Wednesday, 29 April 2015

VoIP- Telephony Basics

1.1What is Trunk circuit? Do telephony switches have addresses?

Trunk circuit is a single time slot, which is used to carry voice bits (PCM) on a trunk (or span or channel) between two switches. Telephony switches are addressable by Point Codes (The SS7 equivalence of IP address).

1.2What is SS7?

SS7 is the classic out-of-band signaling protocol for circuit switched. Mostly it does basic telephony signaling (ISUP), Intelligent network signaling (TCAP) and cellular signaling (MAP).

1.3What is bearer? What’s the difference between bearer and media?

'Bearer' is the circuit way of saying 'media'. In the circuit days media used to be mainly voice (and fax, data and video to a certain extent). In Internet Telephony the term media has much wider implication. It can be anything from white board images to real time text (including of course all legacy types of media).

1.4What is PCM? TDM?

PCM stands for Pulse Code Modulation. TDM stands for Time Division Multiplexing. Both pertain to the way analog signals, such as voice are sampled, encoded as digital bits, and transmitted over the digital wire (or wireless in some cases...). PCM/TDM is the main way circuit switches use to convey the bearer to each other.

1.5What is bandwidth? What is bit rate? What are codecs? any relation between those terms? 

Explanation taken from SIP Illustrated: Bandwidth and audio or video codecs are few of those buzz words, that everyone uses, but NOT too many people seem to know what they really mean…The lengthy explanation which follows, will try to reveal this mystery for you, but you will need to be brave and patient though…In order for any Voice/Media Over IP system to function, the process of Analog to Digital (A/D) conversion needs to take place (for instance, once the SIP signaling has setup a media session, your SIP phone needs to convert your voice/image to IP packets (e.g. G729/RTP/UDP/IP (voice); H261/RTP/UDP/IP (image)), which then need to be transmitted to the other SIP phone, and converted back by the other end point to audio/video information. Voice, audio and video are all sources of analog signal waveforms (i.e., Information, which is continuous in time and value), which need to be converted to digital information (discrete in time and value), so they can be transmitted as IP packets. The process of converting an analog signal to a digital representation (A/D) is consisted of three main phases: 1. Sampling (discretize time) 2. Quantizing(discretize value) 3. Encoding (digitize value).The inverse process of A/D, which is not surprisingly called the D/A process, is consisted of decoding the digital bits into the corresponding analog values and smoothing the rectangular waveform by applying a low pass filter width width Fmax. A device that can both encode an analog signal into a digital representation and decode a digital signal into its equivalent analog signal is called a coder-decoder, or codec for short. (Remember? codecs are these things, which are advertised by the SDP element in an SIP INVITE requests, for instance, “a=rtpmap:0 PCMU/8000”).In order to understand what bandwidth is, we need to realize that based on Fourier theorem any periodic function of a wave form (signal) is made up of sine waves each contribute one frequency component to the signal. The spectrum of these frequencies (to be precise. Fmax - Fmin) is called bandwidth. In the case of “lowpass” signal, i.e., message signals such as data signal, voice, audio or video, bandwidth is approximately equal to Fmax, since Fmin is negligible.This will help us to understand the connection between bandwidth and bit rate in digital systems. Now let’s indeed examine the relation between bandwidth and bit rate in digital systems. In the binary encoding phase of the A/D process, each of the binary digits, which is generated, needs to be transmitted to its digital destination by means of digital pulse. Since we have reciprocal relationship between wave width (period duration) and its frequency, we can say that the smaller the pulse width is - the higher its frequency, and thus the larger its bandwidth (F=Fmax=bandwidth). Since a digital signal is made up of multiple signal pulses of the same duration (and thus same frequency), its bandwidth is the same as the bandwidth of a single signal pulse (reminder: Bandwidth is the range of frequencies and for practical purpose may be approximated by the max frequency element. Therefore the bandwidth of a digital signal is actually determined by the signal pulse it uses (pulse width and shape). Since one pulse is used for one bit, the pulse width is proportional to the bit duration and inversely proportional to the bit rate. Therefore the conclusion is that in digital systems the bandwidth and bit rate are proportional. The general expression, which links bandwidth and bit rate in digital system is given by B = a*R when B is the bandwidth required for relaying a digital signal with a bit rate of R bits per second and 0.5 < a < 1 (depends on the pulse design). Once we digest all of the above, we can realize that codecs would mainly differ from each other by their sampling rate and step size. (To be honest, there are other differences as well (e.g., PCM encoding versus ADPCM encoding, different companding functions, linear quantization versus non-linear quantization, compression!), but we chose not say much about them, in order not to overwhelm you...). The higher the sampling rate and the smaller the step size the more accurate information we get, but the higher the number of steps (per second) is, i.e., the larger the number of quantums (symbols) to be encoded, and thus the longer the code words, i.e., large N, and thus the higher the bit rate, which means higher (PCM) bandwidth requirements. That means there is always a trade off between quality (accuracy) and bandwidth requirements. The more quality you want – the more bandwidth resources you have to throw in (pay!).

1.6Are call setup and call establishment the same? What are they for?

Yes, these are just two synonyms of the same thing. It is needed for making sure the other end point is there and is willing to take the call. In addition it is needed for the telephony network to allocate all necessary resources for the call, make other necessary preparations (e.g. open billing records, open stat records, prepare lawful intercept hooks) and potentially run call setup time services (e.g. translation, call forwarding). In the VoIP world call setup also enables the end points exchange information about their media capabilities (e.g. codecs which can be supported) and media properties (e.g. IP address for media exchange, port number, protocol).

No comments:

Post a Comment