In 1985, the IEEE published a set of four physical and data link layer standards, termed the 802 protocols: 802.2, 802.3 (802.3u), 802.5 and 802.12. The 802.2 is a LLC sub-layer that is concomitantly used with the 802.3 and 802.5 standards. It provides error and flow control and a standard network service that is independent of the other layers. The other 802 standards are different implementations of the MAC component of the data link layer. The 802.3 (and the fast Ethernet 802.3u specification) protocol defines an Ethernet standard that is able to use a variety of network configurations using the CSMA/CD access method. 802.5 employs a token passing access method and can support data speeds of 4 or 16Mbps. In this implementation, the physical network topology can vary although the logical topology is always that of a ring. The 802.12 is based upon a new standard termed 100VG-AnyLAN which uses a contention-based, demand priority arrangement for network access. See also: 10Base2 (collective entry), Ethernet, Logical Link Control, MAC Addressing, Network Topologies and Token Passing
10Base2, 10Base5, 10BaseT, 100BaseT, 100BaseTX
The Ethernet-type networks can be subdivided into several types of network. The IEEE802.3 standard uses the following notation to indicate Ethernet type: x-Base-y, where ‘x’ indicates the data rate (in Mbps), ‘Base’ denotes that baseband transmission is used and ‘y’ either describes the maximum media distance or the cable type. Using this type of notation, the letter ‘T’ shows that twisted pair cabling is used, whereas ‘F’ indicates fiber optic cabling. The numbers 2 and 5 state that the maximum media distances are 200m and 500m respectively. The designation TX (or FX in some cases) highlights that the network is a fast Ethernet network. See also: 802 Protocols, Baseband and Ethernet.
An active directory is the folder that is currently being accessed by a computer
API (Application Programming Interface)
The application layer is the layer at which many TCP/IP services can be run, e.g. FTP, HTTP and SMTP. APIs provide a user with an interface with which to access these services – common examples include the ‘sockets’ program and NetBIOS. See also: HTTP, TCP/IP and NetBIOS.
Although AppleTalk began life in the early 80s as a strictly Apple-specific network protocol, it is portable between different operating systems. This is a cheap solution to networking Apple computers (since they have the technology already built-in) that provides a system that is data-link layer independent and is able to support several networking systems such as FDDI, Ethernet and Token-ring arrangements. In its current state, Phase II, it is capable of supporting up to 16 million nodes (each is assigned an address dynamically) – this is not compatible with older revisions however. This technology however, despite its advantages, is not as popular as the Ethernet and Token-ring architectures. See also: Token-ring Networks.
ARCnet (Attached Resource Computer Network)
ARCnet was developed in the late 70s by the Datapoint Corporation as a method of networking several PC-based systems into small, robust networks using a star or bus physical topology. ARCnet is cheap, simple to implement and has internal automatic configuration technologies to allow it to overcome cable breaks and other network faults. Although Ethernet approaches initially drew attention away from this kind of system, the launch of ARCnet Plus has brought with it increases in bandwidth and allows many more nodes to be attached. See also: Asynchronous Transfer Mode.
ARP (Address Resolution Protocol)
When two systems communicate, an IP address is used to identify the destination machine. However, since data is transmitted on the physical and data link levels. The address resolution protocol performs the task of resolving an IP address to a hardware address. See also: IP Address.
ARPANET was the original prototype network to use the TCP/IP protocol suite. Work began in the late 60s and continued through to the early 70s by the US Department of Defense research group, the Advanced Projects Research Agency (ARPA). ARPANET was first operational in 1972 and was later superceded by MILNET and then NSFNET (National Science Foundation Network), which consisted of a backbone of six supercomputers to which several local networks were subsequently linked. See also: TCP/IP and World Wide Web.
Asynchronous Transfer Mode
This emerging technology offers a highly versatile transmission mode for both LAN and WAN systems. It can support a maximum bandwidth of 622Mbps and has consequently been included as an Internet backbone by many Internet Service Providers. Other slower formats are available, such as the 155Mbps and 25Mbps in order to accommodate Ethernet transmission. Video, audio and data are all supported, with mechanisms for prioritising time-sensitive material. ATM divides information into 53-byte cells containing 48 bytes of data and 5 bytes of header data. This consistency of data size allows for rapid processing and bandwidth can be readily determined. See also: Frame Assembly/Disassembly.
ATM (Asynchronous Transfer Mode) – see also the section in N+ Essentials
This emerging technology offers a highly versatile transmission mode for both LAN and WAN systems. It can support a maximum bandwidth of 622Mbps and has consequently been included as an Internet backbone by many Internet Service Providers. Other slower formats are available, such as the 155Mbps and 25Mbps in order to accommodate Ethernet transmission. Video, audio and data are all supported, with mechanisms for prioritizing time-sensitive material. ATM divides information into 53-byte cells containing 48 bytes of data and 5 bytes of header data. This consistency of data size allows for rapid processing and bandwidth can be readily determined. See also: Bandwidth and Ethernet.
A backbone is a (usually) fast link that connects the various segments of a network. See also: Segment.
Backup Domain Controller (BDC)
A BDC is a type of NT server that contains a copy of the domain security database that is used to authenticate users. However, it is not possible to update details that are stored within a BDC – and the role is generally that of fault tolerance in the event that a primary domain controller should fail. See also: NT Server and Primary Domain Controller.
The bandwidth of a particular medium indicates the amount of data that can be transferred through the medium at a given time. Ethernet connections, for example, usually have a bandwidth of 10 or 100 megabits per second. Bandwidth is also known as the ‘capacity’ or ‘transmission speed’ of a medium. The standard speeds for data transmission are generally in the megabits per second range, although newer gigabit systems are emerging. It should be noted however that speed comes at a price – the faster a transmission speed, the poorer the representation of the data at the receiving end due to cable characteristics and noise. See also: Baseband and Broadband.
Baseband transmission uses the whole of the media bandwidth as a single transmission path. Digital signal networks commonly use this type of transmission. This particular variation upon bandwidth usage is more reliable than the broadband method. See also: Bandwidth and Broadband.
Beaconing is a system of fault detection that is inherent within FDDI and token ring networks. Using this technology, each workstation attached to a network acts as a fault detector and upon perceiving a fault begins to send a unique beacon signal across the network. This continues until the upstream workstation begins to beacon, whereupon the prior workstation stops signalling. This continues until only one workstation is left signalling – the machine immediately downstream of this node is associated with the fault and can be repaired. See also: Token-ring Networks.
A bridge provides a communications link between two or more network segments to form one logical network. Bridging can be advantageous in circumstances where particular sections of the network have to carry a lot of data traffic. Communication between the segments occurs via the bridge although data that does not need to access a workstation on the other side of the bridge is not forwarded. See also: Network Topologies, Routers, Gateways and Switching.
Broadband transmission divides the available media bandwidth into a number of transmission paths. Analog signals generally use this form of transmission. Each channel uses a different range of frequencies. See also: Baseband and Bandwidth.
BNC (British Naval Connector/Bayonne-Neill-Concelman) Connectors
These are twist and lock connectors that are used with 10Base2 network cabling. See also: 10Base2.
A bus topology is a linear network with all nodes attached directly to the main cable (often known as a backbone). The ends of the bus must be terminated so that the signal is absorbed once it has passed all of the connected devices. Signal transmission normally occurs in both directions from the source and this type of topology is easy to install and requires little cable compared to the other topologies. See also: Network Topologies.
CAN (Campus Area Network)
A Campus Area Network is limited in scope to a single geographical location, but this may exceed the size normally defined for a LAN. See also: LAN, MAN and WAN.
Category 3 and category 5 are the most common types of UTP cabling used in network arrangements. They differ with respect to transmission frequency and data capacity – cat5 is a higher cable specification with a greater number of twists per meter and higher-grade insulation. Category 3 has a frequency of 16MHz and a capacity of 10Mbps compared to the 100MHz and 100Mbps that are available using category 5 cable.
CDDI (Copper Distributed Data Interface)
A CDDI network is essentially founded on the same principles as the FDDI arrangement. The two approaches differ however since, whereas the FDDI network uses fiber optic cable, the CDDI uses twisted-pair cabling. It supports the same bandwidth as FDDI, but it cannot be used over the same distance. See also: FDDI, TP cable and Ring Topology.
A client is the term given to a computer that is present on a network that allows users to requested a shared resource contained on or within that network. Another popular term for client is workstation. See also: Host.
In this network system, the high cost of a mainframe computer is removed whilst the ability to support large numbers of simultaneous users is maintained. The client is the user interface that can display and edit data, it is the client that generates data requests that are sent across the network to the structured query language (SQL) server. This server processes the data and transmits only the result back to the interface where the data can then be displayed. In this approach, processing operations are shared between server and client – only queries and results are transferred rather than the data itself (as in a file-server system) – this is ideal in situations where several users require access to a large centralised database.
Coaxial (Coax) Cable
This type of cable is formed from two separate conductors that share a common axis (hence the term co-axial). The outer conductor, a wire mesh, is isolated from the inner conductor, a copper wire, by plastic insulation - electromagnetic interference is limited, in part, by the wire mesh. Two categories of coax cable are available: thick and thin. These are classified using the radio grade rating scale. Twisted-pair cabling has now largely superceded coaxial cable. The technology that underlies the use of coaxial cable is highly developed and can support greater bandwidths than TP cables. Installation is simple, although coax cable can be sensitive to EMI and reconfiguration of an installed coax network is potentially difficult. See also: TP cables, UTP cables, Ethernet and Bandwidth.
Connectionless communication occurs within or between networks that are not physically linked by network cables. Normally, a ‘connectionless’ link is made via radio transmission, or occasionally through infrared signaling. See also: Connection-orientated Communication.
Connection-orientated communication is the opposite of connectionless communication. This type of link is present where two or more network components are joined by a physical link such as a network cable or telephone line. See also: Connectionless Communication.
In a contention-based system, each network device competes with the other connected devices for use of the transmission media. Contention based systems require a set of protocols that reduce the possibility of data collisions, since if the devices compete and simultaneously send data packets, neither packet will reach its’ intended destination. Contention based systems commonly use the CSMA protocols to overcome this problem. See also: CSMA Protocols and Token Passing.
CSMA (Carrier Sense Multiple Access) Protocols
The CSMA protocols allow contention-based networks to successfully communicate by detecting activity on the network media (i.e. ‘carrier sense’) and reacting to this – e.g. if the medium is busy, the device that is attempting to transmit must wait until the media is clear. ‘Multiple Access’ refers to the use of the network by multiple devices and the way in which this particular set of protocols prevents the simultaneous detection of a ‘clear’ medium by several devices at once. Two variations of the CSMA protocol are available; CSMA/CD and CSMA/CA. The first of these forms – Collision Detection – recognizes a signal collision on the basis of electrical fluctuations produced when signals combine. The CA (collision avoidance) form waits for the network to be free before announcing their intention to transmit. If no negative response is received, the unit will transmit. See also: Contention and Token Passing.
The C2 security level was developed by the US National Computer Security Center to define a series of security standards that contain several key features. For networked systems, the C2 grade is required by the US government – both the Microsoft Windows NT and Novell NetWare 4.x systems have achieved this standard. See also: PGP, Firewall and Proxy Server.
CSLIP (Compressed Serial Line Internet Protocol)
Compressed SLIP is an advancement of the SLIP protocol that increases the data flow across serial communications networks by removing repeated header information. However, similarly to the SLIP protocol, there is no support for error checking or correction. See also: PPP and SLIP.
CSU/DSU (Channel Service Unit/Data Service Unit)
The CSU or DSU is used within digital network circuits and it performs an analogous role to the modems used in analogue networks. They convert digital signals from the computer into a digital signal that is more appropriate for WAN environments. See also: Modem.
Current State Encoding
Current state encoding is a form of digital encoding that converts the binary data into representative electromagnetic signals. In binary, data exists as either a 0 or 1 – if you represent this in terms of ‘current state’, ‘1’ can be signified as a positive current and ‘0’ as a negative current. See also: State Transition Encoding.
Cyclical Redundancy Checking (CRC)
A CRC algorithm treats a block of transmitted data as a single large binary number and divides this by a 16- or 32-bit number (called the polynomial). The remainder of this division is termed the checksum. This is transmitted with the data and is compared to the checksum generated by the receiving modem. If the two cyclic redundancy checks are not the same, the data block is rejected and a request for data re-transmission is sent to the source.
DHCP (Dynamic Host Configuration Protocol)
The dynamic host configuration protocol provides and automatic method for allocating IP addresses, subnet masks and optional parameters such as the default gateway, domain name server (DNS) address or NetBIOS name server address. Ensuring each machine is configured with a unique IP address can become a very tedious responsibility for the administrator of a large network. Errors are disastrous for the network users as only the first machine holding the duplicate IP address can connect to the network. DHCP removes many of these configuration problems. See also: WINS and TCP/IP Protocols.
Digital Data Service
This system was created in the late Seventies to allow fast data to be sent in a more rapid and more secure manner than was available through the analogue networks at that time. Typical bandwidths that were available from the telecommunications industry were in the rate of 2.4 to 56Kbps full duplex.
DIX (Digital Intel Xerox) Connector
These connectors are used to connect a cable to a backbone cable. A ‘vampire tap’ is placed around the backbone (e.g. coaxial cable) and the DIX connector links this tap to the cable from the node.
This type of spread spectrum transmission involves sending data as several small ‘chips’ transmitted at different radio frequencies, although the user must be aware of these frequency alterations in order to interpret the data. See also: Frequency Hopping, Spread-Spectrum Transmission, Radio Frequencies and Wireless Bridges.
DMA (Direct Memory Access)
Direct Memory Access is a mechanism that allows hardware devices to transfer data to and from system memory directly, without the need for processor intervention. Typically, this method is employed by hard disk controllers, network cards and tape streamers. In more modern adapters, memory sharing has superceded direct memory access as a method for increasing memory access times.
DNS (Domain Name Service)
The DNS is a distributed hierarchical system for resolving names to IP addresses. It uses a distributed database system that contains information on domains and hosts within those domains. Several name servers each hold part of the database, this allows for the delegation of maintenance and it provides integral backup in the case of link failures. At the top of the DNS hierarchy is the ‘root’ – often represented by a period. Immediately below the root lie the top-level domains. In the US, examples of top-level domains include ‘com’ for a commercial organization and ‘edu’ for an educational establishment. Top level domains are categorized by ISO country codes such as ‘uk’, ‘nl’ and ‘de’ – indicating the United Kingdom, the Netherlands and Germany respectively. Domain information is located by tracking records from the root down through the hierarchy. No name server has complete information about all of the domains, rather each domain points to the information relating to the appropriate sub-domain.
DNS (Domain Name Service) Server
The DNS is a distributed hierarchical system for resolving names to IP addresses. It uses a distributed database system that contains information on domains and hosts within those domains. Several name servers each hold part of the database, this allows for the delegation of maintenance and it provides integral backup in the case of link failures. At the top of the DNS hierarchy is the ‘root’ – often represented by a period. Immediately below the root lie the top-level domains. In the US, examples of top-level domains include ‘com’ for a commercial organization and ‘edu’ for an educational establishment. Top level domains are categorized by ISO country codes such as ‘uk’, ‘nl’ and ‘de’ – indicating the United Kingdom, the Netherlands and Germany respectively. Domain information is located by tracking records from the root down through the hierarchy. No name server has complete information about all of the domains, rather each domain points to the information relating to the appropriate sub-domain. See also: DHCP Protocol, IP Addresses and WINS.
DSx (or T-Carrier)
The foundation level for digital communications is the DS1 (or T1) digital signal circuit that comprises 24 channels (each called a DS0 or KiloStream link) of 64Kbps bandwidth multiplexed into a single digital connection. This connection can be used for voice, video and data signals. Bell Laboratories developed the T-carrier system to allow multiple calls to be conducted on a single cable. T1 lines can be further multiplexed to provide more bandwidth and ratings are numerically increased to account for this, i.e. T1, T2, T3, T4 – in Europe these are termed E1 through to E4. See also: ISDN and DSU/CSU.
Dual-Attached Station (DAS)
In a dual ring network topology, some of the workstations are attached to the primary ring only. Some workstations however are linked to both rings – these dual-attached stations are able to re-establish a connection in the event of a failure simply by switching between the rings. See also: Network Topologies and Ring Topology.
Duplex refers to the ability of a modem to support the simultaneous reception and transmission of data. Previous to the development of this technology, modems would take turns to function in send and receive modes – this is known as half-duplex.
Duplexing is a method of providing fault tolerance within network systems. Duplexing is a mirrored pair of hard disks (with a single disk controller) with an additional disk controller on the second drive. This reduces channel traffic and potentially improves performance. Duplexing is intended to protect against controller failures as well as media failures. See also: Mirroring and Striping.
EEPROM (Erasable Electronically Programmable Read Only Memory)
Erasable EPROMs are similar to EPROM chips except that they can be erased in situ. Many motherboard manufacturers use EEPROM instead of EPROM chips as they can be upgraded using computer programs. In practice however, EEPROM chips can only undergo a limited number of reprogramming cycles and so it is wise to avoid unnecessary updating wherever possible. See also: EPROM.
EISA (Enhanced Industry Standard Architecture)
The Enhanced ISA standard was developed as an extension the original ISA format. EISA provides support for both 16- and 32-bit communication and fast data transfer rates. ISA cards are compatible with this newer standard and ISA cards can be used with EISA bus slots on a computer. See also: PCMCIA, ISA, MCA and PCI.
Electronic Mail (Email)
Electronic mail is the term given to the transfer of documents or similar data between users on the same or different networks. Several protocols have been developed to allow a high degree of flexibility that allows users to download and send email at their convenience, to attach data files to documents (through the use of attachments) and to allow a fast and (reasonably) secure method of communication.
EMI (Electro-magnetic Interference)
Electro-magnetic interference occurs when the electrical signals from one device affect the operation of another electrical device. Fluorescent lights or electric fans can, for example, affect the visual display of computers and may also affect data transmission. EMI does not cause permanent damage to the device and can be eliminated by removing the source of the interference.
Encapsulation refers to the containment of one network protocol within another protocol. Commonly encapsulation is seen in tunneling protocols. See also: Tunneling.
Encryption refers to the encoding of a file or data set so that it is incomprehensible to unauthorized users. Commonly encryption may take the form of a specific key, which is applied in a regular manner to the data within a file thus making the file unreadable. This key may be sent within the file and the intended user can access the data be entering a password. In terms of networks, encryption is often ‘silent’ between source and destination users but prevents any external user from accessing the data during transmission. Currently the ‘pretty good privacy’ encryption system is the de facto standard for email security. See also: Decryption.
EPROM (Electronically Programmable Read Only Memory)
EPROM chips are data-containing chips that are commonly used for the storage of system programs such as the basic input/output system that is required by the computer during startup procedures. Hardware manufacturers use an electronic programming device to program these chips, whilst exposing the chips to ultraviolet light can erase information. In practice, most EPROM chips are not erased, but are instead replaced by newer chips when required. See also: EEPROM.
Ethernet is the most popular type of LAN technology. It is capable of supporting a wide variety of cable options and is based upon inexpensive equipment. It was first created in the 60s at the University of Hawaii and commercially developed in the 1970s by the Xerox Corporation. Ethernet is frame-based and employs a logical bus topology, baseband signalling and the CSMA/CD method for network access. It can handle a bandwidth of 10Mbps and the maximum frame size is 1, 518 bytes. The Ethernet is described by the 802.3 specification standard that is further divided into four categories: 10BaseT, 10Base2, 10Base5 and 10BaseF. In order to use an Ethernet network, an Ethernet compatible network card is required. See also: 802 Protocols, Network Interface Card, CSMA Protocol and VGAnyLAN.
FDDI (Fiber Distributed Data Interface)
FDDI is a high-bandwidth, token ring network that has been developed (by the American National Standards Committee) for use in MAN-like networks. Fiber optic cables within the network permit a maximum ring length of 100km, although repeaters are required every 2km. This expensive technology is used mainly within high-speed backbones, for high-capacity office networks and for back-end networks that connect servers. Token ring schemes are used to access the system and it differs from the 802.5 framework in terms of transmission and the fact that it uses ‘early token release’. FDDI is based around a dual counter-rotating ring topology and this provides an inherent backup against network failure. The physical layout of the network may by either a star or ring topology, although the logical topology is a ring. See also: CDDI, Fiber Optic cable, Repeaters and Ring Topologies.
Fibre Distributed Data Interface
FDDI is a high-bandwidth, token ring network that has been developed (by the American National Standards Committee) for use in MAN-like networks. Fibre optic cables within the network permit a maximum ring length of 100km, although repeaters are required every 2km. This expensive technology is used mainly within high-speed backbones, for high-capacity office networks and for back-end networks that connect servers. Token ring schemes are used to access the system and it differs from the 802.5 framework in terms of transmission and the fact that it uses ‘early token release’. FDDI is based around a dual counter-rotating ring topology and this provides an inherent backup against network failure. The physical layout of the network may by either a star or ring topology, although the logical topology is a ring.
Fiber Optic Cable
Fiber optic cable employs light signals as the basis for data transmission as opposed to the electrical signals that are used by the other main cable types. The light ‘pulses’ are produced by light emitting diodes or by laser diodes – these pulses travel down the glass core of the fiber (known as the waveguide) and the cladding that surrounds this core reflects light back to ensure transmission efficiency. Two main categories of fiber are available; monomode, which uses a single light path through the waveguide, or multimode, which as its name suggests, can allow multiple pathways. At the receiving end of the cable, light-sensitive diodes re-convert the light pulse into an electrical signal. Fiber optic networks are expensive to install and configure although rewards in terms of high bandwidths and immunity to both interference and eavesdropping have justified this expenditure in many cases. Repeaters are required at 2km intervals to boost the transmission signal. See also: Twisted-Pair Cable, UTP cable and STP Cable.
A firewall is a hardware or software filter that is set up on a network that can filter data packets as they pass through a system. Filtering can take place on the following parameters: direction, IP address (either an address or a range of addresses) and using a well-known port number. Firewalls can be designed to accept all packets except those on a ‘reject list’ or conversely to reject all packets except those on an ‘accept list’. This latter arrangement is normally easier to configure. See also: Internet, Proxy Server and IP Address.
A frame is the basic ‘unit’ of data that is transmitted on a network. Frames contain several constant components – the source and target addresses as well as the data and error checking regions. Start and stop signals signify the beginning and the end of the frame respectively. It is the role of the network interface card to construct and understand frame structures. See also: Network Interface Cards and MAC Addressing.
Frame relay evolved for the X.25 protocol and as such, it is based upon the transmission of variable-length packets over a packet-switched network. Frame relay draws upon the advantages of modern digital connections to provide a high degree of reliability and error-control. ‘Permanent vital circuits’ are used to avoid fragmentation and the need for switching – leading to vastly reduced overheads placed upon the network. Frame relay approaches are cheap and can provide high bandwidth transmission – up to 1.544 Mbps in some cases. See also: X.25 Protocol.
Frequency hoping is one format for spread-spectrum transmission. This system rapidly switches from one radio frequency to another and remains at the frequency for a specified length of time – both the sender and receiver must be aligned so that transmission occurs at the same frequency. These systems are generally low range, inexpensive and can allow increases in bandwidth if multiple frequencies are used for simultaneous transmission. See also: Direct Modulation, Radio Frequencies and Spread-Spectrum Transmission.
FTP (File Transfer Protocol)
Due to the enormous possibilities of information storage and retrieval on the Internet, a common protocol was developed to allow users to send and receive data and other types of computer files across standard telephone exchanges. See also: URL, HTTP, HTML and World Wide Web.
FQDN (Fully Qualified Domain Name)
A Fully Qualified Domain Name provides a solution to the problem of host name uniqueness on the Internet. FQDNs consist of a host name and a domain name. An example of a FQDN could be mole.riverbank.co.uk. In this example, ‘mole’ is the host name, and ‘riverbank.co.uk’ is the domain name. The domain name is appended to the host name to define the location of the host and ensure that it is unique. Domain names are registered at a central site to ensure that they are not duplicated and once they have been logged, another user cannot use them. Each FQDN is held together with its corresponding IP address on a DNS server. See also: IP Address, DNS, DNS Hierarchy and World Wide Web.
A gateway is a computer or other device that acts as a translator between two completely dissimilar computer systems. For example, a connection from a PC LAN to an IBM mainframe would require a gateway – this provides translation between the networks. Gateways tend to be slower than bridges or routers. Occasionally, the term gateway (or ‘default gateway’) is used to denote a router on a network that uses the TCP/IP protocol, this however is not a true gateway as described above, since both systems use the common TCP/IP protocol to facilitate communication. See also: Bridging, Routing and TCP/IP Protocol Suite.
GroupWare is a term that describes software that can be jointly used by several individuals connected to a network – examples include scheduling programs, video/voice conferencing packages, library services and various project management systems. Examples of GroupWare are Microsoft Exchange, Novell GroupWise and Lotus Notes.
When data is transferred through a modem to modem connection, each modem must be able to contact the other so as to establish a rate for data flow. Handshaking allows modems to communicate in such as way so that the receiving modem can tell the source modem to ‘send’, ‘stop’, ‘pause transmission’, etc. Hardware and software methods exist to provide handshaking, with each playing a slightly different role in the regulation of transmission.
In TCP/IP networking terminology, a ‘host’ is a device that can directly communicate on a network. In this sense it is similar to a node. See also: Node.
HTML (Hypertext Markup Language)
This is the standard, structured language used for creating documents for publication on the World Wide Web. HTML was developed and continues to be updated by the World Wide Web Consortium. HTML incorporates special (non-displayed) commands, called tags that tell browser programs how to display text, graphics and how to respond to user input amongst many other functions. Specialized HTML document engines have now been produced to make HTML document preparation as simple as possible and they require little or no prior knowledge of HTML tags and construction. See also: HTTP and FTP.
HTTP (Hypertext Transfer Protocol)
This protocol provides a mechanism for the transmission of hypertext documents across the Internet so that they may be displayed using an Internet browser such as ‘Netscape’ or ‘Internet Explorer’. See also: FTP.
Hubs may also be known as ‘multiport repeaters’ or concentrators. They are the central points of connection for segments and act like repeaters so that every segment receives signals sent from any other segment. See also: Hybrid Topology and Segment.
Hybrid networks usually combine the advantages of two different network topologies. Common hybrid networks are the star-bus and star-ring configurations. See also: Bus Topology, Ring Topology and Star Topology.
IAB (Internet Activities Board)
The IAB is the technical sub-division of the ISOC. They are responsible for setting Internet standards and publishing these as RFC documents. The IAB, in turn, controls the Internet Research Task Force, the Internet Engineering Task Force and the Internet Assigned Numbers Authority. See also: ISOC.
ICMP (Internet Control Management Protocol)
This protocol is used to report errors and send messages about the fate and delivery of data packets. It may also be used to test the function of TCP/IP networks. Common examples of ICMP returned messages are ‘Destination unreachable’, which is sent when a router cannot locate the destination and ‘Time exceeded’, which signifies that the ‘time to live’ (TTL) packet has reached zero and the connection has timed out. As the TTL packet crosses a router, its initial value of 255 is reduced by one. If the packet is looping, due to a corrupted routing table, the TTL is eventually reduced to zero and an error is returned to the source host.
IEEE (Institute of Electrical and Electronic Engineers) 802.x Standards
In 1985, the IEEE published a set of four physical and data link layer standards termed the 802 protocols: 802.2, 802.3 (802.3u), 802.5 and 802.12. The 802.2 is a LLC sub-layer that is concomitantly used with the 802.3 and 802.5 standards. It provides error and flow control and a standard network service that is independent of the other layers. The other 802 standards are different implementations of the MAC component of the data link layer. The 802.3 (and the fast Ethernet 802.3u specification) protocol defines an Ethernet standard that is able to use a variety of network configurations using the CSMA/CD access method. 802.5 employs a token passing access method and can support data speeds of 4 or 16Mbps. In this implementation, the physical network topology can vary although the logical topology is always that of a ring. The 802.12 is based upon a new standard termed 100VG-AnyLAN which uses a contention-based, demand priority arrangement for network access.
The ability of a computer to control input and output events is crucial to the ability of that computer to communicate with other compatible devices. Common input and output events include sending and receiving data from a network, the writing and reading information to and from a floppy or hard disk and controlling document printing.
The Internet is a worldwide network of networks that is based on the TCP/IP protocol. The Internet is not owned by a single company or organization. The worldwide collection of computers, networks and gateways that use TCP/IP protocols to communicate with one another. At the heart of the Internet are high-speed data communications lines between major host computers, consisting of thousands of commercial, government, educational, and other computer systems that route data and messages. Currently, the Internet offers a range of services to users, such as e-mail, the World Wide Web, FTP, Usenet newsgroups, Gopher, IRC, telnet, and others. See also: FTP, HTTP, SMTP, TCP/IP Protocol Suite and WAN.
InterNIC (Internet Network Information Center)
This central organization assumes responsibility for the allocation of IP addresses and the assignment of domain names to organizations, companies and educational institutions that wish to connect to the Internet. An important function of the InterNIC is to ensure that each IP address or domain name is unique and is not shared by several users. See also: DNS, DNS Hierarchy and FQDN.
An intranet uses the same technologies as the Internet, but is owned and managed by a company or organization. An intranet is typically implemented as a LAN or WAN. A network designed for information processing within a company or organisation. Its uses include such services as document and software distribution, access to databases, and training. An intranet is so called because it usually employs applications associated with the Internet, such as Web pages, Web browsers, FTP sites, e-mail, newsgroups, and mailing lists, accessible only to those within the organisation. See also: LAN, WAN and Internet.
An IP address is a unique number that is used to identify a particular host connected to a TCP/IP network. The IP address contains both the network number and the host numbers, used to determine both the network and computer. IP addresses, in their raw form, are 32 binary digits (bits) long. In order to make these long strings more memorable, they are split into groups of eight digits (octets) and converted to decimal notation. This resulting format, called dotted decimal notation, makes IP addresses easier to use. In decimal terms, the number 255 cannot be assigned as the representation of an octet since this would mean that each bit within that octet is set to 1. The dotted decimal notation 255.255.255.255 is reserved as a broadcasting address. IP addresses may be divided into class A, B and C addresses. ‘Class A’ addresses use only the first octet, whereas ‘class B’ use the first and second octets and ‘class C’ also includes the third octet. See also: DNS, DNS Hierarchy and FQDN.
IP (Internet Protocol)
Computers must use IP to communicate across the Internet. The standard way of identifying a computer that is connected to the Internet is through the use of an ‘IP Address’, similar to the way a telephone number identifies telephones on a telephone network. The IP address encompasses two pieces of information: the network number – this is common to all hosts on the same network, and the host number, which is responsible for identifying a host on a particular network. An IP address is 32 binary digits long and is used within an IP packet to define the source and destination of the packet. To aid their usability, these digits are arranged into groups of 8 bits called octets. These octets are then converted to decimal notation. The decimal number 255 is used as a reserved decimal representation of an octet. See also: MAC Address.
IPX/SPX (Internetwork Packet Exchange/Sequenced Packet Exchange)
Novell developed the IPX/SPX protocol suite for use on its NetWare network operating system. It describes the functioning of the transport and network components of the theoretical OSI model. Since Novell owns this specification, it has typically been difficult to obtain information concerning this protocol. It is simple to configure for small networks and several networks upon which it was used employed different frame types for data transmission. Therefore a common frame structure must be agreed between the connected machines in order for communication to occur. See also: Frame and OSI Model.
IRQ (Interrupt Request)
An interrupt request line is used by a hardware component (e.g. a keyboard, mouse or hard disk) to inform the processor that it requires attention. This enables the processor to work more efficiently as it does not have to constantly scan every device to determine device status, instead the device interrupts the processor when necessary. Each device operating on a particular system is normally required to have an individual IRQ number although sharing can occur with some components such as tape and hard disk controllers.
ISA (Industry Standard Architecture)
The ISA bus was originally an 8-bit connection used on the IBM PC and XT motherboards. IBM AT and other compatible machines used an improved 16-bit design. Despite continual improvements in design, the ISA bus offers limited performance (8MHz) and alternative bus designs have been implemented. See also: PCI.
ISDN (Integrated Systems Digital Network)
ISDN is the digital version of the PSTN that has been provided by telecommunication companies since the 1980s. It is a digital switched-circuit technology for voice, video and data – hence the term ‘integrated systems’. Full implementation of the system has been delayed due to the time taken to uptake the analog components of the network. The transmission speed is dependent upon the class of ISDN used and although ISDN prices were initially high, subsequent updating of telecommunications networks has lead to competition and a fall in prices. See also: PSTN and Bandwidth.
ISOC (Internet Society)
The purpose of the ISOC is to encourage and promote the development and availability of the Internet. It undertakes technical research (through the IAB), the formulation of solutions to Internet problems and is responsible for assigning a unique number (called a port) to each protocol used on the Internet. See also: IAB.
LAN (Local Area Network)
A LAN is defined as a network that does not exceed a distance of 2km from end to end. The IEEE 802.11 standard provides a definition of wireless LAN networks. See also: IEEE 802.x Standards.
One of the original methods used to resolve IP addresses to specific host names was to use a central server containing a text file list of these data. This file was called a ‘HOSTS’ file. The ‘LMHOSTS’ file is an extension of this concept that contains NetBIOS names as opposed to host names (LM = LAN Manager). LMHOSTS contains an index of all the IP addresses and machine names for all the servers for which access is required. A few problems can be encountered when using the LMHOSTS file, although this is normally due to invalid or misspelled entries or the presence of duplicate entries where only the first entry is ever accessed.
Logical Link Control (LLC)
This is a division of the data link layer described by the IEEE. It is responsible for establishing and maintaining a link between communicating devices for the transmission of frames. This process occurs at a service level (i.e. whether the network is connection-orientated or connectionless) and at the flow and error control levels.
This type of test determines the viability of the connection between source and host computers. Three common forms are used: analog, local digital and remote digital. In the analog loopback test data is sent from the host to the modem, where it is converted into analog form and looped back, converted into digital once again and received back at the computer. This is used to test the circuitry of the local modem (i.e. the connection between the modem and the computer that it is slotted into). Local digital loopback tests involve sending data from the destination computer, through the modem (where it is converted), across a network connection and to another modem. The data is received, decoded and converted back into analog before being sent back to the remote computer. The remote digital loopback test is the exact opposite of this principle.
MAC (Media Access Control) Address
A MAC is a unique hardware address that is hard-coded into a network card by the manufacturer. This is required for directing data frames across a network and for allowing the network card to compare destination addresses (coded into the data frame) and its own unique MAC address. See also: IP.
MAN (Metropolitan Area Network)
A Metropolitan Area Network is defined as a network that covers the area of a city that is no more than tens of kilometers. It can operate at speeds that are comparable with LANs.
MIB (Management Information Base)
The management information base is an essential component of the SNMP protocol that is responsible for controlling and monitoring networks. The MIB is a database that holds statistics relating to the activity of the device e.g. number of frames that are handled by a hub. See also: SNMP.
Message Handling Service (MHS)
The MHS is present within Novell messaging products where it acts like a gateway and can, if required, translate the format of a message. MHS is capable of sending messages to a foreign mail system and it must be able to physically route the message to the destination message store. See also: Message Store and Message Transfer Agent.
An email message store acts as a mailbox for email. Email may be collected from the message store at the users convenience when the user is connected. The message store also assumes the role of a sorting system and ensures that users receive email that is destined for them. See also: Message Handling Service and Message Transfer Agent.
Message Transfer Agent (MTA)
The media transfer agent is responsible for forwarding email documents from one post-office to that which the recipient is connected.
Mesh topologies are most commonly used in WANs and these layouts are often found in public networks such as the Internet. In theory, a mesh network requires that every device has a point-to-point connection to every other device connected to the network. Since this approach is somewhat impractical, a hybrid is used where only the most important devices are interconnected. Mesh networks provide excellent fallbacks in the case of link failures. See also: Network Topologies.
MCA (Micro Channel Architecture)
Micro channel architecture buses were designed by IBM to overcome the problems associated with ISA buses. They offer fast data transfer (running at 10Mhz compared to the 8MHz of ISA buses), and can support both 16 and 32-bit data. However, the incompatibly of MCA and ISA buses and the large royalties demanded by IBM for the use of the technology lead to a lack of support for this bus amongst computer manufacturers. See also: PCI, ISA, EISA, and PCMCIA.
Mirroring requires two hard disks and a single disk controller. It takes place at the partition level and any partition, including the boot/system partitions can be mirrored. This strategy is the simplest way of protecting a single disk against failure. This approach is more expensive than other forms of fault tolerance for networks other than peer-to-peer and modest server-based LANs. During mirroring, data is written to both partitions/disks. See also: Duplexing and Striping.
Modems are devices that are used to convert the digital signals from a computer into the appropriate analogue signal that is required for transmission over public phone lines – this is called modulation. The reverse process, demodulation, occurs at the receiving computer. Modems are available in internal and external forms for different computer expansion slots and vary in terms of speed and data handling capabilities. See also: CSU/DSU.
Multistation Access Units (MAUs)
MAUs are used in IBM and similar token ring networks. They are also known as ‘wiring centers’ or ‘concentrators’. These are used to allow token ring networks to be physically wired using a star topology while retaining the logical topology of a ring within the MAU. The MAU uses electrical relays to allow devices to join the ring. A device must supply a continuous 5 volts to maintain the ring connection – in the event that the device is switched off or during failure, the connection is lost. See also: Ring Topologies and Token Passing.
Negative Acknowledgement (NACK)
Within the TCP protocol, whenever a data packet is transmitted, the sending computer expects an acknowledgement from the destination device upon receipt of the data. If the data is corrupt, the destination device returns a negative acknowledgement to indicate the state of the packet and that the source computer should retransmit the damaged information. See also: TCP/IP.
NetBEUI (NetBIOS Enhanced User Interface)
IBM developed this protocol suite for its PC networks. NetBEUI originally contained the NetBIOS protocol but this was later removed for inclusion in the TCP/IP and IPX/SPX protocols. See also: TCP/IP Protocol Suite and IPX/SPX Protocols.
NetBIOS (Network Basic Input/Output System)
The NetBIOS system provides a basic mechanism for controlling data transmission on networks. This technology facilitates the sending and retrieval of data (input/output) from connected systems. See also: NetBEUI.
NetWare Core Protocols (NCPs)
NCPs are similar in function to SMBs in the NetWare environment. They function at four layers of the OSI model – application through to transport. NCPs provide a group of functions that manage the interchange between client and server. They are responsible for performing all file and print services between clients and servers. See also: SMBs.
NDIS (Network Driver Interface Specification)
All devices that are connected to a computer require a driver in order to function. Originally, device drivers could only allow one particular protocol to be bound to a driver thus limiting its functionality. The NDIS specification was developed by Microsoft to permit the binding of multiple protocol stacks to the same driver. It is similar in function to the ODI driver. See also: ODI.
NDS (NetWare Directory Service)
NDS is a directory system that provides a centralized list of resources that facilitate the viewing and management of network connected devices. The NetWare directory service is also concerned with the processing of requests from clients to uses connected resources. By containing a centralized list of resources, NDS provides a single point for network administration and users can be given permission to use objects across networks regardless of the server to which they are physically linked.
A network monitor is an application that continuously tracks network activity and so provides a picture of the current or past conditions. Baseline network conditions can be determined by repeated sampling of network activity using such monitors, this can, in turn, be used to develop expansion plans and isolate problems. Network monitors, unlike protocol analyzers, do not decode frame content, but instead merely track data transmission levels. See also: Protocol Analyzer.
Network File System (NFS)
This was originally developed by Sun Microsystems to allow sharing of resources between a wide number of systems including OS/2, UNIX and NT. Each system is required to use TCP/IP as the transport protocol and an NFS platform can access any device using the NFS service as if it were a local resource.
Network Interface Card (NIC)
The network interface card allows a physical connection between the computer and the transmission media. The NIC is variously known as the network card, the network adapter card or the transceiver. The network card moves data from the computer to the network and vice versa – to do this the card must convert the data into the necessary format for the computer and the network to understand. Network interface cards can address other cards and can recognize data that is destined for itself. It does this by using a unique address known as the media access control address. Error checking is also performed by the card. Network cards are designed for specific types of networks and do not work on different network structures, e.g. Ethernet and token ring. Network cards are available for several different expansion ports on the computer, including the ISA and PCI ports. See also: MAC Addressing, Frames and Network Topologies.
The shape or structure of a network is commonly described as its’ topology. Topologies may be either physical, i.e. the actual appearance of the network layout, or logical, where the topology refers to the flow of data across the network. Star, bus, ring and hybrid topologies are the more everyday designs – with each kind of arrangement having individual merits and drawbacks. See also: Ring Topology, Star Topology, Bus Topology, Hybrid Topology and Mesh Topology.
NT (New Technology) Servers
Windows NT can be used as the operating system on network servers – these NT servers can have several advantages over other server types. NT servers can be installed and used in one of three main roles: a primary domain controller, a backup domain controller or as a standalone member server. Since NT uses a variety of file systems, which are not entirely compatible with other Windows environment filing systems, initial set- up should be carefully undertaken.
NTFS (New Technology Filing System) Security
NTFS security works on the premise of restricting access to shared resources to only specific groups or users. This does however have several limitations and will only allow protection of a resource when a user connects over a network and not from local machines. Also the restrictions that are applied to a root directory of the share will also apply to all of the subsequent subdirectories and all files will assume similar permissions. See also: NT Servers.
NBTSTAT (NetBIOS over TCP/IP Statistics)
The NBTSTAT utility is used to provide NetBIOS status information when using TCP/IP protocols. See also: TCP/IP Protocol Suite.
‘Node’ is the term given to any device, such as a workstation, server or printer that can communicate on a given network.
ODI (Open Driver Interface)
All devices that are connected to a computer require a driver in order to function. Originally, device drivers could only allow one particular protocol to be bound to a driver thus limiting its functionality. Several companies, including Novell, developed the ODI specification to permit the binding of multiple protocol stacks to the same driver. It is similar in function to the NDIS specification. See also: NDIS.
Optical networks use fibre transmission media and provide bandwidths up to 10Gbps. These type of networks are generally used as the backbones for telecommunications and can offer a hierarchy of bandwidth levels. Optical networks are however expensive and are generally only suitable for large networks were speed and distance are important concerns.
OSI (Open Systems Interconnect) Reference Model
This reference model was developed in 1977 by the ISO to aid the understanding of how a network system functions in terms of both software and hardware components. The OSI divides the actions of hardware and software into seven separate sub-tasks: application, presentation, session, transport, network, data link and the physical layer, each with a separate function. The OSI model serves as a functional reference for network communication but it does not represent any individual standard although many protocols do comply with its’ guidelines. See also: 802 Protocols, TCP/IP Protocol Suite, IPX/SPX Protocols and LLC.
OSPF (Open Shortest Path First)
Routers are able to switch between multiple separate networks. In dynamic routing, routers can learn possible connection paths from each other and this can be supported by protocols such as the OSPF which instructs the router to open up the shortest link between two machines. See also: Routing.
Data is transmitted across networks in small chunks called packets. Each packet has a particular uniform structure containing information about the source computer, the destination, error checking and the data chunk itself. Prior to packet transmission, packets must be correctly assembled and before the data can be read by the destination, the transmitted packet must be broken down into the appropriate fields.
A packet generator is used to create test frames of data that can be used to access network performance and thus network administrators can examine areas of a network that may be causing bottlenecks. See also: Test Frame.
Data parity is an integral feature of network transmission. When a data packet is sent, it contains a specific number of bits – either an odd or an even number of bits are present. If parity is set to odd, the computer that is sending the data adds the required number of bits so that the number of data bits becomes odd. This is the same principle if parity is set to even. Parity provides a simple error checking system through which transmission faults can be overcome. If two computers have been set up to transmit using the same parity, any fault that causes the data to be disrupted may lead to an altering of parity. This discrepancy in parity is used to signify data error.
Password hashing is a mechanism for encoding data that is not required in a decoded form. An example of this might be a file containing passwords for network access – the password would be transformed (hashed) and stored within a file. This file would be inaccessible to users wishing to discover the original password. When a user logs onto a system using their password, the password undergoes the same password as the original and is compared to the ‘hash’ file. If a match is found, the user is allowed access to the network.
PCI (Peripheral Component Interconnect)
Peripheral Component Interconnect was introduced in 1995 with the Pentium processor. It is a local bus, like the Enhanced Industry Standard Architecture (EISA) bus and it connects the CPU, memory and peripherals to a wider, faster data pathway. PCI supports Plug and Play and it supports 64-bit data, which in turn leads to performance improvements in graphics and audio intensive operations. See also: ISA.
PCMCIA (Personal Computer Memory Card Industry Association)
The Personal Computer Memory Card Industry Association was founded to provide a standard bus for laptop computers. Although this type of bus was originally designed to allow memory cards to be used, it has been extended to included modems and network cards. See also: ISA, EISA, MCA and PCI.
In peer-to-peer networks there is no dedicated server, but instead, each computer connected to the network acts as both a server and client – i.e. each computer is a ‘peer’ of the other computers. These types of networks were originally developed as a low-cost alternative to server-based systems for use in smaller companies and organizations where there are approximately ten to fifteen users. A major drawback to this type of network is a comparative lack of security whereby each user must control access to resources on his/her machine. See also: Server-based Networks.
PING (Packet Internet Groper) Utility
This program is a diagnostic utility that is used to test TCP/IP configurations and to identify connection failures. The ‘ping’ command sends a message to a computer using a specific message type, if this message is returned the destination is adjudged to be correct and reachable.
‘Plug and Play’ is an independent set of generic specifications developed by a consortium of hardware and software manufacturers with the goal of allowing changes to be made to the configuration of a PC with minimal intervention from the user, such as the setting up of DMA and IRQ channels. For Plug and Play compliant devices, installation is automatic – plug the device in, turn the system on and it works. See also: DMA and IRQ.
PPP (Point-to-Point Protocol)
The PPP protocol was devised to replace SLIP. It is easier to use and configure and addresses the deficiencies in this other mechanism. PPP allows the user to specify a means of authentication, it provides data compression negotiation, handles error detection and correction and it supports multiple protocols. Point-to-Point protocols use a standard frame format containing a ‘frame beginning flag’, the ‘broadcast address’, ‘control’, ‘protocol’, ‘data’, ‘frame check’ and the ‘end frame flag’. See also: SLIP.
Post Office Protocol (POP)
POP is an email system that allows users to receive and read email at their convenience by downloading it from a POP server. Users log on to a server using a POP client (such as Netscape Mail or Internet Explorer Mail) and they are authenticated by a username and password before they can access their email. See also: SMTP.
POTS (Plain Old Telephone System)
See ‘Public Switched Telephone Network – PSTN’
PGP (Pretty Good Privacy)
This software package allows email to be encrypted and it has become the de facto standard for encoding email documents. It is able to run on multiple platforms and uses existing technologies to provide encryption, the inclusion of digital signatures and data compression. See also: Encryption and Decryption
Primary Domain Controller
The PDC in contrast to the Backup Domain Controller, is the initial server that is installed into an NT network domain. This controller contains the master copy of the security database and it can also act as a resource handler by allowing various file, printing and application procedures. Each domain can only contain one primary controller but several backup controllers may be installed to provide tolerance. See also: NT Server and Backup Domain Controller.
Private/Public Key Encryption
This system of encoding, which is also known as asymmetric encryption, has two main roles: normal encryption and the inclusion of digital signatures. The terms ‘public’ and ‘private’ key refer to the specific (mathematical) code that is used to alter the data to make it hidden. Public keys are given to all users wishing to send encoded messages and each public key is associated with a particular user. Private keys are never transmitted and is normally password protected. Decoding of previously altered data is performed using this key and as such, it is never transmitted or shared across a network (whereas a public key may be freely transmitted). See also: Digital Envelopes, Encryption and Secret-Key Encryption.
Protocol analyzers are similar to network monitors but they permit both traffic analysis and packet capture and decoding. Information can be retrieved concerning the content of the various OSI layers, protocols, packet functions and frame data. Many newer protocol analyzers include a built in TDR system to allow cable fault detection. See also: Network Monitor.
A proxy server, as its name suggests, functions on behalf of a network during Internet access operations. It provides a central point through which communication occurs and can therefore be used to protect unauthorized access to the internal network from the Internet and can be used to log user access. Further protection can be arranged by using Firewall facilities in conjunction with proxy servers. See also: Firewall.
PSTN (Public Switched Telephone Network)
This form of public network is also known as a ‘plain old telephone system (POTS)’. PSTN was designed for transferring voice information between analog telephones. It works by establishing a temporary dedicated path between two locations when the conversation commences and the removes the connection when the conversation ends. Although it was not originally designed to support data transfer, it has been upgraded to allow information transmission. In order to connect to PSTN, which is essentially analog, the computer must use a modem to convert the digital computer signal into an analog signal for transmission. The receiving computer then reconverts this signal back into digital using a modem. See also: ISDN.
Radio transmission can be used to form a wireless connection between two networks or two or more nodes. Useable radio frequencies can range from 10KHz to 1GHz and may fall within the range of LW, MW, SW, UHF or VHF. In order to use most frequencies a licence must be obtained although there are some frequencies that do not require a specific permission to be granted (e.g. 2.4GHz is an internationally unregulated frequency). The choice of frequency that is used normally depends upon distance and cost. See also: Spread-Spectrum Transmission.
RAID (Redundant Array of Inexpensive Devices)
Using RAID technology, users can use multiple hard disks to provide a backup measure for network servers and workstations. Several levels of backup are suggested by this system, ranging from level 0 to level 6, each level representing a particular type of fault tolerance. See also: Duplexing, Mirroring and Striping.
RAS (Remote Access Server) gateway
When a user needs to access a remote network, a RAS server provides the means of achieving this. A RAS server is necessary to receive the incoming signal from a computer. In order to be able to receive and understand an incoming signal, the RAS server and client interfaces must be compatible either by using a modem link or by using ISDN adapters. See also: Modem and ISDN.
In the event that a semi-critical component within a network was to fail, several mechanisms have been introduced to allow a backup device to take over from the faulty device. This is termed redundancy and it is commonly used as a hard-disk backup strategy. See also: Tolerance.
RFC (Request for Comment)
Request for Comment documents are circulated on the Internet and are used to provide a means for developers to test ideas. RFCs are reviewed by independent experts or by task forces and older RFCs are never replaced by newer documents. Each new document is assigned a different number, much like the version numbers found within computer programs. The original development and enhancement of the TCP/IP occurred through the publishing of RFC documents. See also: IAB and ISOC.
In a ring topology, all of the computers are connected in a circle. The ring comprises a series of point-to-point links between each device and the computers are either directly connected or are linked to the ring through a multi-station access unit. Common examples of ring networks include IBM’s token ring and the FDDI and CDDI. Ring networks are relatively easy to install and the system can be developed to incorporate a second ring that acts to preserve network integrity in the event of a media or device failure. See also: Network Topologies.
These forms of connectors come in two main varieties, RJ-11 and RJ-45 (RJ stands for Registered Jack) and are used within twisted pair cable networks. The RJ-11 connector has four contacts and thus uses a two-pair cable – it is the connector that is familiarly seen in telephone connections. The RJ-45 connector has eight contacts that are linked to a four-pair cable and it is this cable type that is usually selected for data networks. Occasionally, this connector has a ‘key’ to prevent accidental insertion into the wrong socket.
Remote Monitoring is a cooperative technology that is responsible for remote network handling. Using this system, network activity can be recorded and MIB data from sub-networks can be collated. Remote monitoring allows a network administrator to act upon trapped operations. See also: MIB.
Routing is similar in principle to bridging although it allows a greater degree of flexibility. Routers are able to link dissimilar networks and can support multiple alternate paths between locations based upon the parameters of speed, traffic loads and cost. Routers form the basic connections of the Internet and can allow data to take multiple paths (hence reducing the likelihood of transmission failure) to reach a destination. Routers function at the network layer and can therefore access source and destination addresses within packets and can keep track of multiple active paths within a given source and destination network. This allows more fault-tolerance than bridges where multiple concurrent paths are not allowed. Routing may be either ‘static’ or ‘dynamic’. See also: Switching.
SAP (Service Advertising Protocol)
This protocol is used by the NetWare environment to advertise services using the application layer of the OSI model. Service providers such as file servers and print servers broadcast a SAP packer every 60 seconds to advertise their presence on the network. This packet informs the client that the service is available. Clients may send a ‘service query packet’ to request information.
SCSI (Small Computer Systems Interface)
The SCSI system operates using a parallel data stream and combines this with hardware handshaking and control signals. It allows the connection of multiple devices to a single SCSI controller with each device being attached by a ribbon cable. The SCSI standard also defines a command language that is used by the controller to detect and identify which devices are connected to the control bus and how they can be accessed. SCSI interfaces may be either internal or external.
This type of encryption involves the use of an identical ‘key’ to encode and decode data. The key is normally transmitted with the file, although a means of conveying the key securely between source and destination is also required. Generally, the number of bits used within an encryption algorithm (key) determines the effectiveness of the coding. Due to the nature of encryption, governments from several companies do not permit the export of very high bit technologies (e.g. using 40 bits). See also: Private/Public Key Encryption and Encryption.
Secure Electronic Transaction (SET)
This system was developed jointly by Visa and MasterCard from existing techniques for banking transactions over the Internet. It covers the purchasing of goods and services electronically (termed e-commerce) and allows the authorization of payment requests and access to purchaser and vendor credentials.
Secure Sockets Layer (SSL)
SSL was designed by Netscape to provide privacy and authentication over the ‘net. It is independent of any application and can be used with range of protocols such as HTTP and FTP.
Segments are components of larger networks that are linked by a connection device such as a hub or bridge. See also: Hub and Bridging.
Servers act as the main link to the shared resources on a network and they control the way in which clients can access and alter information and can control shared resources. See also: Server-based Networks.
A server-based network uses a dedicated central server (or servers) to provide access to available resources on the network. Security is managed centrally by a system administrator who determines access to these resources. The first server-based network was developed by Novell and appeared as ‘NetWare’ in 1985. Server-based networks offer performance, security, central control and scalability, although they are expensive to set-up and they are very complex.
SLIP (Serial Line Internet Protocol)
This very simple protocol is based upon the TCP/IP protocol. It supports the Internet Protocol and provides support for static IP addressing (unlike the improved PPP protocol). It does not support error detection or correction directly, although repeated header information can be removed to add a degree of compression and enhance data flow. See also: PPP.
SMBs (Server Message Blocks)
Microsoft Windows and the OS/2 operating systems use a protocol called ‘server message blocks’. This is used for requesting files from serves and delivering them to clients. SMBs allow machines to share files and printers, thus making them available for other machines to use.
SMDS (Switched Multi-megabit Data Services)
SMDS is a simple, cost-effective high-speed transport service that is based around the same type of cell switching technology as that present within ATM networks. Consequently, SMDS has been used by many as an upgrade route to ATM technology. See also: Asynchronous Transfer Mode.
SMTP (Simple Mail Transfer Protocol)
The SMTP protocol specifies how mail is delivered from one system to another. It is used to deliver messages from the e-mail client to the SMTP server and between SMTP servers. SMTP is not used however to transfer a message from the recipients SMTP server to its e-mail client since it requires both source and destination to be online before it can make a connection. In this later instance, post office protocols allow a user to download messages from his/her SMTP server at their convenience. See also: X.400 Addresses and X.500 Directories.
SNMP (Simple Network Management Protocol)
This protocol provides a mechanism for managing and monitoring devices on a TCP/IP network. It is comprised of three basic components: the ‘SNMP manager’ – this software allows a computer to act as a network management station; the ‘agent’ – software or firmware that allows the computer to store data and respond to queries from the SNMP manager and the ‘management information database’ that provides a database structure into which statistics can be collected.
‘Spread spectrum’ is a type of wireless communication systems that transmits data by simultaneously using multiple frequencies. These frequencies may be either directly modulated or may be randomly ‘hopped’ between. This type of technology makes eavesdropping difficult and can allow highly secure data transmission. See also: Radio Frequencies, Direct Modulation and Frequency Hopping.
SQL (Structured Query Language) and SQL Servers
The front-end requests to a SQL (pronounced ‘sequel’) server are made in the form of SQL statements. IBM originally devised this database query language in the mid-1970s and this was further developed by the Oracle Corporation, who in 1979 released the first commercial SQL database. SQL is now one of the most popular database management systems and its popularity has caused database vendors to ensure SQL-compliance within their programs. The American National Standards Association adopted SQL as the standard database management system in 1986.
In a star network, each node is connected to a central device using a point-to-point link. These central devices are usually called ‘hubs’ or ‘concentrators’. The hub receives a signal from a node and repeats the signal to all of the other connected nodes. Star topologies are easy to reconfigure and any faults in the media or in a particular device can be readily localized. However, in the event of hub failure, all of the connected nodes are affected. See also: Network Topologies.
State Transition Encoding
State transition encoding is a form of digital encoding that converts the binary data into representative electromagnetic signals. In binary, data exists as either a 0 or 1 – if you represent this as a current state transition, a positive to a negative current might be the equivalent of a 1 and the change from negative to positive might signify a 0. This type of encoding is also termed ‘Manchester encoding’ and it is used in Ethernet networks. See also: Current State Encoding.
STP (Shielded Twisted-Pair) Cable
Shielded Twisted-Pair cable, is a form of TP cabling that uses a thick braiding to reduce electrical interference. Although it is less susceptible to external interference, it is more difficult to install than UTP since the shield must be grounded. See also: Twisted-pair Cable, UTP Cable and Fiber Optic Cable.
This is the most common strategy for new fault-tolerance designs. It differs from other levels of RAID (Redundant Array of Inexpensive Devices) technologies in that it writes the parity information across all the disks in the array. The data and parity information are managed so that the two are always on different disks. If a single drive fails, enough information is spread across the remaining disks to allow the data to be completely reconstructed. See also: Mirroring and Duplexing.
Switches look similar to hubs although they are able to function in an entirely different manner. Switches receive incoming data into a buffer and the destination MAC address is compared with an ‘address table’. The data is then only sent out to the port with the corresponding MAC address. In a switched network means that each port is in a separate collision domain and, therefore, collisions cannot occur. See also: Hub, Bridging, Routing and MAC Addressing.
Switching is the process that is used to connect source and destination computers and describes the mechanism by which data is forwarded at intermediary points in this link. Switched networks allow data to be shared between two or more locations and it is more efficient than a series of leased lines for connecting multiple locations. Switched networks can be divided into circuit switched networks or packet switched networks – the former occurs when a temporary dedicated path is established between two locations. Packet switching occurs when data is split into packets each containing a packet address and a route address. This provides a cost-effective solution to the construction of a leased line with users sharing a network infrastructure. See also: Routing.
A subnet mask is used to determine the separate components of an IP address. A subnet mask is similar to an IP address but each decimal octet is set to 255. The value of 255 represents eight positive bits (i.e. each bit is set to 1). When the subnet mask and the IP address are combined, using a bit-wise AND function, the appropriate portion of the address is revealed. See also: IP Address.
System Performance Monitoring
Networks can be monitored to examine data flow between various interconnecting segments. Computer programs, called either protocol analysers or network monitors, can be used to obtain network statistics, determine the presence of data bottlenecks and some can even inform of network connection failures such as kinks or breaks within the cable.
TCP/IP (Transmission Control Protocol/Internet Protocol)
TCP/IP is the ‘de facto’ communications standard used by the Internet. Following its initial development for military installations, it was adopted by other government agencies and universities in the US and was later incorporated into UNIX network operating systems. The TCP/IP protocol is the standard communications protocol used by the Internet and it contains higher level features to accommodate the transfer of files, mail and documents across compatible networks. This protocol maps to a four-layer conceptual model: application, transport, Internet and network interface. See also: IP and Internet.
TDR (Time Domain Reflectometer)
TDRs are able to locate open and short circuits, kinks/sharp bends and other imperfections in cable media that can adversely affect transmission performance. A TDR emits a pulse of known amplitude and duration down a cable and measures the corresponding amplitude and time delay of the signal reflections. Analysis of these reflected signals can indicate and allow the isolation of a particular problem. See also: System Performance Monitoring.
Telnet is a standard program designed to allow users to login to a remote network. It can be used to send and receive email from a remote account and to access data services on remote networks.
Terminators are used within coaxial-based networks to inform the network circuit that the cable has a logical finish. Terminators can be used a means of detecting breaks within network cables by systematically dividing cables at their mid-point. Successive repetition of this strategy allows faults in the cable to be discovered.
A test frame is the data packet sent by a packet generator to assess the integrity and transmission profile of a network. See also: Packet Generator and Network Monitor.
Time-domain Reflectometer (TDR)
TDRs are able to locate open and short circuits, kinks/sharp bends and other imperfections in cable media that can adversely affect transmission performance. A TDR emits a pulse of known amplitude and duration down a cable and measures the corresponding amplitude and time delay of the signal reflections. Analysis of these reflected signals can indicate and allow the isolation of a particular problem. See also: Protocol Analyzer.
Token Ring Networks
Token ring networks use token passing to transmit small data frames called the ‘tokens’, which are passed from device to device. Essentially the token is generated by a computer on the network and is passed from node to node. When a computer is in possession of the token, it can transmit data onto the network. Within this type of system, access times are predictable, data can be prioritised and collisions cannot occur. However, this system requires complex software and reconfiguration may be necessary when devices are added or removed. Token ring networks are generally less popular than the Ethernet types due to their higher cost.
Token passing involves the transmission of a small data frame called the ‘token’, which is passed for device to device. Essentially the token is generated by a computer on the network and is passed from node to node. When a computer is in possession of the token, it can transmit data onto the network. Within this type of system, access times are predictable, data can be prioritized and collisions cannot occur. However, this system requires complex software and reconfiguration may be necessary when devices are added or removed. Token ring networks are generally less popular than the Ethernet types due to their higher cost. See also: CDDI, FDDI, Contention and Ethernet.
The term tolerance refers to the ability of a network (in this case) to withstand a problem that may otherwise cause a failure within the connection. Examples of devices that allow ‘fault tolerance’ include uninterruptible power supplies, drive arrays, backup strategies and regular virus prevention. See also: Redundancy.
This utility allows a user to track the route taken by a data packet as it crosses from network to network to complete its journey from source to destination devices.
Tunneling is a system that is used to access a remote, but identical network, that is connected by a different network system. An example of this could be source and destination networks that use the NetBEUI protocol and are connected via a PSTN link. The tunneling protocol is said to ‘encapsulate’ the original protocol (in the case of the example – NetBEUI).
Twisted-Pair (TP) Cable
Twisted-pair is a common type of cable that has been extensively used for telephone systems. This cable is formed from one or more twisted cable pairs (i.e. where two insulated cables are twisted about each other). The twisting of the wires (which are generally 22 or 26 American Wire Gauge copper wires) acts to reduce interference and crosstalk. Each pair of wires is twisted at a different rate to ensure that the pairs do not interfere with each other. Several drawbacks of TP cabling however are its’ sensitivity to electromagnetic interference and eavesdropping and that it can not be used for long-distance transmission. See also: UTP Cable, STP Cable and Fiber Optic Cable.
UART (Universal Asynchronous Receiver/Transmitter)
A UART chip is the major component within a PCs serial port. It is responsible for processing incoming and outgoing serial data into the correct format. UARTs are also associated with the handling of parity generation and checking, the addition of start/stop bits and serial to parallel conversion (and vice versa). See also: Modem.
Universal Naming Convention (UNC)
Microsoft networks (including Windows 95) use syntax known as the Universal Naming Convention to identify network resources. This is comprised of a double-backslash followed by the machine (server) name, then a single backslash followed by the sharename. For example, a colleague working at a machine named DONALDD has decided to share their documents directory with a share name of WPDOCS. If you connected to that share, the resource would be written as \\DONALDD\WPDOCS.
UNIX was originally developed by the telecommunications company, AT&T during the late 60s and early 70s. UNIX is now a family of more than 20 related operating systems that are produced by various companies. It has become the operating system of choice for many high powered workstations. It is capable of supporting parallel processing and is able to be run on a wide variety of platforms. UNIX offers a multitude of file systems in addition to its native system. UNIX servers are the main types of server that form the Internet and it is able to use the TCP/IP protocol suite to provide compatibility between networks. See also: Internet.
UPS (Uninterruptible Power Supply)
Uninterruptible power supplies provide an alternative AC power supply in the event of power failure. Using this type of device, power surges and spikes are minimized. A UPS requires an array of batteries, a charging circuit, an inverter to convert DC to AC current, a circuit to allow the system to take over from a failing power supply and some degree of spike and surge protection.
URL (Uniform Resource Locator)
The Uniform Resource Locator is an alphanumeric string of characters that contain information that details the Internet address of a resource on the Internet. A resource may be a web page, an FTP or gopher site or another type of data connection. Within the URL, several other characters such as a colon (:), forward-slash (/), tilde (~) or period (.) have specific roles in further defining a particular resource. See also: HTTP, FTP and World Wide Web.
User Datagram Protocol (UDP)
UDP is a connectionless, non-guaranteed method of communication with no sequencing or flow control. The application layer controls delivery reliability, although there is nothing to ensure that messages are delivered or the sequence in which packets are received. UDP is used where data is not dependent upon acknowledgement of delivery, e.g. during the broadcasting of the existence of machines on a particular network.
UTP (Unshielded Twisted-Pair) Cable
Most twisted-pair cable that is used in networks is unshielded. The primary reason for UTP’s popularity is that it is extensively used in many telephone systems (albeit at a lower specification). UTP is widely available and various specifications have been standardized in terms of the frequency and capacity of data that each can transmit.
This new technology has been developed by a group of companies including Hewlett Packard to overcome the limited bandwidth offered by older LAN systems. It is variously known as 100VGAnyLAN, 100BaseVG or AnyLAN. It provides a bandwidth of 100Mbps and it uses voice grade (VG) copper wiring (although fiber optic cable can also be used). This technology is capable of supporting both Ethernet and token ring networks and it uses a demand priority access method to provide high and low priority levels. The IEEE 802.12 standard describes the VGAnyLAN technology. See also: 802 Protocols.
Computer viruses, in their simplest form, are programs that are designed to spread (and hence replicate). The replication and action of viruses may be either silent or obvious to the user, possibly indicated by erratic computer behavior, disk or file corruption and the display of graphics or text on the monitor display. Viruses take several forms that specifically (or randomly in some cases) target computer components, software or boot programs.
Volt-Ohm Meter (VOM)
A Volt-Ohm meter is a basic cable-testing device that is used to access the physical connections within a network. They can indicate cable integrity and the existence of shorts or breaks in the circuit. A break in a circuit is generally associated with an infinitely high level of resistance.
WAN (Wide Area Network)
A Wide Area Network is a network that spans a relatively large geographical area, incorporating more than one site. Connections are made using methods such as telephone lines, ISDN lines, radio waves or satellite links. The distance that a WAN covers is usually greater than 2km and often comprises a series of LANs that have been joined together. See also: CAN, LAN, MAN and Internet.
WINS (Windows Internet Name Service)
The WINS software is provided with the NT server network operating system. Each host on the internetwork is configured with the IP address of at least one WINS server and often a second WINS server is provided to overcome redundancy. When a machine starts or shuts down, it sends a message to the WINS server identifying its name and IP address. The WINS server holds a dynamic database of registered names and corresponding addresses – it does not allow duplicate names to be registered. When a machine requires the IP address of another machine, the query is sent to the WINS server, which can provide this information. This method has the advantage of eliminating much of the broadcast traffic on Windows networks. See also: Dynamic Host Configuration Protocol and DNS Server.
A wireless bridge, as its name would suggest, is a non-physical link between two networks or between two or more nodes on the same network. Common media for wireless bridges include infrared, microwaves, narrow band radio and spread spectrum radio transmission. See also: Radio Frequencies and Spread-Spectrum Transmission.
World Wide Web (WWW)
The World Wide Web (also called WWW, the ‘web or W3) was designed by Tim Berners-Lee in 1989 for the European Laboratory for Particle Physics. It represents the sum total of all of the linked hypertext documents that reside on the HTTP servers across the world. Documents contained on the web, also known as web pages, are written in a standard language called HTML that allows them to contain graphics, formatted and plain text, respond to user input, play animation files and perform many other functions. Facilities on the web also permit users to download files from sites and to send electronic mail messages. See also: Internet and TCP/IP.
X.25 is a protocol that was originally developed to allow connection between remote terminals and a mainframe host without the requirement for a dedicated, leased line. It is a well founded and understood system and X.25 networks are usually universally available and require a low level of maintenance.
The X.400 email protocol was developed by the Consultative Committee for Telephones and Telegraphy and it uses three main components to produce a messaging system: the user agent, the message transfer agent and the message transfer system. X.400 email addresses have myriad optional data fields that generally encompass the following information: country, the provider of the service, the positional hierarchy of the organisation using X.400 – which may be sub-divided into logical groupings and the user of the service. The user is normally identified on the basis of surname and first name, either in full or as initials, or under an allocated pseudo-name. See also: SMTP.
The X.500 standard defines a set of rules that apply to directory services. X.55 employs a hierarchical system that allows users to be located by organisation, department and name and these may b grouped into a single global directory that can give access to thousands of names. An email system can implement this standard within its recipient directory and can integrate with other X.500 compliant services. See also: SMTP.