ACD (Automatic Call Distribution/Distributor)A specialized phone system, or the service it provides, for handling many incoming calls. Typically used by airlines and hotels, it recognizes and answers incoming calls according to instructions in a database, before sending the call to an operator or agent. It also offers management information on the type and volume of calls and efficiency of the agents.

ACF/NCP (Advanced Communication Function/Network Control Program) In host based IBM SNA networks, it is the control software running on a communications controller that supports the operation of the SNA backbone network.

ACF/VTAM (Advanced Communication Function/Virtual Terminal Access Method) In host-based IBM SNA networks, it is the control software running or a host computer that allows the host to communicate with networked terminals.

Actius (Association of Computer Telephone integration Users and Suppliers) A UK forum for users and suppliers to increase awareness of the business benefits of CTI. Act us develops education programs and information campaigns on CTI.

Address One or a group of characters specifying the recipient or originator of transmitted data. An address car also denote the position of data in computer memory or the data packet itself while in transit through a network. IEEE 802.3 and 802.5 recommend having a unique address for each device worldwide.

ADPCM (Adaptive Differential Pulse Code Modulation) A ITU-TS standard technique for voice encoding and compression. It allows an analog to be carried within a 32Kbit/s digital channel.

Adjusted Ring Length When a segment of Token Ring (in practice a dual ring) trunk cable fails, a function known as the Wrap connects the main path to the backup path. In the worst case - the longest path - would occur if the shortest trunk cable segment tailed, so ARL is calculated during network design to ensure the system will always work.

Agent A software-driven process running on a communications or networking device that allows that device to participate in a network management system. For example, an SNMP agent running on a router provides the ability for the router to exchange information with an SNMP network management system through the use of the SNMP protocol.

ADSP (Apple Datastream Protocol) A transport mechanism for interprocess communications between Apple Macintosh and Dec Vax minicomputers.

AFP (Apple Filing Protocol) A standard means of presenting the filing system of a server to the user with a consistent Apple Macintosh interface.

Aggregate bandwidth The total bandwidth of channel carrying a multiplexed bit stream.

Alerts Messages that Microsoft's LAN Manager network operating system sends under certain conditions. The three classes of alerts are admin alerts, error alerts and printer alerts.

Algorithm A process or set of rules necessary for a computer or intelligent device to perform a task, such as voice compression.

Alternate buffer Two buffers are sometimes used to handle data I/0. These are a alternated to achieve continuous throughput.

Alternate routing - Safety technique enabling communication to continue iii the event of node failure or congestion. The network design allows for alternate paths through the network to arrive at the same destination.

Analog An analog (US analog) signal is electrical and varies constantly in voltage, unlike a digital signal which varies between two constant values, usually denoted as 0 and 1. The value of the analog signal varies all the time during transmission, whereas a digital signal changes on y between two set values without intermediate variations.

Ansi (American National Standards Institute) A group that defines US standards for the information processing industry. Ansi participates in defining network protocol standards and represents the US on other international standards-setting bodies like ISO.

Applications Programming Interface (API) Software designed to make computer functions available to an application program PC and network operating systems have them. APIs in a network must be compatible to ensure programs are accessible to machines other than those they reside in. Some APIs, such as NetBios, are de facto standards.

APPC (Advanced Program-to-Program Communications) A set of IBM protocols also known as LU 6.2 and Type 2.1 architectures. It functions within SNA's APPN to support peer to-peer communications between workstations attached to SNA LANs and the applications running on those workstations. It was added to SNA as part of the "new" SNA to support peer to-peer networking, unlike the traditional hierarchical SNA approach in which the mainframe acts as host or master and treats the other computer as a terminal or slave.

APPC/PC A version of APPC developed by IBM to run on PC based Token Ring networks.

APPN (Advanced Peer to Peer Networking) An extension to SNA which routes information around the IBM network without help from the host, allowing systems to adjust dynamically to the topology of the network (dynamic routing). APPN keeps track of network topology, making it easier to connect and reconfigure. It also creates a directory of network nodes and other resources. APPN also allows for dynamic SNA networks, where nodes can join and leave the network as required, and session routes can be selected as needed.

AppleShare Apple system software that allows sharing of files and network services via a file server in the Apple Macintosh environment.

AppleTalk A seven- layer protocol stack developed by Apple for communications between its Apple Macintosh product range. Apple defines it in similar terms to the functionality of the seven-layer OSI model,

Application Layer The top layer in the OSI Reference Model comprising the interface between the OSI environment and a user's application. It does not contain applications, but provides a link from application software on one system to applications an another computer through the OSI environment. Several applications layers support different user tasks such as e mail and file transfer and transaction processing.

ARP (Address Resolution Protocol) The lnternet and TCP/IP protocol used to bind dynamically a high-level IP address, such as an lnternet address, to a low-level physical hardware address. ARP operates only across a single physical network and is limited to networks supporting hardware broadcast.

Arpanet The Advanced Research Projects Agency Network developed by the eponymous research agency in the 1960s as the first, large scale, packet switched network. It is still I in use today, connecting a large number of universities in the US and Europe, as well as commercial users.

ASCII The American Standard Code for Information Interchange developed by ANS I to encode characters in seven bit units. These are normally padded out with an eighth bit that can represent parity to make up an eight-bit byte. This eighth bit can also be used to make ASCII support international character sets, extending the 128 possible seven-bit combinations to 256.

Asic (Application-Specific Integrated Circuit) Pronounced A sick, it is a Very Large Scale Integrated circuit, custom-designed to perform one or more particular functions. Advantages include fewer discrete components, lower power consumption and increased reliability.

ASN-1 (Abstract Syntax Notation-1) A formal language used for describing and implementing ISO OSI protocols used in the automated implementation of protocol software. The protocol data units of most Application Layer standards like ACSE, FTAM, MMS, are defined using ASN-1.

Asynchronous communications A method of transmitting data in which each transmitted character is sent separately. The character has integral start and finish start and stop bits so that the character can be sent at an arbitrary time, and separate from any other character. It is the most rudimentary type of communication as the originating and receiving machines do not have to be synchronized. Cheap, reliable and common among PCs and minicomputers, its disadvantage is the large number of extra bits needed for the data to be interpreted.

AT Modem control language for asynchronous dial-up modems designed by Hayes Micro- computer Products.

ATM (Asynchronous Transfer Mode) A cell- based data transfer technique in which channel demand determines packet allocation. ATM offers fast packet technology, real time, demand led switching for efficient use of network resources. It is also the generic term adopted by ANSI and the ITU-TS to classify cell relay technology within the realm of broadband WANs, specifically B-ISDN. In ATM, units of data are not time related to each other and, as part of the B-ISDN standard, is specified for digital transmission speeds from 34Mbit/s to 622Mbit/s. IBM currently offers ATM at a non standard 25Mbit/s format. ATM will be the high band width networking standard of the decade.

Attenuation The weakening of transmitted signals as they travel away from their point of origin. Amplifiers can recharge the signal up to a point.

AUI (Attachment Unit Interface) The IEEE 802.3 specified cable and connector used to attach devices to a MAU. Defined in Section 7 of the 802.3 standard.

Auto partition A feature of 10 BaseT. When 32 consecutive collisions are sensed by a port in a hub or concentrator from its attached work station or network segment, or when a packet that far exceeds the maximum allowable length is received, the port stops forwarding packets. The port continues to monitor traffic and will automatically begin normal packet forwarding when the first correct packet is received.

BABT (British Approvals Board for Telecommunications) An independent organization that tests telecommunications equipment. Its processes are known for their rigorousness and labyrinthine complexity.

Back end The server part of a client/server application. It provides services across the network that have been requested by the client. For example, a back end may be a database server that responds to SQL requests from a workstation running a front end application.

Back-up server Software or hardware which copies files so that there are always two current copies of each file. Also known as a shadow server.

Backbone A high-capacity network that links together other networks of lower capacity. A local backbone network would typically be an FDDI network acting as an in building backbone to link together multiple LANs. A wide area backbone network would typically use digital leased circuits and multiplexers or routers.

Background Task or Mode A secondary function perforated by a computer without interrupting its current or primary task.

Back-up domain controller A server in a network domain that keeps and uses a copy of the domain's user accounts database to validate logon requests.

Balun A transformer that levels out impedance differences so that a signal generated on to a coaxial cable can transfer on to twisted pa r. Baluns are often used so that IBM 3270 terminals can run off twisted pair, or to allow co-axial Ethernet to be operated over UTP.

Bandwidth The range of frequencies a transmission line or channel can carry: the greater the bandwidth, the greater the information - carrying capacity of a channel. For a digital channel this is defined in bit/s. For an analog channel it is dependent on the type and method of modulation used to encode the data.

10Base2 A form of Ethernet and IEEE 802.3 network cabling using thin coaxial. It refers to I0Mbit/s speed Baseband transmission over 200 meters maximum length in practice 185m) and is commonly known as Cheapernet.

10Base5 A form of Ethernet and IEEE 802.3 network cabling using thick coax. It refers to 10Mbit/s speed Baseband transmission and 500m maximum length.

10BaseT A form of Ethernet and I EEE 802.3 network cabling using twisted pair cabling. It refers to 10Mbit/s speed Baseband transmission twisted pair cable with a maximum segment length of 100m.

100BaseT IEEE standard from proposals by the Fast Ethernet Alliance (including 3Com and SynOptics). It will support Category 3,4 & 5 UTP cabling.

100BaseVG-AnyLan A competing proposal to 100BaseT (promoted by Hewlett Packard, IBM and Proteon among others) to the IEEE for a 100Mbit/s standard over voice grade UTP the cable most users already have installed in existing 10BaseT systems. Based on Quartet Signaling and demand priority protocol, it preserves the infrastructure and will need only a new hub and upgraded adapters in PCs/work stations. It claims support for Category 3,4 & 5 UTP cabling for both Ethernet and Token Ring.

Baseband A term defining any network in which the information is modulated onto a single carrier frequency. The digital input is applied directly to the transmission media without the intervention of a modulating device, which works well if there is wide bandwidth and distances no more than several hundred meters are involved. It is common in LANs and limited distance modems. All stations attached to the network have to participate in every transmission. Simpler and cheaper than Broadband, it permits only one "conversation" at a time as the whole of the bandwidth is used to transmit a single digital signal. Ethernet is a baseband network.

Base station A fixed radio transmitter/receiver which electronically relays signals to and from mobile voice and data terminals or handsets.

Basic Rate Access Two 64 Kbit/s "B" channels + one 16 Kbit/s "D" channel (2B + D), carrying user traffic and signaling information respectively to the user via twisted pair local loop.

Baud A unit of s gnarling speed, expressed in terms of the number of discrete conditions or signal events per second. It is on y the same as bit/s, when one discrete signaling condition is used to transmit a single bit of data.

Beaconing Token Ring process to recover the network when any attached station has sensed that the ring is inoperable because of a hard error Stations can withdraw from the ring if needed. A station detecting a ring failure upstream transmits (beacons) a special MAC frame used to isolate the location of the error using beacon transmit and beacon repeat modes.

Bindery A database that contains definitions for entities such as users, groups and workgroups in the Novel NetWare LAN network operating system environment. The bindery supports the design, organization and secure operation of the NetWare environment.

Bipolar transmission Method of sending binary data in which negative and positive states alternate. Used in digital transmission facilities.

B-ISDN (Broadband ISDN) The proposed advanced version of ISDN, providing speeds of 155.52Mbit/s and higher. Standards and switching technology that will work this fast are under development. It promises universal coverage based on ATM/SDH technologies and optical fiber, supporting data, voice and video traffic.

Bit A binary unit of information that can have two values, 0 or, 1. The word comes from a contraction of binary digit.

Bit Error Rate The percentage of received bits on a digital link that are in error relative to the number of bits received, usually expressed to a power of I 0

Bit Error Rate Tester A device for testing the reliability of a digital datacommunications link. The BERT generates specific data patterns that are routed through a communications device for comparison at the receiving end. The errors are counted by the BERT.

Bit Interleaving A form of Time Division Multiplexing (TDM) for synchronous protocols, including HDLC, SDLC, BiSync and X.25 Bit inter-leaving retains the sequence and number of bits, so that correct synchronization is achieved between both ends.

Bits per second The rate at which individual bits are transmitted across a communications link or circuit; written bit/s. One thousand bit/s is 1 Kbit/s, and one mil ion bit/s is 1 Mbit/s.

Block A group of characters or bytes treated as a unit.

BNC connector A cylindrical push-and-twist connector for connecting thin co-axial cable, such as 10Base2 "thin wire" Ethernet, and to link thin wire Ethernet to network interface cards, transceivers and other network elements. Said to be short for Bayonet Neill-Concelman after the developers of the connector. Also referred to as a Barrel Nut Connector.

Booting Loading a computer memory with information needed for it to operate. Remote booting refers to loading software over the network.

Boundary Routing A 3Com proprietary name for a method of accessing remote networked locations, such as a bank branch office. Effectively a form of bridging, the idea is to reduce the need for technical expertise locally and the cost of equipment at the remote site and manage the communications from head office.

BRA (Basic Rate Access) BRA provides ISDN users with access to two 64Kbit/s data channels, It is defined in ITU-TS Recommendation I.420 which covers a 2B + D-channel where the B channel is a 64Kbit/s channel, and the D-channel is a 16Kbit/s signaling channel.

Bridge Device connecting two separate networks at the OSI Data Link Layer (Level Two Media Access Control Layer). Once bridging is accomplished, the bridge makes interconnected LANs look like a single LAN, passing data between the networks and filtering local traffic. There are two key classifications of bridge: those supporting Spanning Tree and, for Token Ring networks, those supporting Source Routing. Bridges connect networks using dissimilar protocols and do not interpret the data they carry. They control network traffic and security, filtering where necessary to boost network, performance and contain sensitive data to particular LAN areas.

BS5750 A British Standards Institute standard with certification procedures that says an organization is in control of its quality procedures, at least in terms of consistency. Now identical to IS09000.

BSC, BiSync (Binary Synchronous Communications) Rules developed by IBM for the synchronous transmission of binary coded data as a serial stream of binary digits. Synchronization is achieved by using control characters recognizable as bit patterns which do not appear within the body of the message.

BSGL (Branch Systems General License) A license which must be obtained by any organization seeking to link its own private network to the Public Switched Telephone Network (PSTN). A separate license must be held for each individual site.

BSI (British Standards Institute) The UK standards body responsible for input into European and international standards setting bodies like ISO and the ITU-TS.

Broadband Also referred to as wideband. A term describing any network that multiplexes multiple, independent network carrier frequencies on to a single cable. It allows multiple simultaneous "conversations", since the independent networks operate on different frequencies and do not interfere with each other. In LAN terminology, it refers to a system in which multiple channels access a medium, for example co-axial cable, that has a large bandwidth using Radio Frequency (RF) modems. This may allow the co-axial cable to carry multiple separate LANs whose transmission is being modulated at different frequencies. In cable television (CATV), broadband describes the ability to carry 30 or more TV channels and is synonymous with wideband.

Broadcast The simultaneous transmission of data via a network from one terminal to a set of destinations or to all destinations.

Brouter An industry term for a device with the functionality of a bridge and router. It supports more than two LAN connections and uses Level Two addresses for routing. The term is mostly used by bridge vendors.

Buffer A temporary storage place for data, designed to compensate for a difference In transmission speeds or to hold data when there is a difference in timing of events. It can be a software program, a storage facility or a hardware device, ensuring the data always has somewhere to go, even if it has to be held up for while in the buffer until it can be transmitted to the destination.

Bus topology A type of network in which all tie devices are connected in a line to a single cable. A bus network has two distinct ends. All devices which attached to a bus network have equal access to it and they can see all the messages that are put on to the network. Each device determines which messages are intended for it alone, and selects those.

Byte Eight bits forming a unit of data. Usually each byte stores one character.

CAI (Common Air Interface) The CT2 international mobile communications standard which allows any compliant equipment to be used on any network of the same type. CAI compliant telepoint handsets from different vendors may therefore be used on a telepoint network. Vendors with CAI compliant systems include Northern Telecom and GPT.

CATV (Cable TeleVision) Cable system covering defined areas, such as the U K's franchises to install and operate a cable system granted by the Cable Authority and Department of Trade & Industry, offering TV channel output and, increasingly, local loop digital telephony services The Cable Television Association is the CATV industry's representative organization.

Call processing The system and process that sets up the intended connection in a switching system. A system scans the trunk and station ports for any requests for service. It then checks the stored instructions and look-up tables and sets the connection up accordingly.

Carrier signal The underlying frequency or frequencies that are to carry information. They are modulated through one or more modulation techniques to impose information on the signal.

Category 3 cable Standard for UTP voice grade cabling specified by the EIA/TIA 568 standard for use at speeds up to I0Mbit/s including 10BaseT Ethernet.

Category 4 Cabling standard specified by EIA/TIA 568 for use at speeds up to 20Mbit/s including 16Mbit/s Token Ring.

Category 5 Cabling standard specified by ElA/TIA 568 for use t speeds up to 100Mbit/s including FDDI (TP PMD), 100BaseT and 100BaseVG-AnyLan, and potentially ATM at 155Mbit/s.

CCITT (The International Telegraph and Telephone Consultative Committee) Former name for the ITU-TS (International Telecommunications Union), a Specialized Agency of the United Nations.

CCITT Study Groups The ITU-TS (formerly the CCITT) operates as a series of groups considering specialist areas. There are key study groups applicable to networking and communications such as Study Group VII responsible for data communications networks and the X series Recommendations and Study Group XVIII covering digital networks including ISDN.

CCTA Central Computer & Telecommunications Agency - the Government Center for Information Systems.

Cell Relay Generic term for a protocol based on small fixed packet sizes capable of supporting voice, video and data at very high speeds. Information is handled in fixed length cells of 53 octets.

Centronics interface A parallel interface with 36 pins that will transmit eight data bits simultaneously. The interface originates from the Centronics Company, a printer manufacturer. It has become widely used as a parallel interface standard.

CEPT The European Conference of Postal and Telecommunications Administrations. An association of European Telecommunications service providers. It in turn participates in relevant areas of the work of Cen/Cenelec. Formerly extremely powerful and was originally responsible for the Net standards, but these have subsequently been passed on to Etsi.

Character Interleaving A form of TDM used for asynchronous protocols. This can be used either with extra channels, or by carrying RS232-C control signals.

Cheapernet Thin wire Ethernet.

Circuit switching The transmission technique in which a physical circuit is established between sender and receiver before transmission takes place. When the transmission is complete, the circuit is freed.

CLI (Calling Line Identification) A service available on digital phone networks that tells the person being called which number is calling them. The central office equipment identifies the phone number of the caller, enabling information about the caller to be sent along with the call itself. Synonymous with ANI (Automatic Number Identification).

Client/server computing The division of an application into two parts; a front end client and a back end server. It allows multiple front ends running on a PC or Unix workstation (client) to access the same SQL based server database at the same time over the LAN. The aim is to off-load as much processing as possible to the intelligent desktop leaving only the shared information and the software for managing it at the central server. An application that is running in such a fashion with client and server linked by a LAN is termed a bifurcated application.

Cluster controller An IBM device that allows multiple 3270 terminals to be linked directly to a host computer, or into a SNA network through the use of a communications controller. A cluster controller is a Control Unit in IBM speak.

CMIP/CMIS (Common Management Information Protocol/Common Management Information Services) ISO OSI connection oriented network management protocol and set of services. Well accepted in the WAN and telecommunications world, they have not yet been widely adopted for LAN management.

CMOT (CMIP/CMIS over TCP) The use of SO CMIP/CMIS network management protocols to manage gateways in a TCP/IP Internet. CMOT is a co-recommended standard with SNMP.

Co-axial cable A cable comprising a central wire surrounded by a second tubular screening of fine wire. Associated with IBM for linking terminals and other devices needing high-speed links, coax is used in Ethernet. It is difficult to add or remove devices from a coaxial LAN as the cable is unwieldy and thick so is being superseded by UTP.

Collision The result of two devices on a shared transmission medium, like Ethernet, transmitting simultaneously. Data is corrupted and both devices must retry their transmissions. A delay mechanism used by both senders drastically reduces the chances of another collision.

Collision detection Devices at each end of a link are designed to detect collisions instantly and attempt to resend. This is the principle on which CSMA/CD is based and the access control method for Ethernet. An alternative is to resend if there is no acknowledgment of receipt from the remote device.

Communications Controller A switching unit central to the implementation of host-based IBM S NA networks. Typically the network is built around a backbone of interlinked communications controllers to which host computers and Control Units (CUs) are attached.

Communications Manager An individual often underpaid and invariably overworked, dedicated to providing cost effective, ultimately flexible networking to users.

Communications Networks The UK's leading monthly magazine for networking professionals and decision makers.

Communications Server A specialized network server that provides access to external networks, communications facilities and hosts that cannot be directly connected to the LAN. Typically it will enable workstation users running appropriate workstation software, such as terminal emulation software, to access asynchronous communications links and typically modems attached to the communications server.

Communications Toolbox An extension of the Apple Macintosh operating system that provides protocol conversion and the drivers needed for communications tasks.

Concentration A technique used to get the most out of a composite multiplexer link. Usually a statistical multiplexer, or concentrator, is used to focus channel inputs on to the composite ink by removing the portions of the transmission carrying no data.

Concentrator A central chassis into which various modules, such as bridging, supervisory, 10BaseT and other peripheral cards are plugged.

Connection-oriented service The transport of packets of information from one network node to a destination node following an established network connection.

Connectionless Service The transport of a single datagram or packet of information from one network node to a destination node or multiple nodes without establishing a network connection.

Contention The process whereby multiple users make requests for transmission bandwidth across a transmission link but the pool of bandwidth is less than the aggregate amount of bandwidth the users could request between them. Contention is used to resolve which users gain access to the bandwidth. When this s applied to multiplexers, it is concerned with the multiplexer's ability to allow a number of channels to contend for transmission bandwidth that is less than the sum of all the channel rates.

Control point A program that manages an APPN network node and its resources, enabling communications to other control points in the network.

Converter A repeater that also converts from one media type to another, such as from fiber to copper. Often called a media adapter.

CPE (Customer Premises Equipment) Telecommunications- communications equipment, including PBXs and wiring, located in a user's premises.

CRC (Cyclic Redundancy Check) A method of detecting errors in the serial transmission of data. A CRC for a block of data is calculated before it is sent, and is then sent along with the data. A new CRC is calculated on the received data. If the new CRC does not match the one that has been sent along with the data then an error has occurred.

Cross-Connect An ATM switch usually comprising three functional areas. System control The central control unit, which also provides the management interface of the system; the ATM "fabric block" providing the system switching capacity; termination groups to provide the external interfaces and the functions of the ATM layer of the network node. Each of these functional system areas is configured according to the specified needs of the respective network node. Each functional area usually has its own monitoring and control units for safeguarding the high availability of the complete system.

Crosstalk Unwanted interference from another adjacent communications channel . The signal from the adjacent channel is inserted into the original communications channel.

CSMA/CD (Carrier Sense Multiple Access with Collision Detection) The access method used in Ethernet. All nodes are attached to a single cable and contend equally for access to the transmission medium. if two nodes attempt to send data at the same time, they "sense" each other's signal and immediately stop sending. They will both try to send again after Waiting a random number of microseconds.

CSMA/CA (Carrier Sense Multiple Access with Collision Avoidance) A method of network access not covered by OSI standards and used in AppleTalk networks.

CSU (Channel Service Unit) (1) In the US, data transmission equipment to repeat the signal from the carrier and ink to CPE. Vendors add value to CSUs by adding performance monitoring and management. (2) In Europe, CSUs are sold for their value features like diagnostics and performance monitoring. The basic repeating function is prd in the NTU (networking terminating unit). CSUs monitor quality on El, E2 or E3 circuits in terms of transmission and line loading.

CT1 First generation analog domestic cordless telephone (non-cellular).

CT2 Two-way digital cordless telephony technology, particularly relevant to cordless PBXs. In its public guise, it becomes a one way telepoint service now no longer available in the UK but prevalent in the Far East.

CT3 Ericsson's proprietary cordless telecommunications system.

CTI (Computer Telephone Integration) A generic name for the technology automatically relating computers and PABXs via applications such as ACD, power dialing, IVR and other customer facing or agent facing services. A so known by older, proprietary names CIT (Computer Integrated Telephony) and CSTA (Computer Supported Telephony Applications).

Darpa (Defense Advanced Research Projects Agency) Formerly called Arpa, this US government agency that funded research and experimentation with the Arpanet and later, the connected Internet- The group within Darpa responsible for the Arpanet is ISTO (information Systems Techniques Office), formerly IPTO (Information Processing Techniques Office).

DassiII A message based signaling system following the ISO based model developed by BT to provide multi-line IDA interconnection to the BT network.

Data compression A way of reducing the amount of data to be transmitted by applying one of severs techniques that reduce the number of bits needed to represent the information. When the data is received It is decompressed into its original form.

Database server A database installed as a back-end or server component of a client-server system, which can be accessed over a LAN by one or more client, or front-end applications through the use of query language, typically SQL. The server part of the program is responsible for updating the records, ensuring that multiple access is available to authorized users, protecting the data and communicating with other servers holding relevant data. The client end of the program requests records and then modifies them, while the server tracks records down for the client and adds new ones.

Datagram A method of sending data in which parts of the message are sent in random order. The recipient machine has the task of reassembling the parts in the correct sequence. The datagram is a connectionless, single packet message or item of data that can traverse a network at OS I Level Three, the Network Layer. It typically does not involve end-to-end session establishment or delivery-confirmation acknowledgment. As well as the information within the datagram, there is a destination network address and usually a source network address.

Data link A direct serial data communications path between two devices without intermediate switching nodes.

Data Link Layer Layer Two of the ISO OSI model is responsible for the transmission of information over a physical medium. After establishing the link it ensures the error-free delivery of the information through the use of error detection, error recovery and flow control. The contention access methods such as CSMA/CD and Token passing are Layer Two activities.

Data PBX A switching system for data traffic that allows terminals and workstations connected by individual cables to the Data PBX selectively to link to one or more host computers over asynchronous circuits through the use of contention.

DCA (Defense Communication Agency) The US government agency responsible for the installation of Defense Data Networks, like Arpanet and Milnet, and PSNs. The DCA writes contracts for operation of the DDN and pays for network services.

DCA (Document Content Architecture) The IBM approach to storing documents as two types of document group: draft documents and final form documents. For presentation, the draft document is transformed into a final document through an office system.

DCE (Data Circuit Terminating Equipment) Communications equipment installed in a user's premises responsible for establishing, maintaining and terminating a connection. A modem is an example.

DCE (Distributed Computing Environment) A suite of software utilities and operating system extensions that will, in theory, create applications on networks of heterogeneous hardware - PCs, Unix workstations, minicomputers and mainframes. The DCE is the product of the OSF. The DCE is designed to simplify the building of heterogeneous client/server applications and provides seven general services: Remote Procedure Call, Security, Naming (directory), Distributed File System, Threads, Time and PC Integration. DDE (Dynamic Data Exchange) - A Microsoft messaging specification. When DDE-compliant applications are combined, dynamic documents can be created which update each other as data changes.

DDM (Distributed Data Management Architecture) An IBM SNA LU 6.2 transaction providing users with facilities to locate and access data in the network. It involves two structures: DDM Source, and DDM Target. The Source works with a transaction application to retrieve distributed data and transmits commands to the Target program on another system where the data that has been requested is stored. The Target interprets the DDM commands, retrieves the data and sends it back to the Source that originated the request.

DDCMP (Digital Data Communication Message Protocol) The DecNet- specific Link Level protocol that operates at Layer Two of the Digital Network Architecture.

DDN (Defense Data Network) Used generally to refer to Milnet, Arpanet and the TCP/IP protocols those networks use. More specifically refers to Milnet and associated parts of the connected Internet that connect military installations.

DecNet Proprietary peer-to-peer network technology originally developed for use in wide area networking by the Digital Equipment Corporation (Dec) and evolved to include significant Ethernet-based LAN capabilities. It is the implementation of the Digital Network Architecture (DNA).

Dect (Digital European Cordless Telecommunications) A standard governing pan-European digital mobile telephony. Based on advanced TDMA technology, Dect covers cordless PBXs, telepoint and residential cordless telephony.

Demand Priority Access method providing support for time-sensitive applications such as video and multimedia as part of the proposed 100BaseVG standard offering l00Mbit/s over voice-grade UTP cable. By managing and allocating access to the network centrally, at a hub rather than from individual workstations, sufficient bandwidth for the particular application is guaranteed on demand. Users, say its proponents, can be assured of reliable, continuous transmission of information.

Demodulation Technique for retrieving information from a modulated signal. Demonstrated by the eponymous modem (modulator/demodulator).

Des (Data Encryption Standard) An algorithm designed by the US National Bureau of Standards for the encryption and de-encryption of data using a 64-bit key.

Device driver In the context of computer networking a device driver is a software module forming part of a computer operating system, or software that interacts with the operating system. It aims to control communications equipment, such as a LAN network adapter card and facilitate the transfer of information to and from the network. Other examples of device driver programs include software to support the activities of printers, disks and mice.

DIA (Document Interchange Architecture) An IBM term defining the sets of functions needed for document handling in an IBM environment, including storage and distribution.

Digital signal A signal with only two values, normally 0 and 1, during transmission, unlike an analog signal whose values constantly vary.

Direct attachment The IBM term for linking a device or LAN directly to a host computer through an appropriate Control Unit, like a cluster controller.

Disk server A device equipped with disks and a program permitting users to create and store files on those disks. Each user has access to their own section of disk on the disk server. The aim is to give users access to disk space that they would not normally have on their PC. The disk server is linked to the PCs via a LAN. The next level of sophistication would be a file server.

Diskless workstation A PC or workstation attached to a LAN that has neither floppy nor hard disks, but relies on disk storage provided by a file server attached to the same LAN. When the diskless workstation is first initialized it uses a remote boot program stored in a remote boot prom/eprom on its network adapter card to initialize a session with the file server. The workstation then loads its operating system, such as MS-Dos, from the server and executes the normal server login procedure.

Distributed database A database stored on more than one networked computer. The database is split up across these machines, and not replicated.

Distributed name service A technique for storing network node names so that the information is stored throughout the network, and can be requested from, and supplied by, any node.

Disoss (Distributed Office Support Systems) IBM software typically forming part of an IBM Office System Node.

Distributed computing The trend away from having big, centralized computers such as mini-computers and mainframes to bring processing power to the desk top. Often confused with distributed processing.

Distributed processing An approach that allows one application program to execute on multiple computers linked together by a network. The networked computers share the work between them.

DLS (Data Link Switching) An enhancement to source routing which transports source route packets over a resilient IP/OSPF network and provides local termination of LLC2 sessions to avoid LLC timeouts in large or busy networks. It is the ideal mechanism for mixed LAN-to-LAN and interactive SNA traffic since it can recover from network problems quickly using OSPF. It is rapidly becoming accepted as a major standard.

DMA (Direct Memory Access) A technique for high-speed data transfer between a device such as LAN network adapter card and the computer memory. DMA bypasses the Central Processing Unit of the computer, PC or workstation, allowing the device to transfer a block of information directly across the bus into system memory.

DMI (Desktop Management Interface) A set of APIs outlined by the DMTF, comprising three components: service layer, component interface and management interface.

DNA (Digital Network Architecture) The network architecture of Digital Equipment Corporation with eight layers. The DNA is similar in structure to OSI at lower levels, except that the top three layers of the DNA correspond to the top two layers in the OSI model.

DNS (Domain Name System) The online distributed database system used by Internet to map names into IP addresses. DNS servers throughout the connected Internet implement a hierarchical namespace that allows sites freedom in assigning machine names and addresses. DNA also supports separate mappings between mail destinations and IP addresses.

Domain A group of nodes on a network that form an administrative entity. It could also be a number of servers grouped together and named to simplify network administration and security. Every computer on the LAN belongs to at least one domain. Being logged in on one domain, however, does not limit resources in other domains to which the user has access permissions.

Dos - Disk Operating System comprising one or a suite of programs managing a disk-based computer system. Dos schedules and supervises work, allocating computer resources and the operation of peripherals. Versions of Dos from different vendors exist: Microsoft's MS-Dos is the most common. Dos 3.1 was the first version of MS and PC Dos able to support LAN functions separate, of course, from the network's own operating system - notably including record and file locking which is now standard on multi-user systems.

Dos LAN Manager A Dos version of Microsoft's network operating system LAN Manager. It gives Named Pipes (an applications interface) support to Dos machines, enabling them to use the client/server environment.

Downlink Transmission from a satellite to an Earth Station.

DPA (Demand Protocol Architecture) A technique for loading protocol stacks dynamically as they are required. It is associated with adapter cards in workstations and servers. Only the protocol stacks that are needed for a particular communications sessions are loaded. Examples of stacks that could be loaded include TCP/IP, XNS, SPX/IPX and NetBios.

DPNSS (Digital Private Networks Signaling System) Signaling standard for digital private networks within the UK formulated jointly by BT and PABX manufacturers.

DQDB (Distributed Queue Dual Bus) The standard for future Mans which operates as a dual bus, each carrying data in both directions. A queuing system maintains transmission order. Some similarity with ATM encourages evolution between the technologies.

Drop cable A cable that links a network adapter to an external transceiver attached to a co-axial LAN such as Ethernet. Also called an Attachment Unit Interface cable or transceiver cable.

DS1 (Digital Signal 1) Transmission standard at T1 speeds, or 1.544Mbit/s

DS3 (Digital Signal 3) Transmission standard at T3 speeds, or 44.736Mbit/s. DS3 allows the combination of 28 DSls or a single DS3 facility - also known as a T3 circuit.

DSE (Digital Switching Exchange) A node in a telecommunications network.

DSU (Data Service Unit) Data transmission equipment used to interface to a digital circuit at customer site. It converts the customer's datastream, such as X.21 to E1 or T1 for transmission through the CSU, which is often contained, functionally within the DSU device. DSUs can convert data to or from a native port on a router to an E1, E2 or E3 leased line, primary rate ISDN or SMDS, DSU functionality can be built into devices such as some routers or multiplexers. In Europe a DSU can convert El bandwidth into RS.449, X.21, V.35 or other serial interface via a router. A DSU with an HSSI interface will deliver E2 or E3 bandwidth from the WAN to an HSSI router on a LAN.

DTE (Data Terminal Equipment) A piece of equipment where a communications path ends. The user's equipment is collectively termed DTE and can include PCs and display terminals.

DTMF (Dual Tone Multi-Frequency) A term for push button or Touchtone (an AT&T trademark) dialing. The pushed button makes a tone, actually the combination of two tones - of high and low frequency. They are necessary to access advanced network features such as call barring and call forwarding. DTMF penetration in the mass-market the UK is small but growing, but high in the business community.

Duplex Simultaneous, two-way independent transmission of data.

Dynamic node address An Apple-patented feature of AppleTalk under which each node assigns itself a unique address code each time it is initialized. Conventionally, nodes are assigned fixed addresses that do not change.

Dynamic routing A process for selecting the most appropriate path or route for a packet or datagram to travel around a network. At the end of each leg of the journey of the packet across the network the router decides on the most appropriate path for the packet or datagram to follow if there are multiple routes available. This is done using network status information gathered from around the Internet and passed from router to router through the use of routing information protocols.

E1 The European standard for high-speed data transmission at 2.048Mbit/s - 32 64Kbit/s channels are provided.

Earth station Transmitting or receiving point for satellite transmission.

EBCDIC (Extended Binary Coded Decimal Interchange Code) An IBM-developed eight-bit binary code that can represent 256 characters. It allows the representation of control codes and graphics in a logical format. It was created to represent data in particular types of data processing and communications terminal devices.

Echo cancellation A technique used in high-speed modems to isolate and remove unwanted signal energy created by echoes of the transmitted signal. (An echo is the return of a transmitted signal). It would be noticeable by its absence on transatlantic or other time-differentiated phone conversations.

ECMA (European Computer Manufacturers Association) An organization dedicated to the co-operative development of standards applicable to computer technology. It works closely with certain ITU-TS Study Groups and ISO Sub-committees.

EDI (Electronic Data Interchange) The interchange of trading information through the use of an electronic messaging system. A form of paperless trading that is rapidly expanding through organizations such as Geis.

EGP (Exterior Gateway Protocol) The protocol used by a gateway in one autonomous system to advertise the IP addresses of networks in that system to a gateway in another autonomous system.

EIA/TIA The US Electronics Industries Association and Telecommunications Industries Association which have merged. The EIA is a US trade organization for the electronics industry that concentrates on hardware interface standards. Best known for the RS232C and RS422 standards that specify the electrical characteristics of interconnections between terminals and computers or between two computers.

EIA/TIA 568 The EIA/TIA's most famous standard specifying a maximum horizontal cable run of 90 meters, allowing 10 meters for hub and device attachment resulting a total cable length of 1000 meters. Its related Service Bulletins TSB 36 and 40 define three categories of cabling system for data communications: Categories 3, 4 & 5.

EISA (Extended Industry Standard Architecture) A 32-bit adaptation of the 8/16-bit buses originally developed by IBM and now standard in almost all PCs that use Intel's 8086 and 80X86 chips. The EISA bus is a joint development from Compaq and other PC manufacturers. Contrast with Microchannel and ISA.

Electronic mail, e-mail The electronic transmission and reception of messages and text-based information without the need for the recipient to be present at the time of the transmission. Available either as a value-added service from providers such as GEIS, IBM and MCI or on LAN systems such as ccMail. De jure standards are evolving, but gateway software in the interim is now sufficiently mature to cope with most interactivity.

EMS (Element Management System) The level of a network management system concerned with collecting network management information from, and setting parameters on the network elements. Network elements consist of data communications and telecommunications equipment.

Emulation Hardware or software, or a combination of the two, that behaves like another device or program, like PCs emulating dumb terminals.

Encapsulation The process of sending data encoded in one protocol format across a network operating a different protocol, where it is not possible or desirable to convert between the two protocols. For example, where Ethernet LANs attach to an FDDI backbone, it is not possible to convert between the different packet formats, so the Ethernet packet is encapsulated in its entirety inside an FDDI packet as it crosses the bridge on to the FDDI network. When the encapsulated Ethernet packet reaches the bridge connecting the destination Ethernet LAN to the FDDI network, the Ethernet packet is stripped out of the FDDI packet and put, unchanged, on to the destination Ethernet LAN. Also known as protocol tunneling.

Entry point An IBM network management term. An entry point provides management functions for itself and the devices attached to it. It has to be an SNA-addressable unit, allowing it to participate in network management by monitoring its own environment and exchanging information and messages with a Focal point.

Error control A means of ensuring that information received across a transmission link is correct. The techniques involved typically use error detection to detect if the transmitted data has been corrupted. The error control technique involves asking for data to be retransmitted until a correct version is received.

Error correction A technique to restore data integrity in received data that has been corrupted during transmission. Error correction techniques involve sending extra data along with the original data being sent. It allows the correct form of the data to be reconstructed from the extra information if the original has been corrupted. This extra information is calculated using particular error correction algorithms such as Hamming Code. It allow errors to be detected and the original data reconstructed. This is sometimes termed forward error correction.

Error detection A set of techniques that can be used to detect errors in received data. Techniques that are applicable include parity checks involving parity bits, checksums or a Cyclic Redundancy Check.

Ethernet The most widely LAN transmission network. Based on a bus network topology, it runs at a maximum 10Mbit/s - in practice far less - and adopts CSMA/CD techniques operating over convention co-axial cable, thin wire co-axial cable and unshielded twisted pair cabling. A fiber-optic implementation has also been defined. Originally developed by Xerox, Intel and Dec, Ethernet has moved through V1 and V2 proprietary definitions, and has now been standardized by the IEEE as the IEEE 802.3 standard. This has several implementations - 10Base5 for use over conventional co-axial cable, 10BaseF for use over optic fiber, and 10BaseT for use over Unshielded Twisted Pair (UTP) cabling.

Ethernet switching A technique inspired by Kalpana bringing the advantages of a parallel networking architecture to current contention- based Ethernet LANs. Each LAN can be segmented, each with its own 10Mbit/s path. When users on different segments exchange data, an Ethernet switch dynamically connects the two separate Ethernet channels without interfering with other network segments. The switch can create multiple independent connections between separate segments, allowing multiple parallel data exchanges. This multiplies network bandwidth without modification to Ethernet end station hardware or software. Once disparaged by other vendors, it now forms part of the portfolio of the major hub and router manufacturers.

Etsi (European Telecommunications Standards Institute) A European standards body established in 1988 by a decision of the CEPT. It has taken over the work of the CEPT the area of developing the Net-Normes Europeene de Telecommunication, Net standards.

Eutelsat Inter-governmental organization founded in 1983 with the launch of Eutelsat 1, Eutelsat aims to provide and operate a space segment for public intra-European international telecommunications services. The segment is also used to meet domestic needs by offering leased capacity, primarily for television. UK and France are the main shareholders, with about 25 members in total.

Ewos (European Workshop for Open Systems) A forum aimed at promoting OSI standards and undertaking the development of functional profiles. Its work includes OSI Layers One to Four, FTAM, MHS, ODA, Directory Services and the VT protocol.

Fast Ethernet Proposed 100Mbit/s technology for workstation LANs from the eponymous Fast Ethernet Alliance, which includes 3Com and SynOptics. It has been adopted by the IEEE as the basis for the 100BaseT Ethernet standard.

Fast packet switching A WAN technology capable of transmitting data, digitized voice and digitized image information. It makes use of short, fixed length packets (or cells) that are all the same size. The underlying switching technology is based on the statistical multiplexing of data and voice in fixed length cells. Any of these packets could carry digital voice, data or digital image information. All the packets travel at Level Two of the OSI Model, and routing is performed on the basis of the Level Two addressing. Fast packet is an effective way of making best use of available bandwidth. It offers the benefits of conventional multiplexing techniques and circuit switching techniques. It is one of the transmission technologies being developed for use with B-ISDN. The switch used to route packets in a fast packet network is termed a fast packet switch. Also, fast packet technology can carry data transmissions that enter the network using a frame relay access method. For particularly high speed networking, an implementation of fast packet switching known as ATM is being commercially developed.

Fault tolerance A method of making a computer system or network resilient to faults or breakdowns to avoid lost data and downtime. For servers this involves such techniques as disk mirroring, disk duplexing or mirrored servers. For LANs and WANs it may involve the use of multiple redundant transmission links.

Fax server A specialized IVR system which sends facsimile messages to a fax machine designated by DTMF tones. What amounts to a database of fax text resides in the server that is accessed via a user's DTMF phonepad. Requests result in the fax pages being delivered to the chosen fax and the subscriber charged a fee.

FCC (Federal Communications Commission) US regulatory and approvals agency.

FDDI (Fiber Distributed Data Interface) An optical fiber-based token-passing ring LAN technology with dual counter-rotating rings. Each ring carries data at a rate of 100 Mbit/s using a 125MHz transmission frequency. It has been standardized by the American National Standards Institute (ANSI). An FDDI network has two modes of attachment: a device may be a Single Attach Station - attached to one ring - or a Dual Attach Station - attached to both rings. Typical applications of FDDI are in the area of high speed LAN backbones.

File server A computer, attached to a LAN, that runs a Network Operating System (nos). This lets the file server regulate communications among the workstations connected to it across the LAN, and to manage shared resources available on the file server, such as hard disk storage and printers. A file server may be dedicated: the computer is used only as a file server; or non- dedicated: the underlying computer that the LAN nos runs on is used for another task simultaneously, for example as a workstation.

Flow control The procedures for controlling the rate of transfer of data between two points in a data network, such as between a protocol converter and a printer. This avoids data loss when a recipient device's buffer is full. Buffers play an essential role in overall flow control in a network.

FNC (Federal Networking Council) A US group of representatives from those federal agencies involved in the development and use of federal networking, especially those networks using TCP/IP, and the connected Internet. The FNC coordinates research and engineering. Members include representatives from the DoD, DOE, Darpa, NSF, Nasa and HHS.

Focal point An IBM Network management term; it consolidates the functions needed to manage centrally all parts of a network. It provides an end-to-end network view and receives information from entry points and service points. NetView is IBM's key implementation of the focal point.

Foirl (Fiber Optic Inter-Repeater Link) Defined in IEEE 802.3 and implemented over two fiber links, transmit and receive, this medium may be up to 500m and 1 kilometer long depending on the number of repeaters in the network.

Fractional services Bandwidth available from carriers in increments of 64Kbit/s, such as Mercury's Switchband.

Frame A group of bits sent over a link. A frame may contain control and addressing information, as well as error detection - for example CRC information - and forward error correction information. The size and composition of the frame varies according to the protocol. Often used synonymously with packet.

Frame relay A data communications interface originating from ISDN designed to provide high speed frame or packet transmission with minimum delay and efficient use of bandwidth. It is a variation on the X.25 interface and form of fast packet switching. It derives its name from using the Data Link or "frame" OSI layer Two to route or "relay" a packet directly to its destination instead of terminating the packet at each switching node. This eliminates processing overheads and increases throughput speed. Based on the ITU-TS Lap-D standard, it uses variable-length packets and applicable only to sub-broadband, T3/E3 or lower, data transmission. Like Ethernet, or token ring, frame relay assumes that connections are reliable. It does not have error detection and error control within the network, which helps to speed up the protocol. When errors occur frame relay relies on higher level protocols for error control. Frame relay is often viewed as a replacement for X.25, primarily for LAN-to-LAN bursty traffic. Voice over frame relay is available, but the subject of debate. It will also become an access method for ATM-based WANs.

FDM (Frequency Division Multiplexing) A multiplexing scheme in which the available transmission frequency range is divided into narrower bands. Each of these bands is used to carry a separate channel.

FSK (Frequency Shift Keying) A technique for modulating data that use two frequencies. Frequency shifts between the two frequencies are generated when the binary digital level changes. So one particular frequency is used to represent a binary one, and a second frequency is used to indicate a binary zero. FSK is used in low speed modems when, in full-duplex transmission, two different frequencies are used in each direction, resulting in four different frequencies being used.

Front end The client part of a client/server application that requests services across a network from a server, or back end. It typically provides an interactive interface to the user, for example, a data entry front end, allowing data to be entered into a server through the use of SQL.

FTAM (File Transfer Access and Management) ISO 8671 standard which plays a key role in integrated message handling as the vehicle for interchanges of EDI information between applications. FTAM controls the transfer of whole files or parts of files between end systems.

FTP (File Transfer Protocol) The TCP/IP standard, high-level protocol for transferring files from one machine to another. Usually implemented as applications level programs, FTP uses the Telnet and TCP protocols. Full duplex - A channel capable of transmitting in both directions at the same time.

Functional profile A defined stack of ISO OSI Layer elements, such as Gosip, Map or Top. Functional profiles were developed in order to ensure that, when defined, ISO OSI stacks could interoperate. Due to the different protocol elements at each OSI layer, it was possible to define stacks that were syntactically correct, but would not be able to exchange in-formation due to differences at particular layers. A functional profile that has been defined as a standard is a standardized profile. Likewise, an International Standard Profile is an OSI functional profile.

G recommendations A series of standards defined by the ITU-TS covering transmission facilities. They are: G.703 2.048Mbit/s - transmission facilities running at 2.048Mbit/s that use the ITU-TS recommended physical and electrical interface specified in G.703; G.703 641K - likewise for transmission facilities running at 64Kbits/s; G.703 - the ITU-TS standard 1984 current version for the physical and logical traits of transmissions over digital circuits. G.703 now includes specifications for the US 1.544Mbit/s as well as the European 2.048Mbit/s, and circuits with larger bandwidths on both continents. G.703 is still generally used to refer to the standard for 2.048Mbit/s; G.821 - ITU-TS Recommendation that specifies performance criteria for digital circuits for ISDN.

Gateway Network interconnection device and software that operate at OSI Layer Seven. A gateway supports a full stack of the relevant protocol, such as SNA, DecNet, ISO, TCP/IP, and can covert to a non-seven layer protocol, such as async or BSC. It is typically used to provide access to wide area networks over asynchronous or X.25 links from a LAN environment. Examples include pads and protocol converters.

GFI (Group Format Identifier) (X.25 reference) the first four bits in a packet header, GFl contains the obit, Dbit and modulus value.

GGP (Gateway-to-Gateway Protocol) The protocol that core gateways use to exchange routing information, GGP implements a distributed shortest path routing computation.

Gosip (Government Open Systems Interconnect Profile) Country specific ISO OSI functional profiles that have been defined as part of national procurement policies. The US has US Gosip which is defined as a Federal Information Processing Standard. The UK has the UK Gosip which is defined by the Central Computer and Telecommunications Agency. The UK Government will not buy equipment unless it supports OSI as specified in Gosip. Unfortunately, the various Gosips as published by the UK, US and Japanese governments are all slightly different.

GSM (Global System for Mobile communications) A two-way, pan-European digital cellular system. Its specification is in line with ISDN and ITU-TS System 7 signaling and approved by almost all European countries. GSM operates at 900MHz and is a forerunner to the mass-market Personal Communications Networks, based on the same set of standards. GSM services include current digital subscriber services and the unique Short Message Service - a superior form of paging offering up to 160 alphanumeric characters with guaranteed delivery.

GUI (Graphical User Interface) Often pronounced "goo-ee", it describes the screen display that first greets a user and with which the user interacts during the computer session. A GUI makes use of Windows, Icons, Menus and Pointers - so-called WIMP systems - such as Microsoft Windows, GEM and X-Windows. Apple's GUI is particularly notable and be-spoke GUIs are becoming popular.

Group In the context of network security, a group is a set of users who share common permissions for one or more resources. Individually assigned user permissions take precedence over those assigned through groups.

Guard band The unused bandwidth separating channels to prevent crosstalk.

Guard tone A tone generated by a high speed modem dial-up modem to ensure that there is sufficient bandwidth available on the PSTN circuit for transmission.

H channel The ISDN packet switched channel on Basic Rate Interface, designed to carry user information streams at different speeds, depending on type: H11=1536Kbit/s, H0=384Kbit/s and H12= 1920Kbit/s.

H.261 A ITU-TS standard for video compression know as Codes for Audiovisual Services at N x 384Kbit/s. It sets a common algorithm for converting analog video signals to digital, operating at or above 384Kbit/s.

Half Duplex A two-way means of transmission, but data can only travel in one direction at a time.

Half-bridge Apple Computer terminology for a device linking LANs over a low-speed link such as a telephone line or X.25 link. It is termed a half- bridge as one is required at each end of the link.

Handshake Part of the procedure to set up a datacommunications link. The handshake can be part of the protocol itself or an introductory process: the computers wishing to talk to each other set out the conditions they can operate under. Sometimes, the handshake is just a warning that a communication is imminent.

Head-end A central point or hub in broadband networks that receives signals on one set frequency band and retransmits them or another. Every transmission has to go through the head-end in a broadband network. The head-end is the piece of hardware that enables a network to send and receive on the same piece of cable. In CATV technology, the head-end is the control center for a cable system where signals are processed and sent for distribution down the cable system.

Header The control information added to the beginning of a transmitted message. This may consist of packet or block address, destination, message number and routing instructions.

Hertz A measure of radio frequency. One Hz=one cycle per second. KHz = 1,000Hz; MHz =1,000,000Hz; GHz=1,000,000,000Hz.

HDLC (High level Data Link Control) An ISO standard set of protocols for carrying data over a link with error and flow control. Similar to IBM's SDLC, versions of HDLC are under development for multipoint lines. The ITU-TS later adapted HDLC for its Link Access Protocol used for X.25 networks. It is a bit-oriented data link control procedure under which all data transfer takes place in frames. Each frame ends with a frame check sequence for error detection. There is a control field at the start of each frame that allows error detection, and data link setup and data link termination. HDLC is, in fact, a misnomer, as it is not a high level protocol.

HLMS (Heterogeneous LAN Management Specification) Network management specification developed by IBM and 3Com. It provides an underlying structure for the development of network management products which can function with a variety of network operating systems and adapter cards.

Heterogeneous networking The networking of computers from different vendors, or the running of different operating systems.

Hierarchical network A network with one host at its hub, which is the major processing center, and one or more satellite processing units. IBM's SNA was originally designed as an hierarchical network.

Host processor A minicomputer or mainframe attached to a network providing services to network users.

Hot fixing The ability to detect and mark bad sectors of a disk, then assign alternate disk sectors during routine LAN operation. This automatically updates the original defect map.

Hot swap The ability of a device to have parts removed after, for example, a slot-in card or fan failure, without affecting its operation. Many of today's hubs offer hot swapping without bringing down the network to charge a component.

HPFS (High-Performance File System) An OS/2 file system that has faster input/output than the file allocation table file system; it does not restrict file naming to eight characters with a three character extension; and is compatible with the FAT file system.

HSSI (High speed Synchronous Serial Interface) Interface for transferring data to or from a WAN leased line (E2, E3), or to and from a LAN via an HSSI-capable DSU and HSSI router.

Hub The center of a star topology network or cabling system. A multi-node network topology that has a central multiplexer with many nodes feeding into and through the multiplexer or hub. The other nodes do not usually directly interconnect. LAN hubs are increasingly popular with the growth of structured cabling and the need for LAN management.

Brouter or Hybrid Bridge/Router Network interconnection device with a standard mode of operation as a router. It can support multiple OSI Layer Three routing protocols, but also supports bridging capabilities for network traffic that cannot be routed - non-routable traffic.

I-Series recommendations ITU-TS (formerly the CCITT) recommendations on B-ISDN, ISDN and aspects of ATM.

IAB (Internet Activities Board) The technical body that sets policy and standards for TCP/IP and the connected Internet suite of protocols. Its task forces are the IETF and IRTF.

IBMCS (IBM Cabling System) Cabling system introduced by IBM in 1984 based around Shielded Twisted Pair (STP) cabling.

ICMP (Internet Control Message Protocol) Protocol allowing status messages to be generated by an internetworking gateway, indicating a variety of error or unusual conditions.

ICR (Intelligent Character Recognition) Software similar to OCR, but it learns to recognize different fonts and character styles, so is more efficient. Useful to make paper documents accessible via groupware.

IDA (Integrated Digital Access) The means of providing digital access for subscribers to the BT ISDN service. Two versions: single line IDA and multi-line IDA are available.

IDAPI (Integrated Database Application Programming Interface) A standard from Borland and others providing a standard interface to a wide range of databases. It is supposed to be a superset of ODBC.

IEEE Institute of Electrical and Electronic Engineers: a US publishing and standards organization responsible for many LAN standards such as the 802 series.

IEEE 802.2 The Data Link standard for use with IEEE 802.3, 802.4 and 802.5 standards. It specifies how the basic data connection should be set up over the cable.

IEEE 802.3 The IEEE standardization of Ethernet. A Physical Layer definition that includes specification for physical cabling plus the method of transmitting data and controlling access to the cable. It uses the CSMA/CD access method on a bus topology LAN.

IEEE 802.4 The IEEE standardization of Token Bus. A Physical Layer standard that uses the Token Ring passing access method on a bus topology LAN. Used by LANs implementing the Manufacturing Automation Protocol. The older Arcnet operates in a similar way but does not follow 802.4, but Arcnet supporters have been trying to get the technology ratified by IEEE without success.

IEEE 802.5 The IEEE standardization of IBM Token Ring. A LAN Physical layer standard that uses the Token Ring passing access method on a ring topology LAN.

IEEE 802.6 The standard that defines Mans, an SMDS-based, short packet ATM transmission.

IGP (Interior Gateway Protocol) Term applied to any protocol used to propagate network reachability and routing information within an autonomous system. There is no single standard IGP, but RIP is one of the most common.

IN (Intelligent Network) A sophisticated network capable of recognizing the profile (authorization, chosen services) of its users or subscribers. Carriers offering advanced services will increasingly offer IN services, particularly as the boundary between fixed and mobile networks becomes blurred.

In-band signaling A communications technique in which the part of a signal bearing the data control information is contained within the bandwidth of the signal it is controlling.

Installable file system A file system that can be installed in place of the usual file allocation table file system, such as the high-performance file system.

Intelsat (International Telecommunications Satellite Organization) Founded in 1964 to develop a global satellite communications system, Intelsat has some 120 members, including- BT. The Earlybird (Intelsat 1) satellite was launched in 1965 as the first in a series. Intelsat VII is planned to replace the dozen aging Intelsat V-types which carry two-thirds of the world's telephone calls.

Interface The place or piece of equipment where interaction occurs between two regions or systems or processes. A common example is the RS232-C port or the AUI on LANs between the computer and transceiver.

Internet (1) A group of networks that are interconnected so that they appear to be one continuous network, and can be addressed seamlessly at the Network Layer Three of the OSI model. Typical internets are built using routers, either to form a backbone network comprised of routers, or to link together LANs at the Network Layer.

Internet (2) A collection of networks and gateways, including the Milnet and NSFNET, all using the TCP/IP protocol suite. It functions as a single, cooperative virtual network. The Internet provides universal connectivity and three levels of network services: connectionless packet delivery; full duplex stream delivery and application level services including electronic mail and EDI.

Internet address An IP address assigned in blocks of numbers to user organizations accessing the Internet by the US DoD's Network Information Center. Duplicate addresses can cause major headaches on the network, but the NIC trusts organizations to use individual addresses responsibly. Each address is an eight- bit number (0 to 255). There are three classes: A, B and C, depending on how many computers on the site are likely to be connected. The NIC is running out of Class B addresses, so work is in progress to extend the address format to allow expansion into the next century.

Inter-Process Communications (IPC) Communications between several programs based on one computer or across a number of machines. Alternatively, communication across a network between different processes of the same program between different computers running parts of a single program, or between two programs working together. The most common approaches to IPC in networking circles are probably Application Programming Interfaces such as APPC and NetBios.

IGP (Interior Gateway Protocol) The term applied to any protocol used to propagate network reachability and routing information within an autonomous system. There is no single standard IGP, but RIP is one of the most common.

ICMP (Internet Control Message Protocol) An integral part of the Internet Protocol (IP) that handles error and control messages. ICMP also includes an echo request/reply used to test whether a destination is reachable and responding.

IESG (Internet Engineering Steering Group) A US committee consisting of the IETF chairperson and the six area managers. The IESG coordinates activities among the IETF working groups.

IETF (Internet Engineering Task Force) A US committee concerned with short-term and medium-term problems with TCP/IP and the connected Internet. The IETF is divided into six areas which are further divided into working groups and its chairperson is a member of the TAB.

Interoperability The ability to exchange information between two systems so that the information can be processed meaningfully.

IP (Internet Protocol) The TCP/IP standard protocol that defines the IP datagram as the unit of information passed across an Internet and provides the basis for connectionless packet delivery service. IP includes the ICMP control and error message protocol as an integral part. It provides the functional equivalent of ISO OSI Network Services.

IP address The 32-bit address assigned to hosts that want to participate in a TCP/IP Internet. IP datagram - The basic unit of information passed across a TCP/IP Internet.

IRTF (Internet Research Task Force) A technical group working or problems related to TCP/IP and the connected Internet, The IRTF is divided into a set of research groups. The IRTF chairperson is a member of the IAB.

ISA (Industry Standard Architecture) The 8/16-bit bus architecture originally developed by IBM and now standard in almost all PCs that use Intel's 8086 and 80X86 chips.

ISDN (Integrated Services Digital Network) An approach to switched digital networking that can handle a range of digital voice, data and digital image transmission. It is intended to provide end-to-end, simultaneous handling of digitized voice and data traffic on the same digital links via integrated switches. There are different access channels, according to geographic location. These include Basic Rate Access (2 x 64Kbit/s 161 + 16Kbit/s) and Primary Rate Access (1.544 (US) and 2.048 Mbit/s (Europe). ISDN standards are defined in the ITU-TS's I-Series Recommendations.

LAN (Local Area Network) A communications system that links computers into a network, usually via a wiring-based cabling scheme. LANs connect PCs, workstations and servers together to allow users to communicate and share resources like hard disk storage and printers. Devices linked by a LAN may be on the same floor or within a building or campus. It is user-owned and does not run over leased lines, though a LAN may have gateways to the PSTN or other, private, networks.

LAN Manager for Unix Systems An implementation of LAN Manager for use with Unix. Known colloquially as LM/X.

LAN Manager LAN Network Operating System developed by Microsoft which runs or top of OS/2 and recently Windows NT.

LAN Network Manager IBM's network management software for Token Ring networks.

LAN Segment A part of a LAN that is separated from the rest by one or more bridges.

LAN Server IBM version of IBM LAN Manager.

LAP (Link Access Protocol) The Data Link or OSI Layer Two protocol specified by the ITU-TS for the X.25 interface standard.

LAP-B: A link set-up routine to establish and maintain links between DCE and DTE. Lap-D: The Layer Two protocol for an ISDN D-channel specified in the ITU-TS recommendation Q.921. It is a framed, bit-oriented protocol similar to Lap and Lap-B protocols specified for X.25 circuits. Lap-M: A variation of Lap-B used in the V.42 modem error control standard.

LAT (Local Area Transport protocol) A DecNet specific protocol for the exchange of small packets of data between typically Dec Vax host computers and terminal servers in a LAN.

Layer Description of divisions in specifications Such as OSI and SNA communications protocols. Functions are grouped together that comprise one step in the hierarchy necessary for successful data communications.

Line conditioning A collection of techniques for keeping the quality of transmissions within specified parameters.

Line driver A signal converter that conditions the digital signal transmitted by an RS232 interface to extend reliable communication beyond the 50ft RS232 limit up to several miles. It is a baseband transmission device.

Line turnaround The interval on a transmission link between the time one block of data was sent and received and the time the next one can be sent.

Line signal standards In the US: T1 carries data at 1.544Mbit/s and has 24 voice circuits; T1C 3.152 Mbit/s with 48 voice circuits; T2 6.312Mit/s with 96 voice circuits, and T3, 44.736Mbit/s. In Europe, the standards are of the form En. E1 line speed is 2.048Mbit/s with 3C voice circuits; E2 is 8.448Mit/s with 120 voice channels. E3 is 34.368Mbit/s with 480 voice circuits. In the UK, E1 is often referred to as MegaStream, a BT label for its 2Mbit/s leased circuits.

Link state algorithm A routing algorithm such as OSPF which takes into account lowest delay when choosing a route: link speed and congestion as well as hop count.

LLC (Logical Link Control) A data link protocol based or HDLC, developed for LANs by the IEEE 802 Committee and consequently common to all LAN standards for Data Link OSI Layer Two transmission.

Local bridge Bridge that links two local LANs: in the same building, for example.

Local security A security method available for 386 and 486 servers running HPFS386. This method extends LAN Manager security measures to protect the files on a server by restricting access of the users working at the server. With local security, a user must be assigned permissions to access any file or directory in an HPFS386 partition, whether or not the resource is shared as part of a LAN Manager resource.

LocalTalk An Apple cabling scheme underlying its low-cost LANs. A 230Kbit/s baseband network primarily for Mac computers and LaserWriter printers, it uses the CSMA/CA media access method. The current Phase II allows theoretically unlimited networks.

Lobe The cable between a Token Ring station and the Trunk Coupling Unit to which it is connected. Lobe length comprises a patch cable from the TCU to the main wiring panel, the length of the main wiring to the user station's location, then a patch cable from a floor/desk socket to the station.

Logon script A batch program containing LAN Manager, NetWare and other operating system commands used to configure workstations. Logon scripts can be written for one or more users.

Logon server For a domain, a logon server is the primary domain controller and the backup domain controllers. For a user, the server that processes the user's logon request.

Loopback A diagnostic test that returns the transmitted signal back to the sending device after it has passed through a network or across a particular link. The returned signal can then be compared to the transmitted one. The discrepancy between the two help to trace the fault. When trying to locate a faulty piece of equipment, loopbacks will be repeated, eliminating satisfactory machines until the problem is found.

LU (Logical Unit) An IBM SNA network function defined in layers four, five and six (Transmission Control, Data Flow Control and Presentation Services) of the SNA architecture. In the SNA network, corresponding LUs are able to exchange information. Originally, particular types of LU has specific functions: LU1 for printers, LU2 for displays and so on. As SNA has developed, new types of LU have been introduced that support a broader range of communications facilities. LUs are normally associated with particular Physical Units (PUs), or network devices. LUs provide the services required by Applications (APs) in the IBM SNA environment, sitting between the APs and the PUs. A Dependent LU relies on the host for activation, physically and logically, while an Independent LU can initiate a session without host involvement.

LU 6.2 An IBM SNA Logical Unit that provides general communications functions, including the communications functions necessary for peer-to-peer networking. Underlying LU 6.2 is a type of node: node type 2.1, which facilitates peer-to-peer networking. Two SNA units which implement rode type 2.1 can set up a full peer-to-peer session without invoking SSCP capabilities in a host processor. This is defined as SNA Low-Entry Networking.

MAC (Media Access Control) layer A sub-layer of the Data Link Layer (Level Two) of the ISO OSI Model responsible for media control.

MAN (Metropolitan Area Network) A high speed network designed to link together sites in a metropolitan or campus area. The IEEE has defined its 802.6 standard for MANs based on the Distributed Queue Dual Bus technology.

MAP (Manufacturing Automation Protocol) An ISO OSI protocol stack that is defined as a functional profile. Originally developed by General Motors for use in factory floor manufacturing environments it is based around the IEEE 802.4 Token Bus LAN technology.

MAPI (Messaging Application Programming Interface) Microsoft's standard for the applications interface to e-mail. See also VIM.

Master station A device that controls/polls the nodes in multipoint circuits or, in point-to-point circuits, the unit that controls the slave station. In LAN terms, the device on a token passing ring that enables recovery from error conditions, such as lost, busy or duplicate tokens, usually by generating a new token. Servers are sometimes called master stations.

Matrix switching A form of data switching at the heart of ATM, enabling the appropriate of bandwidth to be available end-to-end for the duration of the session without contention.

MAU (Medium Attachment Unit) A transceiver (transmitter/receiver) comprising hardware circuitry that provides the correct electrical or optical connection between the computer and IEEE 802.3 LAN media. Since MAUs typically support only one type of network medium, a choice of MAUs is available to support different media. It detects carrier and collision activity, passing the information to the Computer. It can be a standalone unit or incorporated in a circuit board inside the computer,

Media Access Control driver A LAN device driver that works directly with the network adapter cards, acting as an intermediary between the transport driver and the hardware.

Medium The physical method or equipment used for transmission, from a tangible fiber optic or copper cable to a satellite link. Alternatively, a little old lady in a shawl used as a last resort to retrieve lost data.

Medium Interface Controller (MIC) An hermaphrodite connector on IBM patch parels specified in 802.5.

Member server A server in a domain that keeps and uses a copy of the domain's user accounts database but does rot validate logon requests.

Monolithic driver A network device driver that acts as network adapter card driver and transport protocol driver combined.

MMS (Message Handling System) The engine underlying an electronic messaging system. Also a proprietary e-mail integration scheme developed by Action Technologies.

MIB (Management Information Base) The set of variables or database that a gateway running CMOT, SNMP, or CMIP network management protocols maintains. It defines variables needed by the SNMP protocol to monitor and control components in a network. Managers fetch or store into these variables. MIB-II refers to an extended SNMP management database that contains variables not shared by both CMOT and SNMP. The CMIP and SNMP MIB formats differ in structure and complexity.

Microchannel A proprietary bus developed by IBM for its PS/2 computer family's internal expansion cards. It offers improved performance over ISA-based machines like the IBM PC/AT.

Microwave Wireless transmission at very high frequency to deliver telecommunications services, including TV distribution, between two points. It is dependent on line of sight.

MTBF (Mean Time Between Failures) A term describing the reliability of equipment established by testing kit to its limits and promoting the MTBF as a selling point. As most users know, however, laboratory and operational environments have little in common.

MNP (Microcom Networking Protocol) A series of protocols designed by Microcom to support error control and data compression for asynchronous modem transmission.

Milnet (Military Network) Originally part of Arpanet, Milnet was partitioned in 1984 to make it possible for military installations to have a reliable network service while the Arpanet continued to be used for research. Under normal circumstances, Milnet is part of the Internet.

Modem A device named from an amalgam of the words modulator and demodulator. A modem will modulate an outgoing binary bit stream or to an analog carrier, and demodulate an incoming binary bit stream from an analog carrier.

Modem Approvals Group Established in January 1993 to raise awareness of the UK law against connecting unapproved devices to the PSTN, this group demands a level playing field in the modem market with better law enforcement or more open standards.

Modem eliminator A device that can replace a modem in some instances when the distance to be covered is short. It takes the power it needs to operate from the transmission line.

MSAU (Multi-Station Access Unit) A wiring concentrator on a Token Ring network that allows devices, typically eight to 12 Token Ring stations, to be connected to the ring. Relays in the MAU ensure the integrity of the network when devices are attached or removed. A Managed Multi-Port/Multi-Station Access Unit has built-in network management support.

MS-Net Microsoft DOS-based networking system software product (Microsoft Network).