"B" Networking Definitions & Concepts...

B8ZS (Bipolar 8 Zero Substitution) .. to .. Byzantine Generals Problem(BGP)

Search for Information Technology Items

A signal-encoding scheme in which a 1 is represented alternately as positive and negative voltage, and 0 is represented as zero voltage. B8ZS requires at least one bit of every eight to be a 1. Also see Encoding, Signal.

In the CCITT X.500 directory services model, the more comprehensive of two sets of access-control guidelines. The less comprehensive set is called SAC (Simplified Access Control). Also see X.500.

A central network that connects a number of other, usually lower-speed and lower-capacity networks. Those lower-capacity networks can pass data to each other over the backbone network. The backbone network is usually constructed with a high-speed (i.e. ethernet, or fiber-optic cable) communication medium. The lower capacity networks connect to the backbone using routers, half-routers, or modems. Information sent from one device to another within a network stays in that network, but data sent from one network to another travels over the backbone. Individual devices can also connect directly to the backbone network; they do not have to be part of one of the lower-capacity networks.

The backbone manages the bulk of the traffic, and it may connect several different locations, buildings, and even smaller networks. The backbone often uses a higher speed protocol than the individual local-area network (LAN) segments.

Cabling connecting a main distributing frame to intermediate distributing frames located in telecommunications closets.

A backbone network is one with a central cabling scheme (the backbone) to which other networks are attached. Nodes in one network can talk to nodes in other networks by sending packets across the backbone network.

The networks attaching to the backbone are known as access networks. Access networks may require a gateway or router to attach to the backbone network.

A backbone network can be usful in decentrilized coroprations. For example, a backbone network might be use in a company in which each department has set up its own network and several different architectures are used. Since the backbone network leaves the networks intact, those networks can continue operating as if they were not on the larger network. However, the backbone gives each of the networks access to the resources and data of the other access networks.

One obstacle to a successful backbone network is the high bandwidth that may be required to handle potentially heavy traffic. Because of this consideration, fiber-optic cable is the most sensible cabling for backbone networks.

In a client/server architecture, the portion of an application that runs on the server and does the actual work for the application. The front end runs on the client machine and provides an interface through which the user can send commands to the back end.

A software application that runs on a computer as a background process, allowing the user to work on other tasks while a document is being printed. Background printing may occasionally slow the computer's performance until the document is actually sent to the printer. See also background process. Compare with print server.

A process or program that executes incidentally, while another process or program is operating in the foreground. The foreground process gets the main attention of the CPU (central processing unit), and the back-ground process takes CPU cycles when the foreground process is temporarily idle.

In NetWare's TTS (Transaction Tracking Syetem), the process of abandoning an uncompleted database transaction, leaving the database unchanged. TTS takes this action to ensure that the database is not corrupted by information from an incomplete transaction. Also see TTS (Transaction Tracking System).

A backplane is a circuit borad with slots into which other boards can be plugged, as illustrated in the figure below "A backplane". The motherboard in a PC or a Printer is a backplane.

A segmented backplane is a backplane with two or more buses, each with its own slots for additional boards.

The metal bracket at one end of a circuit bord, usually at the back when the board is plugged into a expansion slot. The backplate, also known as an end bracket or mounting bracket, typically has cutouts for connectors, switches, and interface cards. PCs and Printers usually come with blank backplates over each expansion slot, which are removed when a you plug a board or card into the slot.

In a fiber-optic transmission, light that is reflected back in the direction from which the light came.

A backup is an archival copy that is stored on an external medium. For example, a backup might contain the contents of a hard disk or a directory.

The creation of regular backups is essential in a networking environment. An effective backup system ensures that data stored on the network can be recreated in the event of a crash or another system failure.

Networking packages differ in the type of backup supported, in the media to which material can be backed up, and in the ease with which parts of the archived material can be restored. Backups are generally made to tape or to erasable optical (EO) media. No serious network should be backed up to floppy disks.

Various types of backups are distinguished, including full, differential, and incremental. In full backups, a copy is made of all the data.

In differential and incremental backups, only the data that has been added or changed since the previous backup is included. Differential and incremental backups assume a full backup has been done and they merely add to this material. Such backups use the Archive flag (attribute), which is supported by DOS and most networking environments. This flag is associated with a file and is set whenever the file is changed after the file is backed up.

The backed up material should generally be stored in a different physical location from the original material, and should be protected from disasters such as fire, flood, magnets, theft, etc.

Backup operations should be done at a time when the network is not being used for its ordinary activity, which generally means outside regular working hours. One reason for this is that most backup programs will not back up a file that is open. Truly, the work of a system administrator is never done.

When you restore the data, you restore the last full backup first, then restore each incremental backup made since the last full backup. Also see Archive, with related articles in data protection; disk duplexing; and disk Mirroring.

BACKUP TIPS

• Keep multiple copies of backups; redundancy should be a part of your backup plan.
• Test your backups to make sure that they are what you think they are.
• Store your backups in a secure, off-site location.
• Replace your backup media on a regular basis.
• Consider making incremental backups of critical data at more frequent intervals

Design that takes into account compatibility requirements for earlier models and builds support for those requirements into later generations. The result is a new product that can still be used on an old product platform or file system.

Error correction in which the recipient detects an error and requests a retransmission. The amount of material that needs to be retransmitted depends on the type of connection, how quickly the error was detected, and the protocols being used.

In data protection, the process by which material written to a defective area of the hard disk is retrieved and rewritten to a different, nondefective area of storage. The defective area is identified as such in a bad block table, so that future writes will not be made to the area. Bad-block revectoring is known as Hot Fix in Novell's NetWare.

In storage management, a table in which all known defictive areas of a hard disk are listed to ensure that nothing will be written to these areas. The process of protecting data in this manner is known as bad-block revectoring, or Hot Fix in Novell's NetWare.

A portion of a disk that cannot be used because it is flawed. When you format a disk, the operating system identifies any bad sectors on the disk and marks them so they will not be used. If a sector that already contains data becomes damaged, you will need special software to recover the data.

Almost all hard disks have sectors that are damaged during the manufacturing process, but these are usually replaced with spare sectors at the factory. By the time the disk is shipped, it should be free of bad sectors. If your disk utility starts showing bad sectors, this is a sign that there is something wrong with the disk or disk drive.

A device that connects a balanced line to an unbalanced line, for example, a twisted wire pair to coaxial cable. A balanced line is one in which both wires are electrically equal, i.e., in impedance. In an unbalanced line, such as coaxial cable, one line (the central conductor) has different physical properties from the other (the surrounding concentric conductor), and this impedance is still different from twisted wire pairs. Balun's are also called balanced converter or "bazooka". Basically as stated above it is a device used for matching (impedance) an unbalanced coaxial transmission line to a balanced two-wire system.

It is a hardware device used to adjust impedances in order to connect different types of cable. The name comes from balanced/unbalanced, because the device is often used to connect twisted pair (balanced) to coaxial (unbalanced) cable.

Baluns may have different connectors at each end to make them compatible with the cable types being connected. For example, a balun might have a BNC connector at one end and an RJ-45 connector at the other.

A balun makes it possible to use twisted-pair wiring that may already be installed in parts of a building or office in conjunction with coaxial cable that is coming from elsewhere or that has been installed more recently. The balun controls the electrical signal's passage from one cable type to the other, but does not change the signal in any other way. Similarly, a balum enables you to connect a network interface card designed for use with coaxial cables to a hub tht uses twisted-pair cabling.

Baluns vary with respect to the cable gauge (thickness) supported and to the maximum cable distance over which the signal is supported. This distance may be as high as 360 to 460 meters (1,200 to 1,500 feet). Coaxial boosters may be used to increase signal strength in the coaxial cable, and thus increase the distance over which the signal will be supported by the balun. However, such boosters can cost up to ten times as much as a balun, and will only double the supported distance. Also see connector, and intranetwork link.

WHAT TO LOOK FOR IN A BALUN:

Baluns may include a stretch of cable (at extra cost,of course). Here are some things to consider when you're shopping for a balun:

• Baluns work most reliably when the cable has low capacitance (20 picofarads/foot or less) and when the cable impedance is not too high.
• Baluns are available in different qualities, based on the type and gauge (thickness) of cable at either end. Make sure the balun you select supports the cable properties and distances you need and then some. To be on the safe side, don't use a balun (or any other kind of connector, for that matter) at the maximum rated length.
• Some network interface card manufacturers recommend specific baluns for their boards. Similarly, some manufactures suggest that you do not use baluns with their hubs or cards. Check with the manufacturer to determine whether either is the case with the network interface card or hub you plan to use.
• When using a balun on a network, you'll almost certainly want a balun designed for data transmission, because this type is made for direct (rather than reversed) pin-to-pin connections.
• Baluns pass signals on, so the balun's reliability depends on the signal's quality. For this reason, it's not a good idea to use a balun with passive hubs, which don't clean and strengthen the signal before passing it on.

The capacity of a network to carry information using a particular type of cable, as measured by the maximum number of bits per second (bps) the network can transmit. In a network, the higher the bandwidth, the greater the information-carrying capacity of the network, and the faster data can be transmitted from one device to another.

Put another way, it is the range of transmission frequencies that a network can use. The greater the bandwidth, the greater the amount of information that can travel on said network at one time.

For networks we can also talk of the bandwidth of the signal as being a measure of its frequency spread. That is, the range of transmission frequencies that a network can use. The greater the bandwidth, the greater the amount of information that can travel on the network at one time.

An analog bandwidth is computed by subtracting the lower frequency from the higher one. For example, the bandwidth of the human voice is roughly 2,700 Hz (3,000 - 300).

As stated above, a larger bandwidth means a greater potential for data-transmission capability. For digital signals, a higher bit rate represents a larger bandwidth. However, the higher the frequency, the shorter the wavelength. A higher bandwidth (that is, a higher signal frequency) means faster transmission, which means a shorter signal. With a short signal, there is a smaller margin for error in interpreting the signal. This means that the effects of attenuation and other signal distortion must be kept to a minimum.

A signal traveling along a cable degrades with distance. It is possible to connect the cable to special componensts that can clean up and rejuvenate a signal. High-frequency electrical signals must be cleaned up frequently, which means single cable segments must be short.

Some commonly used frequency bands for analog transmissions are shown in the table below "Bandwidths on the Electromagnetic Spectrum".

Radio Spectrum Bandwidths

Very low frequency (VLF) through super high frequency (SHF) are considered the radio spectrum. The bandwidths are used as follows:

• AM radio broadcasts in the medium frequency (MF) range (535 to 1,605 kHz).
• FM radio and VHF television broadcast in the very high frequency (VHF) range (88 to 108 MHz for FM; the split ranges from 54 to 88 MHz and from 174 to 216 MHz for VHF television).
• Cable stations broadcast over several bands (frequency ranges) in the VHF and ultra high Frequency (UHF) ranges (108 to 174 MHz in the VHF range; 216 to 470 MHz in the VHF and UHF ranges).
• UHF television broadcasts in the UHF range (470 to 890 MHz)
• Radar operates at 10 different bands over a huge frequency range (230 MHz to 3 THz).

Digital Transmission Bandwidths

For digital transmissions, bandwidths range considerably. Here are some examples of bandwidth values for digital transmissions:

• Some digital telephone lines: less than 100 kbps,
• ARCnet networks: 2.5 Mbps,
• ARCnet Plus networks: 20 Mbps,
• Ethernet networks: 10 Mbps,
• Fast Ethernet networks: 100 Mbps,
• Token Ring networks: 1, 4, or 16 Mbps,
• Fast Token Ring networks: 100 Mbps,
• Fiber-optic (FDDI) networks: About 100 Mbps, but can theoretically be several orders of magnitude higher,
• ATM networks: about 655 Mbps, with speeds as high as 2.488 gigabits per second (Gbps) in the future.

Bandwidth also refers to the width, usually measured in Hertz, of a frequency band. It can also be used to describe a signal, in which case the meaning is the width of the smallest frequency band within which the signal can fit. Bandwidth is related to the amount of information that can flow through a channel through the Nyquist-Shannon sampling theorem.

The Bandwidth of an electronic filter is the part of the filter's frequency response that lies within 3 dB of its peak. See below:

On the Internet, a bang path is a series of names that specifies a path between two nodes. A bang path is used in uucp (UNIX-to-UNIX copy program) and sometimes for e-mail (electronic mail) or communications on BITNET. The path consists of domain or machine names separated by exclamation points (!), known as bangs in some computing circles. For example, in a bang path such as hither!thither!yon, hither might be a gateway, thither a computer, and yon a user.

Bang paths go back to the days before automatic routing, because explicit paths were needed when sending to or communicating with another location.

A banner page is output by a printer in a network environment to separate print jobs. A banner page is also known as a job separtor page. Printing of this page is controlled by the network operating system.

A banner page might indicate the name of the user who printed the file and other information. You can eliminate banner pages in NetWare and in most other network operating systems.

With PostScript printers a banner page is only printed after the printer has powered up, and has pasted its POST tests. You can eliminate the banner page by turning it off at the printer or sending down a command from any computer on the network.

A type of network transmission that uses the entire bandwidth of a network to transmit a digital signal. The cables of a baseband network only carry one set of signals at a time. (See broadband, a type of transmission that can send multiple signals simultaneously).

Stated another way, it is a transmission method in which a network uses its entire transmission frequency range to send a single communication or signal. Contrast with broadband.

A fixed project schedule that represents the original plan for the project (including approved changes). The baseline is a yardstick against which the actual project plan is measured to detect deviations. Baselines can take the form of cost baselines, time or schedule baselines, and so on.

The primary function of a local area network is to allow the stations (computers, printers, etc.) that are attached to the network to exhcnage messages. From an architectural standpoint, LANs have been defined in terms of the services that are provided at the two lowest layers of the OSI reference model, the physical layer and data link layer. IEEE Project 802 has defined in detail services to be provided at these levels, and by and large, these services tend to be those that are actually offered by individual LAN implementations. However, these services can be provided in a surprisingly wide variety of ways.

Baud is a measure of signal changes per second in a device such as a modem. It represents the number of times the state of a communication line changes per second. The name comes from the Frenchman Baudot, who developed an encoding scheme for the French telegraph system in 1877.

Baud is no longer used to refer to modem speeds because it does not have a relationship to the number of bits transferred per second. If a modem transferred 1 bit for every signal change, then its bits-per-second rate and baud rate would be the same. However, encoding techniques are employed to make 1 baud or signal change represent 2 or more bits. Two bits per baud is known as dibit encoding and 3 bits per baud is known as tribit encoding.

An internetwork is a collection of subnetworks connected by routers. Routers exchange routing information so they know the current status of the network and how to route packets to their destination. One method for merging router information is the Bellman-Ford distance-vector routing algorithm. It is well defined and used on a number of popular networks. The Bellman-Ford algorithm provides reasonable performance on small-to medium-sized networks. On larger networks, the algorithm can provide slow updates. In some cases, looping occurs in which a packet goes through the same node more than once. Distance-vector routing (DVR) is not suitable for larger networks that have thousands of nodes, dynamic network configurations that require constant updating, and networks that put more focus on routers to handle such tasks as security and congestion management. A more efficient routing protocol is Open Shortest Path First (OSPF).

A standard used to compare computer systems or programs, usually in relation to speed, reliability and accuracy. A point of reference (artifact) to compare an aspect of systems performance (for example, a well-known set of programs). Also, to conduct and assess the computation or transmission capabilities of a system using a well- known artifact.

Basically a computer network can change a group of isolated computers into a coordinated multi-user computer system. A network user can legally share copies of the software with other users if network versions of the software are purchased. Data can be stored in centralized locations or in different locations that are accessible to all users. Printers, scanners, hard-disks, CD-ROM drives, and other peripherals connected to the network are available to all users. Therefore, the biggest benefits for networking is the concept of hardware, and software sharing.

• Hardware Sharing--
• Sharing Hard Disks--Today's sophisticated software applications require large amounts of disk space. As companies require more information about their operation, larger disks are required. Although, the price of disk technology has dropped dramatically in recent years, disks with a capacity to store billions of bytes are still relatively expensive, in the neighborhood of 3,000 to 4,000 dollars. It is not uncommon for microcomputer users to require hard disk capacities of many megabytes. It would be too expensive to purchase large disk space for al users or all possible situations that may arise within a corporation.

Today's networks are based on the concept of sharing access to storage devices. These disks are typically installed on special devices called file servers (see file servers). Sharing disk space has several benefits. The most obvious are:

• Costs
• Integrity of the Data
• Security

Costs are reduced by purchasing hard disks to be shared among all users, instead of purchasing one for each user or location.

The safety of the data is improved over having it on isolated disks, since a network administrator can make constant backups of all files on the device.

Security of the data is enforced by using the network's built-in security systems. Data on isolated disks is an easy target to anyone who wants to damage it.

• Sharing Printers -- Printer sharing is common on networks. Printers can be attached to a file server, or connected to the network independently of the file server. Any user in the network can use any of the printers in the system. Instead of each user having a low-cost printer attached to a terminal or microcomputer, a few high-speed, high-quality printers can be purchased and connected to the network.. Any user that needs a fast printout can send the output to the printer nearest to the station. In addition, other input and output devices can be shared on a network. These include facsimile machines, mono and color scanners, color printers, network interfaces, and plotters.
• Sharing Communication Devices -- Personal computer user on a network often need to access remote systems or networks. One possible solution is to provide them with modems and terminal emulation software to access other systems. This is expensive. Users on a network can share modems, gateways (see gateways), bridges (see bridges), and other network and data communication devices without the need to purchase one for each user.

Therefore, the benefits of sharing hardware on a network are clear:

• Costs can be reduced by avoiding duplicate hardware,
• Users can have access to a variety of devices,
• Data security and safety via network backups, and
• Security measures are easier to maintain.

• Software Sharing -- Instead of purchasing an individual application program for every user in a company, a network version of the program can be obtained. Software designed for networks allows multiple use of the software simultaneously. Users can share the data produced and used by the package. There are many advantages of sharing software. The most important are:

• Cost reduction,
• Legality of the product,
• Sharing data, and
• having current upgrades.

Bridge:Basically it is a device that connects two networks of the same type together (such as two Ethernet networks, or two AppleTalk networks). Thus, bridges are nodes which link together (almost extend) two segments of a network over large distances. Usually, two bridges are linked by a serial line, for example, a leased line or X.25. A separate protocol is used on this line to transport packets from one segment to the other.

Bridges may operate in the network as Internet Routers (B-routers). For LANs (Local Area Networks), usually only so-called MAC (Media Access Control) layer bridges are used to receive data from the link layer of a LAN segment (i.e., Ethernet) and to forward it to another LAN segment (Ethernet).

Therefore, in its basic form, it is a device that links similar networks to each other to allow devices on one network to transmit data to devices on another. Bridges connect only networks that operate under the same communications protocols. However, as explained above, linkages between bridges can be of a different protocol. Compare with gateway, repeater, and router.

NASA-funded high performance operating system for networked workstations.

IP (Internet Protocol) currently uses/provides best-effort service. That is, IP makes every effort to deliver the packets but takes no additional actions when packets are lost, corrupted, delivered out of order, or even misdelivered. In this sense the service provided by IP is unreliable. One may wonder why one would want to build an internetwork to provide unreliable service. The reason is that providing reliability inside the internetwork introduces a great deal of complexity in the routers. The requirement that IP operates over any network places a premium on simplicity.

The design of IP attempts to keep the operation within an internet simple by relegating complex functions to the edge of the network. The connectionless orientation means that the routers do not need to keep any state informtation about specific users or their packet flows. This situation allows IP to scale to very large networks. Similarly, when congestion occurs inside an internet, packets are discarded. The end-to-end mechanisms at the edges of the network are responsible for recovery of packet losses and for adapting to the congestion.

In data transmission and storage, the order in which bytes in a word are processed (stored or transmitted). The term comes from Jonathan Swift's Gulliver's Travels, in which a war is fought over which end of the egg should be cracked for eating. This ordering property is also known as the processor's byte-sex.

In big endian implementations, the high order byte is stored at the lower address. Processor in mainframes (such as the IBM 370 family), some minicomputers (such as the PDP-10), many RISC machines, and also the 68000 family of processors use big endian representations. The IEEE 802.5 (token ring) and the ANSI X3T9.5 FDDI (Fiber Distributed Data Interface) standards use big endian representations. In contrast, the 802.3 (Ethernet) and 802.4 (token bus) standards use little-endian ordering.

The term is used less commonly to refer to the order in which bits are stored in a byte.

COMPARE

Little Endian; Middle Endian

A bitemporal relation is a relation with exactly one slystem-supported valid time and exactly one system-supported transaction time.

Temporal relation, fully temporal relation, valid-time and transaction -time relation, valid-time trnasaction -time relation.

We first discuss the concept; then we discuss the name. In the adopted defintion, "bi" refers to the existence of exactly two times. An alternative definition states that a bitemporal relation has one or more system-supported valid times and one or more system-supported transaction times. In this definition, "bi" refers to the existence of exactly two tpes of times.

Most relations involving both valid and transaction time are bitemporal according to both defintions. Being the most restrictive, the adopted definition is the most descrable: it is the tightest fit, giving the most precise characterization (+E9 -- Names should be precise).

The definition of "bitemporal" is usedas the basis for applying "bitemporal" as a modifier to other concepts such as "query language". This adds more important reasons for preferring the adopted definition.

Independent of the precise definition of "bitemporal," a query language is bitemporal if and only if it supports any bitemporal relation (+E1 -- The naming of concepts should be orthogonal[pertaining to or composed of right angles]. Parallel concepts should have parallel names). (see discussion of Snapshot, Valid- and Transaction- Time and Bitemporal as Modifiers0. With the adopted definition, most query languages involving both valid and transacition time can be charaterized as bitemporal. With the alternative definition, quary languages that are bitemporal under the adopted definition are no longer bitemporal. This is a serious drawback of the alternative definition. It excludes the posibility of naming languages that may be precisely named using the adopted definition. With the alternative definition, those query languages have no (precise) name. What we get is a concepot and name (bitemporal query language) for which there is currently little or no use.

Also, note that a query language that is biterporal under alternative defginition is aslo bitermporal with regard to the adopted definition (but the adopted definition does not provida a precise characterization of this quary language). Thus, the restrictive definition of a bitemporal relation results in a nonrestrictive definition of bitemoral query language (and vice versa).

The name "temporal relation" is commonly used. However, it is also used in a genric and less strict sense, simply meaning any relation with some time aspect. It will not be possible to change the generic use of the term (-E7 [New manes should be consistent with related and already existing and accepted names], and since using it with two meanings causes ambiguity (-E9 [Name should be precise], it is rejected as a name for bitemporal relations. In this respect "temporal relation" is similar to "historical relation".

Next, the term "fully temporal relation" was proposed because a bitemporal relation is capable of modeling both the intrinsic and the extrnsic time aspects of facts, thus providing the "full story." However, caution dictates that we avoid names that are absolute (-E6 [Names should be assigned conservatively. No name is better than a bad name]). What are we going to name a relation that is more general that a temporal relation?

The name "valid-time and transaction-time relation" is precise and consistent with the other names, but it is too cumbersome to be practical (-E2 [Names should be easy to write, that is, they should be short or possess a short acronym, should be easily pronounced (the name or its acronym), and should be appropriate for use in subscripts and superscripts]. Also, it may cause ambiguity. For example, the sentence "the topic of this paper is valid-time and transaction-time relations" is ambiguous.

We choose to name relations as opposed to databases because a database may contain several type of relations. Thus, naming relations is a more general approach.

A raster graphics image, or bitmap, is a data file or structure that consists of a generally rectangular array of pixels, or points of color, on a computer monitor, paper, or other display device. Each pixel has a corresponding red, green, and blue value that combine to determine the colour displayed by that pixel. In this sense, typical raster graphics are said to operate in the RGB color space. This is both the raw format that computer graphics hardware uses to project an image on your monitor, and the basis for many graphics file formats.

A bitmap corresponds bit for bit with an image displayed on a screen, probably in the same format as it would be stored in the display's video memory or maybe as a device independent bitmap. A bitmap is characterised by the width and height of the image in pixels and the number of bits per pixel which determines the number of shades of grey or colors it can represent. A colored raster image (a "pixmap") will usually have pixels with between one and eight bits for each of the red, green, and blue components, though other color encodings are also used, such as four- or eight-bit indexed representations that use vector quantization on the (R, G, B) vectors. The green component sometimes has more bits that the other two to cater for the human eye's greater discrimination in this component.

The quality of a raster image is determined by the total number of pixels (called its resolution), and the amount of information in each pixel (often called color depth). For example, an image that stores 24 bits of color information per pixel (the standard for most high-quality displays in 2001) can represent smoother degrees of shading than one that only stores 15 bits per pixel, but not as smooth as one that stores 48 bits. Likewise, an image sampled at 640 x 480 pixels (therefore containing 307,200 pixels) will look rough and blocky compared to one sampled at 1280 x 1024 (1,310,720 pixels). Because it takes a large amount of data to store a high-quality image, data compression techniques are often used to reduce this size for images stored on disk. Some of these techniques actually lose information, and therefore image quality, in order to achieve a smaller file size. Compression techniques that lose information are referred to as "lossy" compression.

Raster graphics (Bit-maps) cannot be scaled (resized) up without loss of apparent quality (or more accurately, once an image is rasterized, its quality is fixed and cannot improve even on better display devices). This is in contrast to vector graphics, which easily scale to the quality of the device on which they are rendered. Raster graphics work much better than vector graphics, though, for photographs and photo-realistic images. Late 20th century computer monitors typically display about 72 to 96 dots per inch (dpi), while modern printers can resolve 600 dpi or more, so working with images destined for print can be difficult or require large monitors and powerful computers. Monitors with resolutions of 200 dpi were available in late 2001 and higher resolutions are to be expected in future.

To illustrate the matter further - here's the letter "J":

J

Look closely at it... Take a magnifying glass to it if you like, although you may see some chromatic aberration at the edges of the magnifier. You see a "J", the computer sees something more like this, where '.' represents a zero and 'X' represents a one:

....X
....X
....X
....X
X...X
.XXX.

Where you see a zero, the computer instructs its video hardware to paint the current background color. A one calls for the current foreground colour. Yes, it is actually a bit more complicated, but it all basically boils down to one bit or the other making a distinction between the colours of adjacent pixels, which together form an image. This is the basic principle behind drawing on a computer.

In 3D computer graphics, the concept of a flat raster of pixels is sometimes extended to a three dimensional volume of voxels. In this case, there is a regular grid in three dimensional space with a sample containing color information at each point in the grid. Although voxels are powerful abstractions for dealing with complex 3D shapes, they do have large memory requirements for storing a sizable array. Consequently, vector graphics are used more frequently than voxels for producing 3D imagery.

Raster graphics was first patented by Texas Instruments in the 1970s, and is now ubiquitous.

The Because It's Time Network (BITNET) is an international network managed by the Corporation for Research and Educational Networking (CREN). CREN also provides BITNET with information services. BITNET is used mainly for electronic mail services, but offers other services as well. There are gateways between BITNET and the Internet that allow the exchange of electronic mail and other services. BITNET consists of over 2,500 host computers at sites in the U.S., Canada, Mexico, South America, Europe, and Japan. The host sites are typically universities.

BITNET was established in the early 1980s as an outgrowth of the ARPANET. The ARPANET was funded by ARPA (Advanced Research Projects Agency), which later became DARPA (Defense Advanced Research Projects Agency). The ARPANET network linked defense facilities, government research laboratories, and university sites and later evolved into the Internet. BITNET was established as a separate academic network without DARPA funding because DARPA was restricted from providing support outside of the military establishment.

This NetWare 4.X feature maximizes disk space. If there are any partially used disk blocks (usually a block is 8 kilobytes in size), NetWare divides them into 512 byte suballocation blocks for the storage of small files or fragments of files.

BGP is an internet exterior gateway routing protocol that accumulates information about reachability of neighbors from packets as they traverse the network. Route attributes such as the cost or security of a path are also added. BGP reduces the bandwidth required to exchange routing information because the information is exchanged incrementally, rather than by sending the entire database.

A transmission method in which the network's range of transmission frequencies is divided into separate channels and each channel is used to send a different signal. Broadband transmissions is often used to send signals of different kinds simultaneously, such as voice and data. For example, cable television uses the broadband method to carry as many as 100 channels on a single coaxial cable. Contrast with baseband.

In packet based networks, these packets are addressed to all other nodes on a network. Broadcasting offers a way to communicate the same data to every node on a network at once. It is also used for querying purposes. Note that broadcasting results in all nodes on a network receiving the packets regardless of whether a particular node is interested in the contents.

Defines which devices on the Internet gets all of your broadcast messages/datagrams.

Broadcast networks are simple in nature because all information is received by all users, routing is not necessary. A flat addressing scheme is sufficient to indicate which user a given packet is destined to. However, broadcast networks do require a medium access control protocol (MAC) to orchestrate and control the transmissions from the various users. Local area networks (LANs), with their emphasis on low-cost and simplicity hae been based on the broadcast approach.

In broadcast networks a single transmission medium is shared by a community of users. We also refer to these networks as multiple acess networks (MANs). Typically, the information from a user is broadcast into the medium, and all the stations attached to the medium listen to all the transmissions. There is potential for user transmissions interfering or "colliding" which each other, and so a protocol has to be in place to prevent or minimize such interference. The role of the medium access control protocols is to coordinate the access to the channel medium so that information gets through from a source to a destiation in the same broadcast network.

In network traffic, a condition in which packets are broadcast, received, and then broadcast again by one or more of the recipients. The effect of a broadcast storm is to congest a network with redundant traffic. Broadcast storms can arise, for example, in bridged networks that contain loops (closed paths).

In an AppleTalk network that uses the LocalTalk architecture and its LocalTalk Link Access Protocol (LLAP), a transmission sent to each node in the network. Compare broadcast transmission with directed transmission.

A brouter (also known as a bridging router or, less commonly, as a routing bridge) is a device that combines the features of a bridge and a router. A brouter can work at either the data-link layer or the network layer.

Working as a bridge, brouter is protocol independent and can be used to filter local area network traffic. Working as a router, a brouter is capable of routing packets across networks. Other categories to consider are bridge, Internetwork Link, and router.

A short-term decrease in voltage level, specifically when the voltage is more than 20 percent below the nominal RMS voltage. Brownouts can occur when a piece of heavy machinery is turned on and temporarily drains the available power at a site, or when everyone feels the need to run their air conditioners at the same time.

In BSD (Berkeley Software Distribution) UNIX, the layer that represents the API (Application Program Interface) between user applications and the networking subsystem in the operating system kernel.

A UNIX version implemented at the University of California, Berkeley. BSD UNIX introduced several enhancements to AT&T's original implementation, including virtual memory, networking, and interprocess communication support.

The sum total of all approved cost estimates for project activities completed during a particular time period.

The sum total of all approved cost estimates for project activities that are planned or scheduled to be performed during a particular time period.

This is when terminals would generate messages in a bursty manner, that is, the message transmissions would be separated by long idle times. Basically this is a mode of transission for early terminals to a central computer or mainframe.

In general, bus mastering is a bus-access method in which a card or device takes control of the bus in order to send data onto the bus directly, without help from the CPU (central processing unit). In a network, the network interface card takes control of the bus.

Generally, MCA (Microchannel Architecture) and EISA (Extended Industry Standard Architecture) machines support bus mastering, but ISA (Industry Standard Architecture) machines do not. VL (VESA Local) and PCI (Peripheral Component Interconnect) buses also support bus mastering.

Bus mastering can improve throughput considerably, but only if the board and the computer support the same bus-mastering method, and if the bus mastering doesn't conflict with the hard-disk controller.

Several types of transfer modes are possible with bus mastering, including burst mode, streaming data mode, and data duplexing. A particular bus-mastering scheme may support some or all of these modes.

A layout scheme in which devices on a network are connected (via taps) along the length of a main cable (which carries the network's signals), or bus, rather than in a daisy chain or loop. This is also known as a backbone cable. This network configuration is also known as a tree topology. Therefore, data may pass directly from one device to another without the need of a central hub. With some applications, however, the data must first be moved in and out of a central controlling station.

In the typical implementation of the bus configuration, all nodes on the bus have equal control. One end of the bus is the head end. The head end returns the message back into the bus traveling in the opposite direction. Most personal computer networks use the bus topology.

The reliability of bus networks is good unless the bus itself malfunctions. Losing one node does not have an effect on the rest of the network.

The most commonly used controls are scroll bars and buttons. Buttons are usually part of a dialog or alert, but scroll bars are frequently used in applications windows t enable users to move and scroll over a document and need to be defined separately from any dialog. The button function returns TRUE if the mouse button is currently down, and FALSE if it isn't.

Buttons cause an immediate or continuous action when clicked or pressed with the mouse. They appear on the screen as rounded-corner rectangular objects with a title centered inside.

Usually called after a mouse-down event, StillDown tests whether the mouse button is still down. It returns TRUE if the button is currently down and there are no more mouse events pending in the event queue. This is a true test of whether the button is still down from the original press -- unlike button, which returns TRUE whenever the button is currently down, even if it has been released and pressed again since the original mouse-down event.

Radio buttons also retain and display an on-or-off setting. They're organized into groups, with the property that only one button in the group can be on at a time. Clicking one button in a group both turns it on and turns off the button that was on, like the buttons on a car radio. Radio buttons are used to offer a choice among several alternatives. On the screen, they look like round check boxes; the radio button that's on is filled in with a small black circle instead of an "X".

A collection of -- usually eight -- bits (but rarely worth a dollar anymore). A byte generally represents a character or digit of eight bits.

In networking, a situation in which a node fails by behaving incorrectly or improperly, rather than by breaking down completely and disappearing from the network. A network that can keep working even if one or more nodes is experiencing Byzantine failure has Byzantine robustness.

Creating redundant computers has been greatly helped by better analysis techniques. There are proof techniques that allow pruning of the unworkable failure trees by assuming so called "Byzantine" failure models. These techniques allow strong statements to be made about the redundancy properties of designs. The heuristic part is trying to verify the absence of "common-mode-failures", or failures in which several redundant and supposedly independent components fail at the same time for the same reason.

A byzantine failure is one in which the failed component does the worst possible thing to the system. It is as if the component were possessed by a malign intelligence. The power of the technique is that it lends itself to certification, at least within the confines of well-defined models.

This is in relation to Byzantine Faults and Failure Models.

From L. Lamport, R. Shostak and M. Pease. ACM Transactions on Programming Languages and Systems, 4(3):382-401, July 1982

Notes by Indranil Gupta, March 08, 1999; Cornell University.

Adapted from:

• original notes by Xun Wilson Huang.
• original notes by Lawrence Kesteloot.

Thanks to: Xun Wilson Huang, Lawrence Kesteloot.

Problem Statement: Byzantine Generals Problem (BGP)

The setting: There are n generals, one of them the commanding general. Generals can send (and receive) messages from other generals.

The problem: Develop a communication protocol for the commanding general to send an order to the n-1 lieutenant generals so that:

1. All loyal lieutenants obey the same order
2. If the commanding general is loyal, every loyal lieutenant obeys the order he sends.

Adversary: Any of the generals could be traitors i.e., could send inconsistent messages regarding the order to other generals.

Impossibility Results:

• For n = 3 generals and 1 traitor, there is no solution (protocol). This is because a loyal lieutenant cannot distinguish who is the traitor when he gets conflicting information from the commander and the other lieutenant. Let's call this the 3-Generals Problem.
• BGP for n < 3m+1 generals and m traitors can be reduced to the 3 - generals problem, with each of the Byzantine generals simulating at most m lieutenants and taking the same decision as the loyal lieutenants they simulate. Thus BGP for n < 3m+1 and m traitors is not solvable.
• Reaching approximation is as hard as reaching agreement.

A solution with oral messages for n > 3m:

A solution for BGP with n > 3m and upto m traitors, is given.

Oral message system properties:

1. Answer 1. Every message that is sent is delivered correctly. -> No message loss.
Answer 2. The receiver of a message knows who sent it. -> Completely connected network with reliable links(due to A1).
Answer 3. The absence of a message can be detected. -> Synchronous system only.

Every general can send a message to every other general.

Solution in brief:

• uses a function majority which takes in a set of values and returns the value that is the majority among them (a possible implementation - median of the values).
• uses 'rounds' in each of which a general broadcasts the value he has received in the earler round to all the other generals through whom the value has not passed before he received it.
• when returning from the round, for each j, any two loyal lieutenants receive the same vector of values {v1, ... v(n-1)}. As the majority of the loyal lieutenants' values in these is ensured, applying the majority function on {v1, ... v(n-1)} to obtain vn preserves the above fact (that any two loyal lieutenants receive the same vector of values {v1, ... vn}). This ensures that BGP is solved.

Note: If the commander is not a traitor, we can be done in 2 rounds. If the commander is a traitor, you may need upto m+1 rounds.

A solution with (unforgable) signed messages:

The difficulty of BGP is in the ability of a traitor lieutenant to lie about the commander's order. If we can restrict this ability by making the following assumptions, BGP is solvable with any number of traitors as long as their maximum number is known.

Signed messages:

4. Answer 4. In addition to the 3 assumptions made in the solution with oral messages, we add the following assumption.
1. A loyal general's signature cannot be forged, any alteration can be detected. -> can drop a message, but can't change it
2. Any one can verify the authenticity of a signature. -> no one can fool a general

Again, assume a fully connected message graph among the generals.

Solution in brief:

Uses a majority-like function called choice.

The solution:

• the commander sends a signed order to lieutenants
• if a lieutenant receives an order from some one (either from commander directly, or from other lieutenants), he verifies it and then puts it in a set V if it's not already there. Relay the order if there are less than m distinct signatures on the order.
• Everyone halts at round m+2, and use choice(V) as the desired action

The algorithm is to make all loyal lieutenants keep the same set of V, thus choice(V) is the same. If the commander is loyal, the algorithm works because all loyal lieutenants have the correct order by round 1 and by unforgablity no more orders can be produced. If the commander is not loyal, by running the algorithm to round m+1, at least one loyal lieutenant will get the order before round m( because there are only m traitors). And that loyal lieutenant will send it to all others. In short, if one loyal lieutenant gets an order, all loyal lieutenants will get it in the next round.

Relaxing the assumption on full-connectivity of the generals graph - extending above solutions:

The previous 2 solutions can be extended to relax the assumption that the message graph among the generals is fully connected.

• Oral messages: Solution with oral messages is extended to solve BGP with upto m traitors in a p-regular graph with m>0 and p>3m-1.
• Unforgable messages: Earlier solution with signed messages solves BGP with upto m traitors in (m+d-1) rounds, where d is the diameter of the subgraph of loyal generals. Assumption here: subgraph of loyal generals is connected (this can be relaxed by relaxing the problem statement of BGP)

Practical use of BGP in building real life systems:

The best way to provide faul-tolerant decision-making in redundant systems is by majority voting. A faulty input device may generate meaningless inputs, but majority voting would ensure that the same meaningless values are used.

For majority voting to yield a reliable system, the following 2 conditions must be satisified --

1. All non-faulty processors must use the same input value
2. If input unit is non-faulty, then all non-faulty processes use the value it provides

But these are just the requirements of the BGP!

So we can apply the above solutions to the BGP in real-life. Now what about the practicality of the assumptions made by those solutions?

• About Answer 1: In real life, link failures occur. However, link failures are indistinguishable with failures of processors, therefore we can count the link failures as one of the m. Signed message is insensitive to link failures because no message can be forged even if links fail.
• About Answer 2: What is actually required is that no traitor can forge a non-faulty process' message. Answer 2 not needed in the solution with signed messages.
• About Answer 3: In an asynchronous system, this condition cannot be satisfied. It is usually implemented via time-outs.
• About Answer 4: Signing message has 2 aspects:
• If processor is non-faulty, then no faulty processor can generate S(M). This can never be solved in real-life - only its probability of failure reduced.
• Given M and X, any one can verify if X == S(M). This is doable in real world.

Further Observations:

• Optimizations for the BGP solution.
• combine messages to reduce the total number of messages.
• reduce the amount of information transferred.

• BGP required in the most general undecidable case of process failure.
• Solution presented is optimal because Fischer and Lynch have proved that any solution to the BGP necessarily has each lieutenant wait for a message that has passed through the hands of at least m generals after the commander.
• Solutions to clock synchronization (needed for the implementation of above BGP solutions) - very similar to the solutions for BGP.
• Further impossibility results
• BGP with messages transmitted arbitrarily quickly with upper bound on message transmission delay.
• Consensus with restricting traitors to fail-crash only.

• BGP works but is inherently expensive, especially in terms of the number of messages O(m!). So it's a trade-off between performance and reliability. If you want more reliability in the most general failure conditions, you have to settle for a (costly) BGP solution. If, however, you can relax the failure conditions in your systems (ex. assume only fail-crash processes in a synchronous system and leave it to God to ensure that), you can go for cheaper solutions.

Critique and Questions:

• Graph connectivity. Are p-regular topologies that frequent? Can we extend the BGP solutions to any network topology? Has it been extended to any other topologies?
• Value of m: How would one obtain a reasonable value for maximum m in a practical system (note that this maximum number is required even in the solution with signed messages).
• Synchronous/asynchronous systems: How many synchronous system do we really use (SMP machines, and?) How about asynchronous systems?
• Further work after this paper:
• What other solutions to BGP have been proposed after this paper?
• Has any attempt been made to extend the BGP solutions to asynchronous systems to ensure 'some degree/probability' of reliability?

• Bounds on best possible BGP solution (in terms of messages)?

Resulting thoughts on the Byzantine General Problem:

Thus -- Reliable computer systems must handle malfunctioning components that give conflicting information to different parts of the system. This situation can be expressed abstractly in terms of a group of generals of the Byzantine army camped with their troops around an enemy city. Communicating only by messenger, the generals must agree upon a common battle plan. However, one or more of them may be traitors who will try to confuse the others. The problem is to find an algorithm to ensure that the loyal generals will reach agreement. It is shown that, using only oral messages, this problem is solvable if and only if more than two-thirds of the generals are loyal; so a single traitor can confound two loyal generals. With unforgeable written messages, the problem is solvable for any number of generals and possible traitors.

Search for Information Technology Items

Return back to Network & Concepts Index

Networking "B" Definition and Concepts

This site is brought to you by
Bob Betterton; 2001 - 2011.

This page was last updated on 09/18/2005
Copyright, RDB Prime Engineering