VoIP: Technology and Terminology
There are 2 main type of VoIP telephones; Soft phones and hard phones, although hard phones are often just called “phones” because that’s what they look like.
A soft phone is a piece of software that runs on a PC which then uses the audio capabilities of the PC to take audio data, convert it into digital data then send it out the network port, while doing the reverse at the same time. The audio hardware can be a simple speaker and microphone, or a headset connected to the on-board sound device, or it may be a USB connected speaker and microphone unit.
Note that when we say PC here, we really mean almost any form of computing platform – a Desktop system running Windows, or an Apple Mac, or a Laptop running Linux. Even some handheld devices have VoIP soft phone software avalable to them.
A hard phone, is a physical device that looks just like a normal telephone, but instead of a wire to connect into a phone socket, it would connect into an Ethernet switch. These devices sometimes need an external power supply (no different to a conventional DECT cordless phone), and can sometimes take their power over the Ethernet line, but this then requires a special Ethernet switch, or power injector device back in the central switch location. The Power over Ethernet solution, while more expensive is ideal for offices where you need to minimise cabling under each desk. Some phones have a built-in Ethernet switch, so you’d typically just un-plug the Ethernet wire going to your PC, and plug it into the back of the phone, then plug a new lead from the phone to the PC. This can be a boon in office environments where cabling may be an issue.
Other types of hard IP phone can connect via Wi-Fi, and look just like the current generation of mobile phone handset, and there are “bar” and “clam-shell” designs.
A recent development is in the mobile phone arena where some mobile phones have a soft phone client built into then and Wi-Fi connectivity, enabling the use of VoIP to make calls rather than the traditional GSM network connectivity!
Examples of VoIP phones. Desk phone on the left, and a Wi-Fi phone on the right
The Soft Phones that run on your PC/Laptop/Mac, etc. are designed to look like physical counterparts phones. They have a keypad, on and off-hook buttons, volume controls, and so on. Some of the USB handsets have their own keypads and with the right drivers, the keypads can connect directly into the soft phone application and make the system work just as you would expect.
A soft phone application and a USB “phone”
Finally, it’s worthwhile mentioning a device called an ATA – Analogue Telephone Adapter. These devices allow you to convert your existing analogue phone into an IP phone – they have a normal phone socket on one side, and the Ethernet (and power) connection on the other. These may be used to connect in an existing DECT phone, expensive polycom conference phone or a specialist phone used by someone with disabilities. (big buttons, special ear pieces, and so on), or maybe someone just has a nice phone and they don’t want to get rid of it! Some even have a way to connect to an existing PSTN telephone system, so you can use an existing analogue phone with an ATA and make calls either over your PSTN connection, or over the Internet. Ideal for working from home when you want a real phone rather than a soft-phone, but don’t have the desk space to cope with another phone in addition to your normal house phone.
Terminology
The world of telephony has more acronyms, buzzwords and technobabble than almost anything else, and VoIP is no exception. See the glossary for more buzzwords and acronyms, but we’ll mention the important ones here:
Starting with the way the phones connect to each other and the various VoIP exchanges, we have SIP, H.323 and IAX. SIP and H.323 are the common ones. SIP stands for Session Initiation Protocol. H.323 is the number of a published standard which does the same thing as SIP (and more) The important thing to know is that SIP and H.323 do not actually transmit the voice data – they merely let the 2 end-points know how to do the data transmission. They control things like sending a number, putting a party on hold, transferring a call, and terminating a call. H.323 is the more complex and full-specification of the two, SIP is newer and easier to implement but may lack some of the more esoteric features of H.323. It’s a bit like Betamax vs. VHS. Betamax is superior to VHS, but VHS was easier, and cheaper to build at the time and so became the popular standard. SIP is the VHS of the VoIP world, and there are undoubtedly many more SIP implementations than H.323, and almost all VoIP phones (both soft and hard) support SIP, while hardly any support H.323.
IAX is a new protocol and it was designed to allow the Asterisk PABX system to talk to other Asterisk systems. It’s very compact and network efficient and some VoIP phone manufacturers are starting to implement it in their systems. It has a huge advantage over SIP and H.323 in that is is designed to work with NAT firewalls. SIP and H.323 have difficulties passing their data through a NAT firewall.
Asterisk is an open source PABX implementation that is designed to run under the Linux operating system on ordinary PCs. It is a full featured system capable of handling anything from one to thousands of calls with every feature you can possibly think of, and when it doesn’t have a feature you think of, it can usually be programmed in relatively easily.
The actual voice data
As mentioned above, SIP and H.323 have nothing to do with the actual speech data, just the setting up of the calls. The data is carried independently of the signalling and there are a multitude of formats to encode and decode the speech into digital data.
The most common method is the same method that the global telecommunications companies use themselves – this is called G.711. Unfortunately there are 2 variants of G.711. Sometimes known as G.711a and G.711u, or often referred to as uLaw or aLaw. North America uses uLaw, the rest of the world uses aLaw, although converting between the 2 is very fast and efficient with no loss of sound quality. These are what you might call the “baseline” CODECs (another buzzword – CODEC is COmpression DECompression) and are not really compressed data at all. Data is sampled 8000 times a second into 8-bit samples which gives us a data rate of 64Kbps. When we put this on the Internet, we need to add a little bit more data which takes us up to approximately 80Kbps. If your outgoing ADSL line is 256Kbps, then you can only place 3 calls using the G.711 codec over your ADSL line before it is full.
We can get round this by compressing the data. Just as MP3 compresses music data, there are ways to compress speech data. GSM is a very common one, it’s data rate is just 13Kbps, and most people can’t really tell the difference between a mobile phone call and a landline, so it works reasonably well. (It is actually possible to use GSM over a standard 56Kbps dial-up modem connection)
There are many other CODECs, some even provide better quality than G711, but that comes at a price of requiring much more bandwidth. Some use even less than GSM and sound better, but these have patents attached requiring licenses for each one you use. (E.G. the G.729A CODEC gives you excellent speech quality, uses less bandwidth than GSM, but it’ll cost you $10 for each license you require)
The telephony world is riddled with patents, with everyone trying to lock you into their own systems, and prevent anyone else from using their systems, or charging them money to do so.
How do we make VoIP calls?
To call someone using VoIP, you need 2 compatible VoIP phones, and some network connectivity between them. At the simplest, you can just key in the IP address of the remote phone and it should ring. This is hardly ideal if you are in an office and making frequent calls – an IP address is 12 digits long, so you need some sort of exchange to manage the calls, and give you additional features like voicemail, diverts on busy or no-reply, call transfer, 3 (or more) way calling, and so on, it might even have a gateway to the PSTN, so from your VoIP phone on your desk, you can, for example, dial 9 for an outside line and then make a call to anyone on the planet who has a “real” phone.
The exchange or PBX/PABX (Private [Automatic] Branch Exchange)
Because we are dealing with the Internet, we could have the VoIP exchange located somewhere out there on the Internet. There is no real need for it to be physically in your office, although there are several distinct advantages to this.
Using a virtual exchange means that you are utilising the services of a company to run the exchange for you. The traditional telephony equivalent is the Centrex exchange, where the telco (or telephone company; E.G. BT) would run what would appear to be an in-office PABX for you. You can place calls from extension to extension and the virtual PABX arranges the setup of the calls, manages voicemail (and voicemail to email or even to a web interface) and all the usual features that you’d expect from an in-office PABX.
In the VoIP world, the problem area is then connecting the virtual PABX back to the PSTN. It’s easy for BT to do as they are the phone company, but for independent VoIP operators it’s harder, but not impossible. This is where moving the exchange into an office can have it’s advantages; The exchange can connect on one side to your Local Area Network, and on the other side it can connect to your existing analogue or digital (ISDN) lines and bridge the two. You now get the same functionality as a traditional PABX, but your data to/from each phone and to/from phone to the PSTN goes via your in-office LAN.
The advantage of this is that you can plug your hard phone in almost anywhere. With the right setup, your phone does not need to be connected to your office LAN, it can be anywhere in the world where you have a broadband Internet connection. If you are running a soft-phone application on your laptop then you could have your office phone right with you. Working from home or sitting in an airport and being able to call any office extension, or even make an outgoing call on the office PSTN lines is now a reality.
