Sams Teach Yourself MCSE Windows NT Server 4 in 14 Days
(Publisher: Macmillan Computer Publishing)
Author(s): David Schaer, et al
ISBN: 0672311283
Publication Date: 12/15/97

Appendix A

Glossary

ACL (Access Control List)

Every object in Windows NT has a list, called the Access Control List, which lists the Security IDs (SIDs) that have rights to that object, as well as which rights each SID has. See SID.

ActiveX

A Microsoft trademark for object services and interchangeable software controls, including COM and embedded Internet software components.

Alpha

A series of processors made by Digital Equipment Corporation that, at the time of this writing, held the distinction of being the fastest processors in the world. Alternatively, the first test cycle of an application in development.

API (Application Programming Interface)

A set of functions that gives application programmers access to common system functions. The native API in Windows NT is the Win32 API, although NT also supports subsets of the MS-DOS, Win16, OS/2, and POSIX.1 APIs through subsystems.

AppleTalk

A network protocol stack for Macintosh computers and printers. Originally designed to work only over Apple’s serial cable technology (see LocalTalk), AppleTalk now is supported over EtherNet and Token Ring. Although it is traditionally a relatively chatty and inefficient protocol, most recently AppleTalk can be carried over TCP/IP, greatly improving its throughput. Windows NT’s Services for Macintosh, however, does not yet support AppleTalk/IP.

ARP (Address Resolution Protocol)

A protocol TCP/IP hosts use to discover the hardware address of a destination node when only the IP address is known. Windows NT includes a command-line tool for examining the system’s local ARP cache; for more information, type arp -? at a command prompt.

ARPA (Advanced Research Projects Agency)

A government agency responsible for the development of robust internetworking technologies that led to the ARPANET, a predecessor to today’s Internet.

ATM (Asynchronous Transfer Mode)

An advanced packet-switching networking technology that provides phenomenal throughput for data traffic, particularly for the intense needs of videoconferencing and other multimedia tasks.

auditing

The process of tracking access to objects. Auditing can track success and failure on a per-object, per-user basis. Auditing most often is used to help trace security problems and breaches, although it also can be useful for tasks such as billing for printing services.

authentication

The process of identifying oneself to a security provider. NTLM (NT/LAN Manager) authentication, the current standard method supported by Windows NT, generally associates a user object’s SID with a username and password.

b-node

Short for broadcast node, a method for NetBIOS name resolution over TCP/IP. B-node resolution uses NetBIOS broadcasts over TCP/IP to resolve names. See h-node, m-node, and p-node.

b-tree

Short for balanced tree structure, a b-tree is a data structure used to represent directories and files in file systems such as HPFS and the Macintosh HFS (Heirarchical File System).

BDC (Backup Domain Controller)

An NT Server that maintains a copy of the domain’s SAM (Security Accounts Manager) database. BDCs provide an alternative source for authentication services in a domain, as well as a backup copy in case the domain’s PDC (Primary Domain Controller) suffers catastrophic failure. See PDC, SAM.

bindery

The authentication database on a Novell NetWare 2.x or 3.x server. See bindery emulation.

bindery emulation

A NetWare Directory Services (NDS) server can use bindery emulation to allow non-NDS–aware NetWare clients to authenticate to the server.

bridge

A device that enables two or more networks, each using different types of data link protocols (such as Token Ring and EtherNet), to form a single logical network.

browser

A NetBIOS client that attempts to contact other NetBIOS resources on a network using the browse list maintained by one or more master browsers. The browse list contains a list of active systems, identified by their NetBIOS network names. See master browser and network browser.

bus

An architecture for communication between multiple devices, whether on a system board or across a network. A bus architecture is a shared medium in which all devices on the bus must contest for communication time, and all devices can hear all transmissions.

bus mastering

A process by which a device can gain temporary control of the bus to direct data transfer, thus reducing processor overhead during I/O operations.

caching

The process of storing commonly used data from one data storage medium in another, faster medium.

CHAP (Challenge-Handshake Authentication Protocol)

An encrypted password authentication protocol, for use with dial-up connection protocols such as RAS and PPP.

CISC (Complex Instruction Set Computing)

A philosophy for CPU design which states that the processor’s instructions should be very powerful, allowing large tasks to be completed in fewer instructions. The Intel x86 family of processors is an example of a CISC architecture, although later models also have included RISC innovations. See RISC.

Class (A, B, C)

A type of IP address that determines the range of digits available to a particular installation. A Class A number is preallocated only the farthest-left octet of the dotted-decimal IP address, leaving the farthest-right three fields free for suballocation (for example, 110.0.0.0). Similarly, a Class B address is preallocated the farthest-left two octets (133.129.0.0), and a Class C address is preallocated the farthest-left three octets (205.122.89.0). A Class A number can have a maximum of 16,777,214 nodes, a Class B number 65,534, and a Class C 254. IP network numbers are currently assigned by the InterNIC.

coaxial

A type of cable used for data transmission and networking. Coaxial cable can support long lengths due to its shielding but is less cost-effective and harder to maintain than twisted-pair. See STP.

collision

On a CSMA/CD network, a collision is a jamming signal sent when multiple workstations attempt to access the transmission bus at the same time. When a collision occurs, all transmitting workstations wait a random amount of time before re-attempting transmission. See CSMA/CD.

COM (Component Object Model)

A model developed by Microsoft to develop, register, track, and use interchangeable software components on computer systems. COM is part of the larger product ActiveX. See ActiveX.

connectionless

A connectionless network protocol is responsible only for the transmission of data and has no role in the proper ordering, collection, or verification of data receipt. Data link protocols such as EtherNet and Token Ring, as well as transport protocols such as IP are examples of connectionless protocols. See connection-oriented.

connection-oriented

A connection-oriented protocol contains facilities for ordering, verifying, and retransmitting data to guarantee delivery (or report an error if delivery cannot take place). Transport protocols such as TCP are examples of connection-oriented protocols.

console

A console application runs from the command prompt and uses basic text input and output methods rather than the Windows Graphical User Interface. Alternatively, it is the physical location of a system.

CSMA/CD (Carrier Sense-Multiple Access/Collision Detection)

A bus contention communication method used by network technologies such as EtherNet. In a CSMA/CD network, workstations first attempt to determine whether a transmission is taking place (carrier sense) before trying to transmit. When the wire is free, any workstation can attempt to transmit packets (multiple access). If, however, multiple workstations attempt to transmit at the same time, a collision occurs (collision detection), and the workstations wait a random time to retry communication. CSMA/CD networks are not suitable for real-time data transmission because no workstation is guaranteed the capability to transmit its packets, but CSMA/CD networks have proven simple and reliable enough for most data transmission needs.

CSNW (Client Service for NetWare)

A network client included with Windows NT that enables the workstation to connect to NetWare resources. The client included with NT 4.0 enables the workstation to authenticate to NDS resources as well as bindery resources.

data link layer

The second level of the OSI network model, this layer is responsible for the formation and transmission of packets over layer one, the physical layer. The data link layer has no facility for connection-oriented communication. EtherNet and connectionless, and connection-oriented.

DDE (Dynamic Data Exchange)

A protocol for inter-process communication on Windows systems.

debugger

A tool for examining source code to determine the point of failure in an application, or to examine the status of a system to determine where a fatal system error has occurred.

DHCP (Dynamic Host Configuration Protocol)

A method for dynamically allocating IP addresses. DHCP can be run on servers such as Windows NT, UNIX, and NetWare, and can dynamically provide IP addresses and stack configuration to several different network clients, thus avoiding the process of manually configuring the stack on each workstation. DHCP supports the leasing of addresses to specific machines for set periods of time.

differential backup

A backup method whereby only the files modified since the last full backup are copied to tape.

DirectX

A series of services for providing high-performance access to video, network, input, sound, and other services. Primarily for game developers.

DLC (Data Link Control)

A non-routable network protocol used primarily by Hewlett-Packard JetDirect print servers, as well as IBM mainframes.

DLL (Dynamic Link Library)

A library file containing functions that can be used by many different programs, allowing those programs to be smaller and to load into memory only the shared execution code when needed.

DMA (Direct Memory Access)

A process that enables data to be moved from a device across the system bus to memory, or vice-versa, with minimal processor attention, thus greatly increasing system throughput.

DNS (Domain Name Service)

The standard for resolving IP addresses to canonical host names on TCP/IP systems. DNS relies on static tables that map addresses to host names, although standards for Dynamic DNS, designed to be used in conjunction with DHCP and NetBIOS naming, are under development. See WINS.

domain

A logical structure that allows for centralized administration of user and group objects and resources, as well as centralized authentication and rights management. A domain must have at least one Windows NT Server system serving as a domain controller. See workgroup.

domain master browser

A system that maintains a master NetBIOS browse list for an entire domain.

dotted decimal

The name for an IP address, rendered in four octets and represented in decimal form. For example, 123.14.5.115.

driver

An application or file that provides an abstracted interface to low-level hardware functions, allowing higher-level operating systems and application programs to use the same functionality independently across devices.

duplexing

The process of maintaining the same data on multiple disks across multiple disk controllers, protecting against data loss due to disk or controller failure. Alternatively, the process of printing on both sides of a sheet of paper. See RAID.

EISA (Extended Industry Standard Architecture)

An enhanced version of the ISA bus standard that uses a 32-bit bus with bus mastering to transfer data at up to 66MB per second while maintaining backward compatibility with ISA cards. EISA is a relatively rare bus standard, except in server-class systems, where it can still be commonly found. EISA bus devices can be configured automatically by using software utilities. See ISA, micro channel, PCI.

election

The process of determining, based on values such as operating system version and system response time, which system will serve as master browser. See master browser, browser.

environmental subsystem

A Win32 application that provides an emulated version of another operating system environment. Examples include the WOW (Windows-on-Windows) Win16 Subsystem, the OS/2 Subsystem, and the POSIX Subsystem.

EtherNet

A system for CSMA/CD networking, originally developed by Xerox in the 1970s and now adopted in many implementations as an industry standard data link layer networking protocol. See CSMA/CD.

extent

A portion of a file spread across a block or group of disk blocks. See fragmentation.

FAT (File Allocation Table)

A file system, originally developed for MS-DOS, that is still used by DOS, Windows 95, and Windows NT systems. FAT indexes all files and extents in a table called the FAT table. FAT requires little system overhead, making it simple and fast to implement, but it includes little fault tolerance, wastes disk space with inefficient block allocation, and suffers heavily from disk fragmentation. See NTFS.

fault-tolerant

A scheme that is not subject to data loss or system downtime due to the failure of one or more hardware devices.

FDDI (Fiber Distributed Data Interface)

A fiber-optic network technology for high-speed WAN interconnectivity that uses a token-based access method. FDDI is commonly used for high-speed, fault-tolerant interconnection between LANs in a Wide-Area Network.

fiber

A lightweight thread, lacking many thread features such as automatic processor scheduling. Fibers are provided by Windows NT to ease the porting of applications from other multitasking operating systems with weaker threading support. See thread.

finger

A protocol for checking to see whether a user is logged into a multiuser system over TCP/IP, as well as viewing some user-specified information (referred to as a plan). Finger can be a serious security hole on many shared-time network operating systems.

FPNW (File and Print Services for NetWare)

An add-on tool for Windows NT Server that enables the server to emulate a NetWare Bindery server in order to provide file and print services to NetWare Bindery client workstations. This tool now is included in the Services for NetWare package from Microsoft.

fragmentation

The process of spreading the extents of a file across the free space on a hard disk rather than placing them in contiguous blocks. Fragmentation generally occurs when a file system inefficiently plans space for file saves and the growth of existing files. Different file systems are affected differently by fragmentation. Fragmentation also can take place within the virtual memory address space of an operating system, but does not seriously affect operating systems with efficient memory managers.

frame type

One of several different standards for the layout of an EtherNet frame, based on the type of protocols the frame can carry. Examples of frame types are 802.2, 802.3, and Ethernet II.

FTP (file transfer protocol)

A protocol for performing file downloads and uploads over TCP/IP, supporting both raw binary and ASCII file transfer. See TFTP.

gateway

A device that provides access to outside networks. A gateway can be a dedicated hardware device, such as a network router; or a software service, such as an SMTP gateway for Microsoft Mail.

GSNW (Gateway Service for NetWare)

A service for Windows NT Server that enables NT clients to access remote NetWare hosts transparently, with all actual NetWare NCP traffic taking place only between the gateway server and the NetWare resources. GSNW can provide convenient access to NetWare resources without requiring NetWare client installation on NT clients but can act as a performance bottleneck, as well as making it simple to violate NetWare licensing restrictions.

h-node

Short for hybrid node, a method for NetBIOS name resolution over TCP/IP. A client configured to use h-node resolution first will attempt to use p-node resolution; if that fails, it then will attempt b-node resolution; finally, if all else fails, it will check the local LMHOSTS file for addressing information. See b-node, p-node, m-node, and LMHOSTS.

HAL (hardware abstraction layer)

A software interface that handles all hardware-specific programming code, allowing an operating system to be written in non-system–specific high-level languages, thus minimizing system dependency. This allows for simpler porting to other platforms, as well as better modularity and stability.

hive

A root-level subtree in the registry that contains keys and values.

I/O port

A unique memory address through which input and output to a hardware device takes place.

IDE (integrated device electronics)

A standard interface for mass storage drives on PC systems that is based on the original PC hard disk controller interface. EIDE (Enhanced IDE) improves still further on the IDE model.

incremental backup

A backup method that backs up all files that have been modified since the last backup. Unlike differential backup, however, an incremental backup sets the archive file for files it backs up.

infinite loop

A classic programming error in which a program enters a looping situation from which it cannot exit, causing the application to run in a loop until forced to quit by the user or operating system. See infinite loop.

interrupt

A signal to the system’s CPU to gain processor time to execute an I/O or other function. Operations that generate excessive interrupts can negatively impact system performance. Interrupts can be signaled either by hardware access or by software command.

IPX/SPX (Internet Packet Exchange/Sequenced Packet Exchange)

A LAN protocol designed by Novell Corporation for use with NetWare, IPX/SPX is a routable protocol that occupies the Transport and Network layers of the OSI model. IPX/SPX provides dynamic address configuration and high-speed networking, at the cost of chattiness and the incapability to route packets beyond a set hop limit. IPX/SPX can be used for NetBIOS communications.

IRQ (Interrupt Request Queue)

A dedicated queue over which a specific hardware device can interrupt the CPU.

ISA (Industry Standard Interface)

The bus designed for AT-class systems, still commonly in use today. ISA buses have a 16-bit data path, with a maximum theoretical transfer rate of 16MB/sec, without bus mastering. ISA buses do not, by default, have the capability to perform automatic adapter configuration, although the plug-and-play standard, on a system with a plug-and-play BIOS, does allow for automatic configuration of those devices that support the standard.

ISDN (Integrated Services Digital Network)

A high-speed service that can provide data rates much larger than those available over standard telephone lines, as well as the capability to use some bandwidth to support simultaneous phone and data connections. ISDN is slower than most LAN technologies but faster than standard modem communications.

kernel

The core of an operating system, the kernel provides essential services. Operating systems such as NT use a so-called microkernel, which provides only very basic services such as thread scheduling, for the sake of efficiency and stability.

key

A subfolder in the registry that can contain values. Also, a unique security token identifier.

LAN (Local-Area Network)

A network for allowing computers to communicate with one another. See WAN.

LANMAN(LAN Manager)

A term referring to network standards, such as Windows NT’s current model, which are based on the LAN Manager product developed by Microsoft and IBM in the 1980s. LANMAN networking will be phased out in Windows NT 5.0, though it still will be supported.

lease

The period during which a client is associated with an IP address using DHCP. See DHCP.

LMHOSTS(LAN Manager Hosts)

A static list of NetBIOS names and their corresponding IP addresses. The use of LMHOSTS files predates the availability of the WINS service. See WINS.

LocalTalk

A standard for serial communications developed by Apple Computer. LocalTalk allows for simple network communications between Macintosh workstations and devices such as modems and printers; however, its slow data rate makes it unsuitable for LAN networking.

LPC (Local Procedure Call)

A facility for passing messages between client and server processes on a system. See RPC.

LPD (Line Printer Daemon)

A protocol for serving printers over TCP/IP originally developed for UNIX systems.

LSL (Link Support Layer)

Part of an ODI network stack, the LSL provides an interface between network device drivers and upper-layer networking protocols.

m-node

A method for NetBIOS name resolution over TCP/IP. A client configured to use m-node resolution first will attempt to use b-node resolution; if that fails, it will then attempt to use p-node resolution. See b-node, h-node, p-node, and WINS.

MAC (Media Access Control)

The unique hardware address of an Ethernet workstation; every Ethernet interface has a unique MAC address.

master browser

A system that maintains the list of all NetBIOS resources for a particular entity, whether a subnet or a domain. Browsers contact the master browser for the list of available NetBIOS resources, and master browsers are determined via the election process. See domain master browser, election, and browser.

master domain

A domain that provides user authentication services for subsidiary resource domains. See domain.

micro channel

A proprietary bus architecture designed by IBM that uses 32-bit width and automatic adapter configuration via software configuration files. Micro channel supports bus mastering and can support communication rates ranging from 10MB/sec to 160MB/sec. Micro channel no longer is common in PCs but can be found in other IBM computer systems.

MIPS

A series of processors made by the company MIPS (a subsidiary of Silicon Graphics) using RISC architecture. Future versions of Windows NT no longer will support MIPS systems. Also, an acronym for Millions of Instructions Per Second, a method of rating a processor’s speed. MIPS measurements do not include data on how much work can be done in every instruction cycle, but merely list, theoretically, how many instructions the processor can perform each second.

mirroring

The process of maintaining multiple copies of data across multiple disks, thus reducing the risk of data loss due to a single disk’s failure. See RAID.

multihoming

The process of placing a system on multiple subnets by using more than one network interface card. A multihomed system can perform services such as protocol routing.

multitasking

The process of switching system resources between processes and threads, thus seeming to allow multiple tasks to take place at once.

NCP (NetWare Core Protocol)

The protocol used for NetWare file, print, bindery, and NDS services.

NDIS (Network Driver Interface Specification)

The network driver model used by LANMAN and NT-based network clients, such as Windows for Workgroups, Windows 95, and Windows NT. See ODI.

NDS (NetWare Directory Services)

The replacement for the bindery, NDS uses an advanced X.500-based directory service to maintain a tree-hierarchy database of all network resources, including users, groups, servers, volumes, and printers, for ease of management and administration. NDS also provides distributed authentication and security management, as well as fault tolerance. NDS requires higher hardware and network overhead than other network implementations, but provides functionality not found in many other operating systems (although NT 5.0 promises the active directory, a similar service). NDS is based on pioneering directory services such as Banyan’s StreetTalk, a part of its VINES UNIX-based network operating system, and acts as the model by which current and future directory services (such as the Active Directory) are judged.

NetBEUI (NetBIOS Enhanced User Interface)

Originally an enhanced version of the NetBIOS protocol designed by Microsoft and IBM, NetBEUI is now a highly efficient but unroutable protocol that replaces the original role of NetBIOS. NetBEUI can be used for NetBIOS networking in small, single-subnet workgroups, and provides efficient throughput and dynamic addressing. See NetBIOS.

NetBIOS (Network Basic Input/Output System)

Originally a subnet-only network protocol designed by IBM in the early 1980s, NetBIOS is now a standard for network APIs that can be transported over other protocols, such as NetBEUI, IPX/SPX, and TCP/IP. See NetBEUI, IPX/SPX, and TCP/IP.

NetWare

A Novell product for providing LAN file, print, and security services.

network browser

The service that implements NetBIOS browsing functionality on NT systems.

NFS (Network File System)

A UNIX standard that uses TCP/IP to mount remote file systems over the network as though they were local file systems.

NIC (Network Interface Card)

A generic term for any device, whether an actual adapter card or simply a port on a gateway device, which connects a computing system to a network.

NOS (Network Operating System)

An operating system with integrated support for networking services.

NTFS

The Windows NT File System. NTFS improves upon older file systems by providing highly efficient block sizing, file- and directory-level ACL security, fault tolerance via transaction logging, support for multiple streams within files, and advanced caching. NTFS uses more system overhead than FAT, but provides better reliability and higher performance on systems with enough resources to take advantage of its features. See FAT.

NWLink

Short for NetWare Link, the services and protocols provided by Microsoft to provide IPX/SPX support and communication with NetWare bindery services.

ODI (Open Data Interface)

A standard for implementing a network protocol stack that is used by NetWare and other network stacks.

OLE

Originally Object Linking and Embedding, OLE provides distributed object communication services. OLE is now a part of the larger service known as COM. See COM and ActiveX.

OSI (Open Systems Interconnection)

A reference model for implementing network protocol stacks. No real-world network stack is entirely OSI-compliant, although the OSI model remains a standard for describing how network stacks function.

p-node

Short for point-to-point node, a method for NetBIOS name resolution over TCP/IP. A client using p-node resolution will attempt to resolve NetBIOS names using a WINS server only. See b-node, m-node, h-node, and WINS.

pagefile

A file used by a virtual memory system to hold memory pages that temporarily are not being used by active processes, but which can be swapped back into RAM when called for. Also known as a swap file.

PAP (Password Authentication Protocol)

A password authentication protocol for use with dial-up connection protocols such as RAS and PPP.

PCI (Peripheral Component Interconnect)

A system bus designed by Intel but administered as a current industry standard, using 64-bit data width, with automatic adapter configuration and bus mastering. It supports data rates of up to 264MB/sec, although most PCs currently cannot transmit data across the bus at this rate.

PDC (Primary Domain Controller)

An NT Server that maintains the primary copy of a domain’s SAM database. A PDC is required to create an NT domain, though some authentication services can be provided by backup domain controllers (BDCs). See BDC.

physical layer

The lowest layer of the OSI model, this layer refers to the actual physical cabling, as well as the method, whether electrical, optical, or otherwise, for transmitting digital data in a network.

POSIX (Portable Systems Interface for Computing Environments)

A standard for developing applications that can run across multiple operating systems without modification, needing only a recompilation of executable images. Windows NT supports the POSIX.1 standard, meaning that applications written to the POSIX.1 standard can run under Windows NT, requiring only a recompilation.

PowerPC

A series of processors made by IBM and Motorola that are used most commonly in Macintosh computers. Future versions of Windows NT no longer will support PowerPC systems.

PPP (point-to-point protocol)

A protocol designed to allow the encapsulation of other protocols, such as TCP/IP, over a modem connection.

process

A separate unit of application memory space, resources, and security access; each application has a process. A process consists of at least one thread of execution. See thread.

protocol

A standard for communication, defining rules for transmitting data in a fashion understandable to all parties supporting that protocol.

RAID (Redundant Array of Inexpensive Disks)

A system of standards for enhancing the performance and reliability of disk subsystems by using multiple disk drives or disk controllers.

RAS (Remote Access Service)

A dial-up server service that enables NT systems to serve connections over telephone lines, ISDN, or X.25 networks. RAS supports multiple dial-up protocols, including PPP, and can work using either multiport serial boards or via dedicated modem switching hardware.

redirector

A component that determines whether requests for mapped resources should be sent to the local file system or to the network stack, thus making access to those resources transparent to applications.

registry

A database that contains information used by the operating system and its applications, ranging from dynamic hardware configuration and system runtime information to application customization settings.

replication

The process of maintaining identical copies of a file system or parts of a file system on multiple NT systems to provide multiple access points and fault tolerance.

RISC (Reduced Instruction Set Computing)

A philosophy for CPU design which states that computing is best done by breaking up programs into small, efficient, low-functionality instructions that can be executed at very high clock rates. Examples of modern RISC chips include the DEC Alpha, the MIPS line of processors, and the PowerPC. See CISC.

router

A computing device designed to route network traffic from one subnet to its destinations. Any system with two or more NICs and the proper software can act as a router, though systems designed and dedicated to network routing are far more efficient and cost-effective.

RPC (Remote Procedure Call)

A facility for passing messages between client and server processes on systems over a network. See LPC.

SAM (Security Accounts Manager)

A component of Windows NT that maintains the list of security IDs for user and group objects and performs the authentication and verification for those objects. This term also refers to the database of users and groups for a system or a domain. See SID.

SCSI (Small Computer Systems Interface)

A bus architecture for mass storage and other devices, commonly used in servers and other systems where performance and extensibility are important.

server

An NT (or other) system designed and tuned to provide centralized network services, such as network authentication, file and print, or application services. Alternatively, a version of the Windows NT product, which can act as a domain controller, designed specifically to provide these types of services. Alternatively, the service that enables clients to connect to an NT system’s shared resources. See workstation.

SID (Security ID)

A randomly generated number that represents a user, group, or other security object in a SAM database. See SAM.

SMB (Server Message Block)

The protocol used by NT for file, print, and other services. See NCP.

SMP (Symmetric Multiprocessing)

The capability of an operating system (such as Windows NT) to dynamically schedule individual threads on any processor on the system.

SMTP (Simple Mail Transfer Protocol)

The protocol most commonly used for Internet e-mail transfer.

SNMP (Simple Network Management Protocol)

A protocol used to allow network management of remote system and network resources.

socket

A virtual connection using a network protocol such as TCP/IP. Alternatively, a standard for connecting a CPU to a mainboard.

striping

The process of splitting data into chunks that are spread across multiple drives, thus increasing access speed. See RAID.

STP (Shielded Twisted Pair)

A standard for cabling that uses a twisted pair of cable terminated with an RJ-45 connector. Similar to the RJ-11 standard used for telephone wiring, twisted pair is cheap and robust, and enhances the inherent interference-resistant qualities of a twisted wire pair by adding an external shielding wrapper. See UTP.

subnet

A portion of a network separated by a router or other such device from the rest of the network. Alternatively, the addressing of that subnet in TCP/IP.

TCP/IP (Transfer Control Protocol/Internet Protocol)

The standard protocol for communications over the Internet, TCP/IP is a protocol that supports efficient routing over large internetworks and generally requires static configuration of addressing information (although newer standards such as DHCP eliminate this need). See DHCP.

Telnet

A protocol for terminal emulation communications over TCP/IP.

TFTP (Trivial FTP)

A protocol for file transfers over TCP/IP that is based on FTP. TFTP, however, provides no security.

thread

The smallest-possible unit of execution. Each process has at least one thread, but a process can have multiple threads to allow more efficient processing of complex tasks. See process and fiber.

token passing

A method for allowing each workstation on a network an equivalent number of opportunities to transmit on the wire, by using a “token” that is passed from workstation to workstation in sequential order. Contrast with CSMA/CD. See Token Ring.

Token Ring

A network standard developed by IBM that uses a token-passing method rather than competitive network access. Token Ring is more suitable for guaranteed workstation access to the network, though it can be less efficient under low-traffic situations. See token passing.

trust relationship

A secure connection established between two domains that enables the trusting domain to authenticate its resources to accounts existing in the trusted domain. Trust relationships can be two-way, but are not transitive in the current NTLM network model.

UNC (Uniform Naming Convention)

A standard for providing uniform resource names that are valid no matter where the resource is located or what type of resource it describes.

unicode

A standard for providing larger storage for characters, thus giving applications the capability to support multiple language sets much more easily than with ASCII.

UPS (Uninterruptible Power Supply)

A device designed to provide constant power to a workstation, thus allowing it to continue to function after a power failure. More advanced online UPS devices run on battery constantly, thus providing a steady source of power free of spikes and sags, and keep the batteries charged with the utility current supplied from the power jack.

UTP (Unshielded Twisted Pair)

A standard for cabling that uses a twisted pair of cable terminated with an RJ-45 connector. Similar to the RJ-11 standard used for telephone wiring, twisted pair is cheap and robust, and highly resistant to interference due to its twisted-pair configuration. See also STP.

virtual memory

A process for providing memory spaces that are not directly related to system RAM. This process allows memory spaces far larger than system RAM and removes the responsibility of an application to track its own memory addresses in RAM. See pagefile.

WAN (Wide-Area Network)

A series of interconnected LANs that provide networking services to a large campus or series of installations. See LAN.

WINS (Windows Internet Name Service)

A service for resolving NetBIOS names to TCP/IP addresses, WINS uses dynamic address configuration rather than static tables. See DNS.

WINSOCK

Short for Windows Sockets, a standard for network socketed communications over TCP/IP. The latest WINSOCK standard, Winsock 2.0, allows for socketed connections independent of network protocols.

workgroup

An organization of systems using peer-to-peer network resources. Workgroups lack centralized administration of accounts and access privileges. See domain.

workstation

An individual computer. Alternatively, the standard workstation product in the NT product line, capable of peer-to-peer networking as well as connection to larger, server-based networks and domains. Also the service that enables an NT system to act as a client to network resources on other NT systems. See server.

WOW (Windows-on-Windows)

The subsystem in Windows NT that provides an emulated DOS/Windows 3.x environment.

x86

Term for Intel’s 80x86 line of processors, beginning with the 286 (though most system requirements that state an x86 processor actually require a 386 or higher). Later versions of these processors use names rather than numbers to enhance their trademarkability. The line includes the 286, 386, 486, Pentium, Pentium Pro, Pentium MMX, and Pentium II processors.

zone

A virtual organization of AppleTalk devices on a network. Also, in the context of DNS, a subtree of the domain name space. See DNS.