Glossary¹¹

Artificial Neural Networks (ANN) -

Also referred to as connectionist architectures, parallel distributed processing, and neuromorphic systems, an artificial neural network (ANN) is an information-processing paradigm inspired by the way the densely interconnected, parallel structure of the mammalian brain processes information. Artificial neural networks are collections of mathematical models that emulate some of the observed properties of biological nervous systems and draw on the analogies of adaptive biological learning. The key element of the ANN paradigm is the novel structure of the information processing system. It is composed of a large number of highly interconnected processing elements that are analogous to neurons and are tied together with weighted connections that are analogous to synapses.

Asynchronous Transfer Mode (ATM) -

A network technology for both local and wide area networks (LANs and WANs) that supports realtime voice and video as well as data. The topology uses switches that establish a logical circuit from end to end, which guarantees quality of service (QoS). However, unlike telephone switches that dedicate circuits end to end, unused bandwidth in ATM's logical circuits can be appropriated when needed. For example, idle bandwidth in a videoconference circuit can be used to transfer data. ATM is widely used as a backbone technology in carrier networks and large enterprises, but never became popular as a local network (LAN) topology. ATM is highly scalable and supports transmission speeds of 1.5, 25, 100, 155, 622, 2488 and 9953 Mbps. ATM is also running as slow as 9.6 Kbps between ships at sea. An ATM switch can be added into the middle of a switch fabric to enhance total capacity, and the new switch is automatically updated using ATM's PNNI routing protocol.

Audio Codec -

A hardware circuit (chip) or software routine that converts sound into digital code and vice versa. The first step is to convert the analog sound into digital samples, using PCM or ADPCM. The next step is to use perceptual audio coding to further compress the amount of digital data. If the codec is specialized for human voice, it is also known as a "speech codec," "voice codec" or "vocoder."

Audio Compression -

Encoding digital audio data to take up less storage space and transmission bandwidth. Audio compression typically uses lossy methods, which eliminate bits that are not restored at the other end. ADPCM and MP3 are examples of audio compression methods.

Automatic Repeat reQuest (ARQ) -

A method of handling communications errors in which the receiving station requests retransmission if an error occurs.

Bit Rate (BR) -

The transmission speed of binary coded data.

COder-DECoder (Codec) -

Hardware or software that converts analog sound, speech or video to digital code and vice versa (analog to digital- digital to analog). Codecs must faithfully reproduce the original signal, but they must also compress the binary code to the smallest number of bits possible in order to transmit faster. As network bandwidth increases, so does the demand for more audio and video, so compression is always an issue. Codecs can be software or hardware. Software codecs are installed into audio and video editing programs as well as media players that download audio and video over the Web. Software codecs rely entirely on the PC for processing. Hardware codecs are specialized chips built into digital telephones and videoconferencing stations to maximize performance. Although hardware codecs are faster than software routines, faster desktop machines are increasingly enabling software codecs to perform quite adequately.

Forward Error Correction (FEC) -

A communications technique that can correct bad data on the receiving end. Before transmission, the data are processed through an algorithm that adds extra bits for error correction. If the transmitted message is received in error, the correction bits are used to repair it.

IP Telephony -

The two-way transmission of audio over a packet-switched IP network (TCP/IP network). When used in a private intranet or WAN, it is generally known as "voice over IP," or "VoIP." When the transport is the public Internet or the Internet backbone from a major carrier, it is generally called "IP telephony" or "Internet telephony." However, the terms IP telephony, Internet telephony and VoIP are used interchangeably. IP telephony uses two protocols: one for transport and another for signaling. Transport is provided by UDP over IP for voice packets and either UDP or TCP over IP for signals. Signaling commands to establish and terminate the call as well as provide all special features such as call forwarding, call waiting and conference calling are defined in a signaling protocol such as H.323, SIP, MGCP or MEGACO

International Telecommunication Union (ITU) -

Formerly the CCITT (Consultative Committee for International Telephony and Telegraphy), it is an international organization founded in 1865, now part of the United Nations System, that sets communications standards for global telecom networks. The ITU is comprised of more than 185 member countries. The Union began the 21st century streamlined into three sectors: Telecommunication Standardization (ITU-T), Radiocommunication (ITU-R) and Telecommunication Development (ITU-D). The oldest of these is the ITU-T, which produces more than 200 standards recommendations each year in the converging areas of telecommunications, information technology, consumer electronics, broadcasting and multimedia communications.

Internet Protocol (IP) -

The protocol that is used to route Internet traffic. IP is an unreliable protocol; higher layer protocols such as TCP insure that IP successfully delivers all data to the intended recipient. The Internet protocol defines how information gets passed between systems across the Internet.

Jitter -

A flicker or fluctuation in a transmission signal or display image. The term is used in several ways, but it always refers to some offset of time and space from the norm. For example, in a network transmission, jitter would be a bit arriving either ahead or behind a standard clock cycle or, more generally, the variable arrival of packets. In computer graphics, to "jitter a pixel" means to place it off side of its normal placement by some random amount in order to achieve a more natural antialiasing effect.

Mean Opinion Score (MOS) -

The quality of a digitized voice, audio, video or multimedia signal. It is a subjective measurement that is derived entirely by people scoring the results from 1 to 5, with a 5 meaning that the quality is perfect (other quality scales are also available). The MOS is an average of the numbers for a particular codec. There are several recommendations provided by the ITU concerning MOS and the diffrent ways to carry out subjective quality tests.

Multimedia -

Information in more than one form. It includes the use of text, audio, graphics, animated graphics and full-motion video.

Neural Network (NN) -

A modeling technique based on the observed behavior of biological neurons and used to mimic the performance of a system. It consists of a set of elements that start out connected in a random pattern, and, based upon operational feedback, are molded into the pattern required to generate the required results. It is used in applications such as robotics, diagnosing, forecasting, image processing and pattern recognition.

Packet Loss -

The discarding of data packets in a network when a device (switch, router, etc.) is overloaded and cannot accept any incoming data at a given moment. High-level transport protocols such as TCP/IP ensure that all the data sent in a transmission is received properly at the other end.

Packet Switching -

A networking technology that breaks up a message into smaller packets for transmission and switches them to their required destination. Unlike circuit switching, which requires a constant point-to-point circuit to be established, each packet in a packet switched network contains a destination address. Thus all packets in a single message do not have to travel the same path. They can be dynamically routed over the network as lines become available or unavailable. The destination computer reassembles the packets back into their proper sequence. Packet switching efficiently handles messages of different lengths and priorities. By accounting for packets sent, a public network can charge customers for only the data they transmit. Packet switching has been widely used for data, but not for realtime voice and video. However, this is beginning to change. IP and ATM technologies are expected to enable packet switching to be used for everything.

Quality of Service (QoS) -

refers to the capability of a network to provide better service to selected network traffic over various technologies, including Frame Relay, Asynchronous Transfer Mode (ATM), Ethernet and 802.1 networks, SONET, and IP-routed networks that may use any or all of these underlying technologies. The primary goal of QoS is to provide priority including dedicated bandwidth, controlled jitter and latency (required by some real-time and interactive traffic), and improved loss characteristics. Also important is making sure that providing priority for one or more flows does not make other flows fail. QoS technologies provide the elemental building blocks that will be used for future business applications in campus, WAN, and service provider networks.

Random Neural Network (RNN) -

The RNN model is introduced by E. Gelenbe in 1989, and 1990. In the RNN model signals travel as voltage spikes. This model represents more closely the manner in which signals are transmitted in a biophysical neural network than widely used artificial neuron models in which signals are represented by fixed signal levels. Signals in the form of spikes of unit amplitude circulate among the neurons. Positive signals represent excitation and negative signals represent inhibition. Each neuron's state is a non-negative integer called its potential, which increases when an excitation signal arrives to it, and decreases when an inhibition signal arrives. Thus, an excitatory spike is interpreted as a ``+1'' signal at a receiving neuron, while an inhibitory spike is interpreted as a ``-1'' signal. Neural potential also decreases when the neuron fires. Firing occurs at random according to an exponential distribution of a constant rate and signals are sent out to other neurons or to the outside of the network. A backpropagation type learning algorithm for recurrent RNN model is introduced by Gelenbe in 1993.

Realtime Transport Protocol (RTP) -

An IP protocol that supports realtime transmission of voice and video. It is widely used for IP telephony. An RTP packet rides on top of UDP, the non-reliable counterpart of TCP, and includes timestamping and synchronization information in its header for proper reassembly at the receiving end. Realtime Control Protocol (RTCP) is a companion protocol that is used to maintain QoS. RTP nodes analyzes network conditions and periodically send each other RTCP packets that report on network congestion.

Speech Codec -

Also called a "voice codec" or "vocoder," it is a hardware circuit (chip) or software rouine that converts the spoken word into digital code and vice versa. A speech codec is an audio codec specialized for human voice. By analyzing vocal tract sounds, a recipe for rebuilding the sound at the other end is sent rather than the soundwaves themselves. As a result, the speech codec is able to achieve a much higher compression ratio which yields a smaller amount of digital data for transmission. However, if music is encoded with a speech codec, it will not sound as good when decoded at the other end.

Streaming Audio -

Audio transmission over a data network. The term implies a one-way transmission to the listener, in which both the client and server cooperate for uninterrupted sound. The client side buffers a few seconds of audio data before it starts sending it to the speakers, which compensates for momentary delays in packet delivery. Because of the buffering, streaming audio can be delivered over a slow network. Audio conferencing, on the other hand, requires realtime, two-way transmission, which means the bandwidth must support the speed of both incoming and outgoing audio streams without buffering any of it. Streaming audio differs from downloading an audio file that can be played later, because the latter remains in the computer. The streaming audio is stored as a temporary file that is deleted when the media player is closed.

Streaming Video -

Video transmission over a data network. It is widely used on the Web to deliver video on demand or a video broadcast at a set time. In streaming video, both the client and server software cooperate for uninterrupted motion. The client side buffers a few seconds of video data before it starts sending it to the screen, which compensates for momentary delays in packet delivery. Because of the buffering, streaming video can be delivered over a slower network. Videoconferencing, on the other hand, requires realtime, two-way transmission, which means the bandwidth must support the speed of both incoming and outgoing video streams without buffering any of it. Streaming video differs from downloading a video file that can be played later, because the latter remains in the computer. The streaming video is stored as a temporary file that is deleted when the media player is closed.

Transmission Control Protocol/Internet Protocol (TCP/IP) -

It is the protocol of the Internet and has become the global standard for communications. TCP provides transport functions, which ensures that the total amount of bytes sent is received correctly at the other end. UDP, which is part of the TCP/IP suite, is an alternate transport that does not guarantee delivery. It is widely used for realtime voice and video transmissions where erroneous packets are not retransmitted. TCP/IP is a routable protocol, and the IP part of TCP/IP provides this capability. In a routable protocol, all messages contain not only the address of the destination station, but the address of a destination network. This allows TCP/IP messages to be sent to multiple networks (subnets) within an organization or around the world, hence its use in the worldwide Internet. IP accepts packets from TCP or UDP, adds its own header and delivers a "datagram" to the data link layer protocol. It may also break the packet into fragments to support the maximum transmission unit (MTU) of the network. Every client and server in a TCP/IP network requires an IP address, which is either permanently assigned or dynamically assigned at startup.

User Datagram Protocol (UDP) -

A protocol within the TCP/IP protocol suite that is used in place of TCP when a reliable delivery is not required. For example, UDP is used for realtime audio and video traffic where lost packets are simply ignored, because there is no time to retransmit. If UDP is used and a reliable delivery is required, packet sequence checking and error notification must be written into the applications.

Video Compression -

Encoding digital video to take up less storage space and transmission bandwidth.

Video on Demand (VoD) -

The ability to start delivering a movie or other video program to an individual Web browser or TV set whenever the user requests it.

Videoconferencing -

A video communications session among three or more people who are geographically separated. This form of conferencing started with room systems where groups of people meet in a room with a wide-angle camera and large monitors and conference with other groups at remote locations. Federal, state and local governments are making major investments in group videoconferencing for distance learning and telemedicine.