landline the old-fashioned way).
Photo: A gateway allows VoIP Internet phones linked to computers and routers (left) to communicate with ordinary landline phones connected to the PSTN (right). Two landlines can communicate directly over the PSTN (yellow line), just as
two VoIP phones can connect directly over the Internet (red line) in something like a Skype call. But if a landline wants to communicate with a VoIP phone (green line),
it has to go via a gateway (blue box).
Types of VoIP
Simply speaking, there are three different kinds
of VoIP. The simplest are VoIP telephone handsets that look and work much
like traditional telephones, except that instead of being wired to a
telephone line, they're either directly connected to your computer
(by something like a USB cable) or indirectly connected to it by a
wireless (Wi-Fi) router. You'll find a slightly different kind of
VoIP on cellphones (mobile phones). You make and receive mobile VoIP calls much
the same as normal cellphone calls but, instead of calls being sent and received on a
permanently open line, like a traditional cellphone call, they're
broken into packets and sent back and forth—rather like a web page
that you're browsing with something like a smart phone.
In other words, they're using packet switching over the cellphone network.
A third kind of VoIP is entirely computer based. When you call someone, the
VoIP software running on your computer (known as a client) sets up a more or
less direct connection (known as a peer-to-peer or P2P)
connection with someone else's computer, across the Internet. You send and receive text messages, voice data, or webcam chat over this
direct link. Apart from the initial logging on process, there is no intermediate computer managing the
communication between the sender and receiver, which makes this relatively secure
compared to other forms of telephone communication.
Skype worked this way until a few years ago but now uses a much more centralized
cloud-computing system.
(Zoom also uses cloud-mediated connections, rather than peer-to-peer, for reasons
it explains in this
blog post.)
How does Skype™ work?
Skype was the original peer-to-peer VoIP software—indeed, the technical nitty-gritty that
it uses is even known as the "Skype protocol." After Microsoft bought Skype in 2011, it slowly
began the process of changing it over to a more centralized, client-server model, which is how
it works today. "Original Skype" (as I'm going to refer to it from now on) was quite different.
Original Skype
Photo: Skype is the best known (although by no means the only) VoIP system. You can call any user just by entering their Skype username. Or you can use the Skype pop-up keypad to make calls to any phone, anywhere in the world, using the ordinary telephone network (PSTN).
"Original Skype" was a proprietary VoIP system using its own protocol based on peer-to-peer (P2P) networking;
essentially, it worked by creating ad-hoc, direct communication between two computers on the Internet in
a similar way to file-sharing systems such as KaZaa (developed by Niklas Zennström and Janus Friis—the same people who developed Skype). Apart from a logon server that grants access to the network, assigns unique usernames, and so on,
"Original Skype" was completely decentralized and distributed: there was no centralized "Skype control system."
At any given moment, there were something like 100 million "Original Skype" users logged on worldwide.
When you signed on to "Original Skype," your computer became one node in a global network of equal peers. Each user ran
a piece of software called a client that allowed them to send messages to other Skype users, make calls, send files,
and play real-time games. Each of the clients became an active part of the network and, whether it was actively
sending messages or not, helped the network as a whole to locate and route traffic to other users. Within the
network, some of the users with highest bandwidth and best connectivity, known as supernodes, acted as traffic hubs.
The network as a whole was made up of supernodes connected to one another
(something like 50,000 of them), with each supernode linking to many ordinary nodes.
Unlike other instant messaging programs (such as the Yahoo! and Microsoft Live Messengers and AOL's AIM),
"Original Skype" was much more adept at communicating through firewalls by random selecting the
ports
it would use. As a consequence, it was much harder for system administrators to detect and block "Original Skype" than
traffic between other Internet chat programs. "Original Skype" also used encrypted communication between peers, which also made it
highly secure—and relatively hard for random eavesdroppers or law-enforcement agencies to monitor.
Modern Skype
Microsoft's acquisition of Skype changed all this. First, around 2012, the company
restructured Skype so it worked using a network of 10,000 supernodes entirely under its own control,
apparently for security reasons, but prompting concerns about privacy.
Later, Microsoft switched from the anarchic peer-to-peer model to
a more centralized, cloud-based client-server model,
completing the transition in around 2016.
This prompted further fears of systematic privacy invasion and surveillance,
although Skype's developers argued that the change had really been made to
improve performance, particularly for mobile users.
Advantages and disadvantages of VoIP
The biggest plus point of VoIP is call cost, which
is typically either free or much less than making traditional calls
over the PSTN. (That's a huge plus point for customers, but a huge drawback for
the big telephone companies, who've been forced to regear their
businesses to meet the threat from Internet telephony.) VoIP is easy
and often immediate to set up, and generally requires no
long-term contract (although you do need to set up an account of some
kind to create a phone number or user name where people can call
you). You can usually send any kind of data over VoIP, from text
and images of your computer desktop to voice and webcam chat. Another
big plus is that VoIP liberates you from a fixed, physical location;
if you have a Skype username, for example, you can sign in with it
and receive calls from anywhere in the world.
Photo: You can make a VoIP call from anywhere you can connect to the Internet, whether or not there's a telephone network or cellphone mast nearby. That's why VoIP has proved a big hit with the military. Here, a soldier is making a VoIP call with a laptop linked to the Internet via a radar dish and satellite connection. Photo by Teddy Wade courtesy of US Army.
The biggest drawback of VoIP is call quality,
which is neither as good or as reliable as you'd get with a direct call
between two landlines. Although the sound quality itself may be
poorer (it varies considerably according to the CODECs that are used),
this is not usually much of an issue since
most people are used to the highly variable quality of cellphone
calls. Since VoIP calls travel back and forth as streams of packets,
network problems that lead to the total loss of packets cause a
degradation in call quality and a loss of communication—though
that's also a problem people are used to with cellphones and poor
signals. A much bigger issue is call latency, where delays in
sending data across the Internet (coupled with the time it takes for
the CODECs to process them) result in a significant lag between the
sender saying something and the receiver hearing it (similar to a
really bad international telephone call), which can lead to people
talking on top of one another. A related problem called jitter
can make snippets of conversation arrive in irregular bursts,
separated by silences (it happens because a certain number of
digital packets have to arrive from the Internet and be assembled
before they can be converted into audible sounds), and this can also be
very confusing to the people involved in a conversation. The "geographical freedom" of VoIP can also be a
drawback in an emergency, because if you make an emergency call from
a VoIP phone the emergency services cannot automatically figure out
where in the world the call has come from. By the same token, nuisance
calls made over VoIP may be much harder to trace or block.
The growth of VoIP
VoIP has grown enormously since it was first developed in the mid-1990s, especially with business customers.
By 2013, the most popular VoIP service, Skype, claimed to have about 300 million active users (my own experience suggests around 50–100 million people may be logged in at any time); the current overall total is hard to establish, not least because some people have multiple Skype accounts (I must have over a half dozen for various purposes). In 2016, Skype reported massive growth in mobile-app use: "We've reached one billion Skype mobile downloads," and
revealed that its hundreds of users had made "nearly two trillion minutes of video calls" since its video calling was released 10 years before. Even so, these vague figures (how many people downloaded twice on the same device or multiple times on different devices?) reveal very little about worldwide VoIP usage. That's not really surprising when you consider that this is a much more ad-hoc technology than traditional telephony, where there are actual, physical phone lines you can count.
What can we glean about overall VoIP use worldwide? Here are a few statistics I've managed to pull together over the last few years:
In July 2009, the International Telecommunication Union (ITU) reported that there were 34.6 million VoIP subscribers in Europe, with about a quarter of all fixed-line telephone subscribers using VoIP instead of the PSTN. Unfortunately, the ITU news and statistics blog tells us very little about current figures.
A similar picture has emerged in Japan, where fast cellphone networks are widely available, and around a third of all telephone calls are now made by VoIP.
In the United States, manufacturers such as 3Com report that something like a half of all new business telephones sold are VoIP models. In 2012, according to the New York Times, 42 million people were regularly using VoIP phones in the United States.
Telephone companies offering only VoIP services are still much smaller than their PSTN rivals; one of the leading providers, Vonage, had 2.4 million registered customers as of June 2011, but in February 2017 was reporting "nearly 2 million customers worldwide" (it's uncertain whether these are directly comparable figures, but they came from the same sentence of the "About" page on their website at different times).
Industry estimates suggest there were somewhere between 100 and 200 million VoIP subscribers worldwide by the end of 2015, although a mere six countries (the USA, China, Japan, South Korea, France, and Germany) represent about 80 percent of the market.
A study by Future Market Insights forecast the global VOIP market would be worth $204.8 billion by 2020, comprising around 200 million subscribers.
In the future, as broadband Internet and traditional telephony continue to converge, the strict split between the PSTN and the Internet is likely to disappear. It will be much more common to see, for example, Web pages with clickable links that make immediate VoIP calls to sales agents. Telephone calls are also likely to become more sophisticated, for example, with multi-way video calling over VoIP increasingly replacing two-way voice-only calling over the PSTN, and stereo VoIP calls replacing mono PSTN calls.
Who invented VoIP?
Here's a quick summary of some key moments in VoIP history.
1994: Two young Israeli computer scientists, Alon Cohen and Lior Haramaty, invent an audio transmitter-receiver that can send and receive voice data in real-time over the Internet.
1995: Cohen and Haramaty's company, VocalTec, develops the first Internet telephone software (also known as a VoIP client).
1996: The first versions of the VoIP call-signalling standards H.323 and SIP are published.
2002: The final version of SIP is published as international standard RFC 3261.
2003: Skype is launched.
2005: The commercial importance of VoIP becomes apparent when the eBay auction site buys Skype for $2.6 billion.
2005: Skype announces that 100 million people have downloaded its free VoIP software.
2007: Skype downloads reach 500 million.
2011: Microsoft buys Skype for $8.5 billion and its partner, Facebook, announces that it will incorporate Skype video calls into
chats between Facebook friends. Skype announces 663 million registered users.
2013: Metro, the US-based cellphone network, launches the first commercial VoLTE service (essentially, VOIP for cellphones).
2014: WhatsApp, the popular mobile messenger, announces a new version incorporating VoIP voice calling.
2015: Global VoIP subscribers are predicted to reach 200 million.
2016: Skype announces that its mobile apps have been downloaded one billion times.
2016: Microsoft finishes moving Skype to a cloud-based (Azure) service.
2020: Zoom, a leading provider of video-conferencing software,
posts extraordinary, quarterly growth rates of between 169 per cent and 367 per cent. Zoom and Skype now hold the lion's share of the video-conferencing market,
with rivals like Google Hangouts, Cisco Webex™, and GoTo Meeting trailing in their wake.
What about VoLTE?
Artwork: The VoLTE icon indicator (top, red) and settings (bottom, blue) on a Samsung Android cellphone.
Simply speaking, VoLTE is a kind of VOIP designed for cellphones, in which voice calls travel as Internet data,
packet-switched over a mobile network. Where VOIP uses your home broadband line or work Internet connection
VoLTE operates in an analogous way over a 4G or 5G mobile cellphone network. While VoLTE can offer greatly improved
call quality compared to standard mobile calls ("I'm going through a tunnel... my signal is poor..."), and faster call connections,
it's still less reliable than "desktop" VOIP, which uses a much more stable, fixed Internet connection.
If you have a reasonably new phone, you'll probably find a little icon in the status area indicating that VoLTE is
enabled—and an option in your settings to turn it on and off. For the time being, VoLTE is a handy option for mobile users, but the plan is for VoLTE/VOIP to be the default method of making mobile calls in the future.
Sponsored links
Find out more
On this website
You might like these other articles on our site covering similar topics:
Comparison of VoIP software: This useful Wikipedia article lists and compares widely available VoIP software for PCs, Macs, cellphones, and servers.
Books
Simple introductions
VoIP for Dummies by Timothy V. Kelly. John Wiley and Sons, 2011. A simple, hands-on guide to using VoIP with your computer, written in the classic, lighthearted Dummies style.
Voice Over IP Fundamentals by Jonathan Davidson. Cisco Press, 2006. A short (96-page), concise, technical overview of VoIP, including how the PSTN works, basic VoIP technology, VoIP protocols such as H.323 and SIP, and gateway controls.
VoIP and Unified Communications: Internet Telephony and the Future Voice Network by William A. Flanagan. Wiley, 2012. A useful briefing aimed mainly at a business audience. Positioning VoIP as a disruptive technology, it explains the conceptual transition from circuit to packet switching, and describes how VoIP works with such things as voice, data and fax transmission.
Switching to VoIP by Ted Wallingford. O'Reilly Media, Inc., 2005. A practical guide to how you can migrate a business telephone system to VoIP
Articles
Moving Your Number to Google Voice by J. D. Biersdorfer. The New York Times, December 28, 2017. How can you use Google Voice to integrate your various telephone numbers and services?
Skype use may make eavesdropping passe by Peter Svensson (Associated Press), USA Today, 16 February 2006. Explores the difficulty of monitoring illicit communications over Skype.
Technical
US Patent 5,825,771: Audio transceiver by Alon Cohen and Lior Haramaty, Vocaltec Ltd, October 20, 1998. The original VoIP patent, which describes how real-time audio can be sent across the Internet.
SIP: Official 2005 specification (RFC-4168) from the Internet Engineering Task Force (IETF). There's also a Wikipedia overview of SIP.
Please do NOT copy our articles onto blogs and other websites
Articles from this website are registered at the US Copyright Office. Copying or otherwise using registered works without permission, removing this or other copyright notices, and/or infringing related rights could make you liable to severe civil or criminal penalties.
