The lowest level of the T-carrier hierarchy.
Overview
T1 is part of the T-carrier digital transmission architecture developed for the Public Switched Telephone Network (PSTN) in the 1960s. A T1 circuit (also called a T1 line) is formed from a combination of 24 DS-0 (Digital Signal Zero) channels, each having a bandwidth of 64 kilobits per second (Kbps), for a total bandwidth of 1.544 megabits per second (Mbps). These 24 DS-0 channels can either be used separately for carrying 24 separate voice circuits (called channelized T1) or aggregated into a single data stream (called unchannelized T1) for high- speed wide area network (WAN) connections.
T1 (sometimes called T-1) actually stands for T-carrier Level 1, but it is almost never referred to in this way.
Uses
T1 is the preferred technology used by enterprises for combining voice, fax, and data transmissions. This is because T1 is "trunking" technology that enables a single physical circuit to support as many as 24 separate virtual circuits, a process which is generally cheaper than provisioning 24 separate physical links. T1 lines are also typically used
To provide enterprises with dedicated leased-line WAN links among remote locations-for example, to connect a branch office to corporate headquarters.
To provide corporate users with high-speed access to the Internet.
Architecture
Like other members of the T-carrier family, T1 uses time-division multiplexing (TDM) to interleave multiple DS-0 channels into a single bit stream (called a DS-1 circuit). DS-0 generates 8 bits (1 byte) every 125 microseconds, or 8000 DS-0 frames per second. The bandwidth of a DS-0 channel is therefore
DS-0 = 8 bits x 8000 per second = 64,000 bits per second (bps) = 64 kilobits per second (Kbps)
Because T1 multiplexes 24 DS-0 channels together, a single T1 frame (or DS-1 frame) should equal 24 x 8 = 192 bps. The T1 specification, however, adds an extra bit to each frame to ensure that transceivers at each end of the line maintain their synchronization. This extra bit is added at the start of each DS-1 frame, which makes the length of a DS-1 frame equal to 192 + 1 = 193 bits. Using the same transmission rate of 8000 frames per second, this means that the total bandwidth of a T1 circuit is
T1 = 193 bits/frame x 8000 frames/sec = 1544000 bits/sec = 1.544 Mbps
TDM is applied to the individual DS-0 channels in such a way that each DS-0 channel is located at the same position of each DS-1 frame generated.
To package binary information into electrical signals, T1 originally used the Alternate Mark Inversion (AMI) line coding mechanism in which a voltage represents a binary 1 and no voltage represents zero. The problem with this mechanism was that it was hard to maintain synchronization between transceivers at opposite ends of the T1 circuit when a large number of successive 0s or 1s were transmitted. A scheme was therefore devised whereby bits were "robbed" from certain parts of each frame to ensure that synchronization could be maintained and to allow for control and signal maintenance information to be carried in-band within the circuit. The net result of this bit robbing was to reduce the data- carrying capacity of each DS-0 channel within DS-1 from 64 Kbps to only 56 Kbps. However, this bit- robbing scheme has no discernable effect on voice transmission.
You can work around the capacity-robbing effect of this bit robbing by replacing AMI line coding with Bipolar with 8-bit Zero Substitution (B8ZS) line coding. B8ZS substitutes a special byte if eight consecutive zero bits are detected to maintain a specific ones density to help maintain synchronization. This approach is called "ones density" and allows a T1 channel service unit/data service unit (CSU/DSU) at the customer premises to recover the data clock reliably when synchronization is lost with the T1 multiplexer at the telco central office (CO). The result of using B8ZS is that each DS-0 channel can carry the full 64 Kbps of data. An alternative scheme to B8ZS that is also commonly used is Zero Byte Time Slot Interface (ZBTSI) line coding.
Bellcore also developed an alternate scheme whereby a 2 Binary 1 Quaternary (2B1Q) line coding scheme was employed. 2B1Q is the same signal encoding mechanism employed by Integrated Services Digital Network (ISDN) and encodes 2 bits/baud instead of the 1 bit/baud supported by AMI. This new technology was called "repeaterless T1" because it eliminated the necessity of regenerating T1 signals every 6000 feet (1830 meters) using repeaters, a process that made original T1 deployments complex and expensive. Repeaterless T1 needed repeaters only every 12,000 feet (3660 meters) and transmitted data at only 784 Kbps over each twisted pair. Because two pairs of wires are used for T1, this new technology also carries data at T1 speed of 1.544 Mbps. This new technology is now commonly referred to as High bit-rate Digital Subscriber Line (HDSL). A telco will often provision customers with HDSL and call it T1 instead, because it is functionally equivalent in speed and framing to T1.
Implementation
T1 cannot operate over analog Plain Old Telephone Service (POTS) telephone lines. Instead, it must be deployed using specially conditioned copper twisted- pair lines, with two pairs of wires (four wires) being used for a single T1 circuit. To support full-duplex communication, two of these four wires are used for transmission (TX interface) and the other two for receiving (RX interface). T1 lines typically terminate at the customer premises with an RJ-48 connector, which looks like an RJ-45 connector but is pinned differently. T1 lines are generally unshielded twisted-pair (UTP) cabling but other media can be used, including coaxial cabling or fiber-optic cabling.
T1 usually cannot run over existing local loop wiring because:
Bridge taps installed by telcos to trunk telephone traffic in neighborhood wiring causes distortion of T1 signals, so these must be removed to allow the circuit to carry T1 signals.
Loading coils, which are used to reduce signal distortion for analog phone lines, have the opposite effect of increasing distortion of digital signals, and these also must be removed.
To deploy T1 as a solution for multiplexing voice traffic, a T1 channel bank is generally installed at the customer premises. This channel bank can be connected to a Private Branch Exchange (PBX), which then connects to digital telephone and fax equipment. For WAN data links the scenario is usually somewhat different, using customer premises equipment (CPE) such as
A T1 CSU/DSU for connecting bridges or routers to T1 circuits
T1 bridges and routers with integrated T1 CSU/DSUs
A T1 multiplexer (MUX), a multiplexer for aggregating several T1 circuits for even higher-speed communication
T1 access routers, which support multiple remote access links over a single T1 line
To test T1 equipment such as channel banks and CSU/DSUs, use a cable simulator, which is a passive device that simulates a standard 22-gauge twisted-pair T1 line that is 1310 feet (400 meters) long (the alternative is to use 1310 feet of actual 22-gauge twisted-pair wiring). Connect two cable simulators between your CPE and your T1 test equipment using the TX and RX interfaces to analyze your device's performance. A "wet" T1 line carries a small DC current of about 140 mA (milliamperes) at several hundred volts for powering the CSU/DSU at the customer premises. "Dry" lines carry no current, so CSU/DSUs must be powered from the customer premises. Do not touch a T1 line-a wet line can give you a serious shock!
Marketplace
The cost of provisioning T1 is complex and depends on whether you are using it for high-speed Internet access (T1 local loop connections between the customer premises and the telco CO) or for building a high-speed WAN (long-haul T1 lines crossing large geographical distances). A good rule of thumb for T1 WAN links is that the long-haul cost is about $2.50 per mile, which means a 2000-mile T1 leased line would cost about $5,000. These figures were for the year 2000, and the good news is that T1 prices have been falling about 10 percent per year for the last couple of years.
The cost for a T1 local loop connection to provide your company with dedicated high-speed Internet access is generally between $1,000 and $1,500 per month. These prices seem not to be changing much, despite forecasts that Digital Subscriber Line (DSL) technologies will cut into the T1 market, the main reason being the greater reliability of T1 compared to newcomer DSL.
The primary reason T1 lines are so expensive is that they are always "on" regardless of whether they are being used. This is characteristic of leased lines and provides both the benefit of availability and the cost of underutilization. A cheaper solution for many companies that do not require full T1 capacity is to lease a fractional T1 service such as 4 x DS0 = 256 Kbps from their carrier and then have them upgrade it to higher speeds as their WAN traffic grows. Fractional T1 is usually cheaper than using individual DS0 circuits multiplexed together.
T1. Some different WAN scenarios using T1 lines.
Notes
T1 and PRI-ISDN both carry data at around 1.5 Mbps, but they are incompatible so far as their framing formats are concerned. For example:
T1 multiplexes 24 DS-0 channels using TDM for carrying data and adds a control bit to each T1 frame (and may use bit robbing to gain additional bandwidth for control purposes)
PRI-ISDN multiplexes 23 DS-0 channels for carrying data and adds a 24th DS-0 channel dedicated to carrying control information.
The European E1 specification avoids the bit robbing used in American T1 by adding a 16-bit control header to each E1 frame instead of the single bit added to T1 frames.
See Also Channel Service Unit/Data Service Unit (CSU/DSU) , Digital Subscriber Line (DSL) ,DS-0 ,DS-1 ,High-bit-rate Digital Subscriber Line (HDSL) ,Integrated Services Digital Network (ISDN) ,leased line ,line coding ,PRI-ISDN ,Private Branch Exchange (PBX) ,Public Switched Telephone Network (PSTN) ,
Customer premises equipment (CPE) used to terminate a T1 line and make it available across an organization.
Overview
T1 channel banks are typically used to enable T1 lines to connect to
Data terminal equipment (DTE) such as routers and access servers
Private Branch Exchange (PBX) units that provide integrated phone/fax services
A typical T1 channel bank consists of a modular chassis unit to which you can add various expansion cards to provide digital communication services for CPE. The modular chassis allows customers to add channels and upgrade fractional T1 services to full T1 or higher. It also allows customers to multiplex several channels to provide higher bandwidth for high-speed data connections to routers, Web servers, and other DTEs. The chassis typically includes a built-in T1 Channel Service Unit (CSU) for terminating the T1 circuit at the customer premises, plus a number of slots capable of holding expansion cards for various uses.
T1 channel bank. Using a T1 channel bank to connect a router and PBX to a T1 line.
Each expansion card in a T1 channel bank typically handles either one or two DS-0 channels, which means that different channels can supply different services (such as voice, fax, or data connections). Typical types of expansion cards include the following:
Data service cards: These usually have a dual channel format that supports two DS-0 channels and employ a serial interface such as RS-232, RS-530, or V.35. These interfaces are then used for directly connecting the unit to bridges and routers having integrated Channel Service Unit/Data Service Units (CSU/DSUs).
High-speed data cards: These support up to 1.544 megabits per second (Mbps) in 64-kilobits per second (Kbps) or 56-Kbps increments (the speed depends on how DS-0 is provisioned by the carrier) by multiplexing DS-0 channels.
Voice cards: These are used to connect the unit to a PBX or directly to a telephone using standard 4-wire connections.
Modem cards: These convert the channel bank into a modem pool to support corporate remote access needs.
Some T1 channel banks can support as many as four T1 lines, which can be configured for both active and backup purposes to provide redundant wide area network (WAN) connections.
See Also Channel Service Unit (CSU) , Channel Service Unit/Data Service Unit (CSU/DSU) ,customer premises equipment (CPE) ,data terminal equipment (DTE) ,Private Branch Exchange (PBX) ,
Part of the T-carrier hierarchy.
Overview
T3 represents the "next step up" for enterprises that want to build their wide area network (WAN) connections using dedicated leased lines. Although the commonly used and relatively inexpensive T1 lines used in enterprises carry traffic at 1.544 megabits per second (Mbps), T3 lines support a much faster speed of 44.736 Mbps, well above standard 10Base2 Ethernet speeds and almost comparable to Fast Ethernet. This huge jump in speed, however, comes at a significant cost and with some associated issues:
T3 requires fiber-optic cabling to be provisioned from the telco central office (CO) to the customer premises, because T3 cannot run over existing copper local loop wiring even if it is properly conditioned. This up-front cost of laying fiber must be factored into the cost of deploying T3 in the enterprise.
The cost of a dedicated T3 line is generally between $25,000 and $35,000 per month, a hefty price tag compared to the $1,000 to $1,500 cost of individual T1 lines. Many companies have a difficult time justifying the cost of upgrading from T1 to T3.
Although T3 operates over fiber-optic cabling, there is no universal specification for how physical layer signaling occurs with this system. As a result, different telcos and telecommunications equipment vendors have developed many proprietary T3 signaling schemes, and most of these schemes cannot interoperate. This means that if you want to deploy T3 you must "buy in" to equipment from a single vendor (or lease equipment from your telco).
Despite these issues, T3 has grown in popularity in the last few years, particularly for large enterprises to connect their data centers to the Internet. The main problem faces companies whose WAN or Internet access needs are too great for a T1 line to satisfy yet do not require the capacity (or cannot afford the cost) of a full T3 line. The emerging solution to this problem is for telcos to provision services that bundle multiple T1 links for greater throughput. Cable and Wireless is one provider that offers a dedicated Internet access service called NxT1 that can aggregate from two to seven T1 lines into a single fat data pipe carrying up to 10 Mbps. This system employs Cisco 7500 routers running Multilink Point-to- Point Protocol (MPPP) for link aggregation. The disadvantage of this scheme is that customers must order additional T1 port connections to the provider's network, which adds to the cost. Nevertheless, the cost of this scheme is generally less than using fractional T3, which requires a full T3 interface at the customer premises.
See Also Multilink Point-to-Point Protocol (MPPP) ,
A family of standards for multiuser conferencing and collaboration over a data network.
Overview
T.120 represents a series of eight International Telecommunication Union (ITU) standards that define real-time multipoint communication over a network such as the Internet. T.120 can be used for such tasks as video conferencing, data exchange, or interactive gaming. The T.120 standards define such things as
Multipoint services for conferencing
Standard network services
Guidelines for defining data channels
Whiteboard methodologies
Application-sharing protocols
File transfer methodologies
A related standard from the ITU is the H.323 standard for video and audio conferencing.
Architecture
The architecture of the T.120 standard follows that defined by the Open Systems Interconnection (OSI) reference model for networking. The T.120 architecture can be divided into two parts:
Network-layer and transport-layer standards (T.122 through T.125): Allow data to be transmitted and received among conferencing nodes over a variety of supported network connections. These standards also provide platform independence and the capacity for simultaneously managing multiple participants running on different operating system platforms and conferencing software.
Application-layer standards (T.126 through T.128): Support multiuser conferencing functions such as whiteboarding, file transfer, and application sharing across different platforms and networks.
The following table shows the details of the various standards included under the T.120 umbrella.
Standard | Description |
T.121 | A required standard for T.120 applications that defines how conference nodes register themselves with a T.120 node controller. Also defines the generic application template (GAT) for building T.120 application protocols and management facilities. |
T.122 | Defines multipoint communication services (MCS) over various topologies to enable multiple participants to send data as part of a conference. The MCS defined by T.122 are implemented by T.125. |
T.123 | Defines flow control, error control, and sequencing mechanisms for connect, disconnect, send, and receive functions over different network connections. |
T.124 | Defines how multipoint conferences are initiated and administered and defines the generic conference control (GCC) that manages and monitors users, address lists, data flow, and MCS resources. |
T.125 | Defines how data is transmitted during a conference, specifying the private and broadcast channels that transport conference data. T.125 implements the MCS defined by T.122. |
T.126 | Defines mechanisms for transmitting and receiving whiteboard information among conference nodes and managing the multiuser whiteboard workspace. |
T.127 | Defines mechanisms for file transfer among conference nodes in either broadcast or directed mode. |
T.128 | Defines mechanisms for application sharing among conference nodes so that users can share their local programs with others for collaborative purposes. |
Notes
T.120 also forms the basis of the Remote Desktop Protocol (RDP), which is used by the Terminal Services of Microsoft Windows 2000, Windows XP, and Windows .NET Server.
See Also H.323 , International Telecommunication Union (ITU) ,Open Systems Interconnection (OSI) reference model ,Remote Desktop Protocol (RDP) ,
Stands for Terminal Access Controller Access Control System, a security protocol supported by Cisco routers.
See Also Terminal Access Controller Access Control System (TACACS)
Assuming the role of an object's creator, thus having the associated rights and privileges that this role incurs.
Overview
Ownership describes the highest level of permissions that can be granted to objects. On the Microsoft Windows 2000, Windows XP, and Windows .NET Server platforms, these objects can include files and folders, Active Directory directory service objects, and so on. For example, assuming ownership of an object such as a file on an NTFS file system (NTFS) volume gives one the right to share the object and assign permissions to it. Normally, the user who creates a file on an NTFS volume is the owner. Other users can take ownership of the file provided the user is either a member of the Administrators domain local group, has NTFS full control permission on the object, or has explicit permission to take ownership of the object.
Notes
Ownership can only be taken; it cannot be assigned.
See Also NTFS permissions (Windows 2000,Windows XP ,and Windows .NET Server),NTFS permissions (Windows NT) ,NTFS special permissions (Windows 2000,Windows XP ,and Windows .NET Server),NTFS special permissions (Windows NT)
A device used to back up data to magnetic tape.
Overview
Tape drives and their larger cousins, tape libraries, form the backbone of the disaster recovery plan for most enterprises. Tape drives are distinguished from one another by a variety of factors:
Recording technology: There are several different ways in which data can be written to magnetic tape, including linear-scan, helical-scan, or hybrids of these two basic technologies. For more information about these various recording technologies and the tape formats from different vendors that support them, see the following article, "tape format."
Capacity: The capacity of a tape drive is the amount of data it can store on a single tape cartridge. This capacity is usually measured in gigabytes (GB) and can be expressed as either native capacity for uncompressed data or compressed capacity. Tape drives for large enterprise networks may have capacities exceeding 50 GB, but drives for departmental or workgroup use may have capacities of only a few GB. Compressed capacity is usually specified as twice the native capacity-in other words, a tape drive with 50 GB native capacity would be rated as having a compressed capacity of 100 GB. The actual capacity when compression is used, however, depends on the type of data being backed up.
Transfer speed: This is the speed at which data can be buffered by the tape drive and written to tape. For enterprise-class tape drives, transfer speeds exceeding 25 megabytes per second (MBps) are possible, but for workgroup or small business use the capacity is often measured in megabytes per minute (MB/min) instead and is considerably less.
Cost of media: If you are backing up large amounts of data frequently, the cost of individual tape cartridges can be a significant expense that needs to be budgeted for accordingly. Cost for tapes range from about $10 to $100, depending on the tape format used.
Input/output (I/O) interface: Most enterprise tape drives use Small Computer Systems Interface (SCSI) as their data interface, but some cheaper drives for small business use have ATAPI/IDE, parallel, or Universal Serial Bus (USB) interfaces.
Interoperability: Usually a tape cartridge produced by one vendor will not work in a tape drive from a different vendor. The only difference is for tapes and drives that adhere to the new Linear Tape Open (LTO) standard developed by Hewlett-Packard, IBM, and Seagate Technology. LTO drives and cartridges have only recently appeared on the market, but they are likely to become a dominant format in the years to come.
Software driver: Before buying a tape drive for your servers, make sure that your backup software has a suitable driver for this hardware.
Marketplace
The tape drive market is basically divided into three categories:
Enterprise: These drives have the largest capacities and best performance to meet the demanding backup needs of large companies. Some popular drives in this market include DLT 8000 drives from Quantum Corporation, which cost about $5,000 and have a native capacity of 40 GB; Exabyte Mammoth-2 and Sony AIT-2 drives, which are comparable in cost and capacity to the DLT 8000; and the new SuperDLT drives from Quantum and LTO Ultrium drives from IBM, which have higher capacity and cost.
Departmental: These medium-capacity drives include the Tandberg SLR100, Benchmark DLT1, Ecrix VXA-1, and drives from other vendors. The cost for departmental drives is usually $1,000 or slightly higher, and capacity is measured in tens of GB.
Workgroup: For small business settings, these are a wide range of popular drives, including digital audio tape (DAT) drives from a number of vendors, OnStream's ADR50 drives, and Ultrium 3580 from IBM. Drives for workgroup or desktop use typically have capacities of a few GB and cost several hundred dollars.
Notes
Here are some tips on getting the most from your tape drive:
Make sure you clean your drive's read/write heads regularly, usually every 10 hours or so. Some newer drives automatically clean themselves as needed, while others display a light-emitting diode (LED) indicator when cleaning needs to be done.
Avoid exposing both the tape drive and the tapes to stray magnetic fields such as those from computer monitors. When some tape formats are exposed to such fields, data can be lost-for others, the entire tape can be rendered unusable.
Replace your tapes regularly according to the mean lifetime of the particular tape format you use. Most tapes can be used about 50 times before they wear out and become unreliable.
See Also backup , disaster recovery ,
Techniques for storing digital data on magnetic tape for backup purposes.
Overview
Magnetic tape is the medium used by most companies for archiving valuable business information. No single standard exists, however, for how information is stored on tape. As a result, different vendors have developed a variety of tape formats for use by small, medium, and large businesses. These formats differ in capacity, format, and ease of use, and you need to weigh these factors when deciding which format is appropriate for your organization's needs.
The two main technologies used for storing data on magnetic tape are as follows:
Helical-scan: This method pulls the magnetic tape out of the cartridge that houses it, tensions the exposed tape using capstans and rollers, and wraps the tape around a rotating drum that contains multiple read/write heads. Data is then written in diagonal stripes across the width of the tape, and the entire tape is filled in one pass. This technology is essentially the same used by common VCRs and was developed in the late 1980s. One problem with this technology is that the high speed of the drum rotation (7000 rpm or higher), coupled with the tensioning of the tape, can stretch the tape over time or even cause it to break, but most modern helical tape technologies are designed to minimize these effects.
Linear-scan: This method leaves the tape in the cartridge while passing the small exposed portion of tape over stationary read/write heads. Data is written on one track until the end of the tape is reached, and then the tape reverses direction and the next track is written. This "serpentine" writing action continues until the entire tape is filled. Many linear tape formats use multiple heads to write several tracks simultaneously for greater throughput. The stationary heads and minimal tape tensioning make this approach more reliable than helical-scan technology, but helical-scan generally provides greater performance and supports higher capacity. Linear-scan is also an older technology than helical- scan and is better established, especially in the enterprise arena.
Marketplace
A number of vendors have developed both linear-scan and helical-scan tape technologies, with the result that numerous linear and several helical tape formats are on the market. The following are some of the popular linear-scan tape formats offered by different vendors:
Quarter-inch cartridge (QIC): This is the oldest linear tape format, developed in the early 1980s by Tandberg Data. QIC cartridges have capacities up to 40 gigabytes (GB) with transfer speeds up to 10 megabytes per second (MBps). QIC cartridges originally had a large 4-by-6-by-5/8-inch form factor, but this is now considered obsolete and has been largely replaced by smaller minicartridges, especially those from Travan using technology developed by 3M Corporation/Imation Corporation.
Digital linear tape (DLT): This linear tape technology was first developed by Conner in 1991 and acquired by Quantum Corporation in 1994. DLT has been the dominant tape format used in enterprise backup solutions for many years. DLT cartridges generally have a form factor of 5.25 inches and come in various types. For example, the Quantum DLT 8000 tape format is popular in the enterprise environment and has a native capacity of 40 GB per cartridge, supports transfer rates of 6 MBps, and costs about $100 each. The DLT 8000 format achieves its high transfer rate by writing four tracks of data simultaneously across the tape. Other popular DLT formats include DLT 7000 and DLT 4000.
Super DLT (S-DLT): This format was recently developed by Quantum and uses a combination of magnetic-head and laser-guided technologies to improve on the earlier DLT architecture. Super DLT cartridges have a native capacity of 100 GB and support transfer speeds from 10 to 40 MBps.
9840: This linear format uses half-inch tape and was developed by StorageTek for their tape automation market. The 9840 format is widely used in mainframe computing environments and has a native capacity of 20 GB and a transfer rate of 10 MBps.
Mammoth-2: This linear tape format from Exabyte Corporation is a competitor of Quantum's DLT format and has a native capacity of 60 GB and a transfer rate of 12 MBps and costs about the same as comparable DLT tape.
SLR100: This format from Tandberg Data has a native capacity of 50 GB and transfer speed of 5 MBps. SLR100 tapes cost about $100 each and are popular in the departmental backup arena.
The following are some of the popular helical-scan tape formats offered by different vendors:
Advanced Intelligent Tape (AIT): This technology from Sony Corporation was the first helical-scan tape technology developed for tape backup. AIT has become a popular alternative to DLT in the enterprise and comes in several versions. AIT-2, for example, has a native capacity of 50 GB and a transfer speed of 6 MBps and costs about $100 per cartridge. The newer AIT-3 has a native capacity of 100 GB and a transfer speed of 11 MBps.
8-millimeter: This helical tape format has the same width as standard videotape and supports native of 20 GB and higher, with transfer speeds of 3 MBps and higher. Sony and Exabyte are two popular manufacturers of 8-millimeter tapes.
Digital data storage (DDS): Often wrongly called digital audio tape (DAT), this is a helical tape format based on technology originally developed for the professional audio market. DDS comes in several popular versions, including DDS-1 (2 GB native capacity) DDS-2 (4 GB), DDS-3 (12 GB), and DDS-4 (20 GB). Transfer rates range from 1.5 to 2.4 MBps. DAT cartridges use 4-millimeter tape in a 3.5-inch form factor and typically cost about $50 each.
Some other tape formats include
Advanced Digital Recording (ADR): This is a technology originally developed by Phillips and now available from OnStream that combines aspects of both linear-scan (stationary heads) and helical-scan (multiple tracks written simultaneously). For example, the ADR50 format has a native capacity of 25 GB and a transfer rate of 2 MBps and is well-suited to the workgroup arena.
VXA: This new format developed by Ecrix Corporation writes data to tape in packet form instead of the streaming method used by both linear-scan and helical-scan tape formats. VXA uses a variable tape speed in contrast to the fixed speed of the standard formats. VXA also employs a parity striping format called discrete packet format (DPF) to optimize data integrity. VXA-1 tapes have a capacity of 33 GB and a transfer speed of 3 MBps and cost less than $100.
Prospects
A new development in this field has been the emergence of an open standard for linear-scan tape media called Linear Tape Open (LTO). This standard was developed jointly by Hewlett-Packard, IBM, and Seagate Technology and is intended to bridge the interoperability gap that exists because of each tape vendor developing its own proprietary technology. The aim is that LTO tape from one vendor would work equally well on an LTO tape drive from a different vendor. There are actually two different LTO standards:
Accelis: A fast data-access linear tape technology for small and mid-sized businesses.
Ultrium: A high-capacity tape technology for large enterprises.
The Ultrium tape format has a native capacity of 100 GB and a transfer speed of 15 MBps and is likely to emerge as the main competitor to market leader DLT and its successor, Super DLT.
See Also backup , storage ,
A tape backup device that can retrieve and load tapes automatically.
Overview
When enterprises have hundreds of gigabytes (GB) to several terabytes of data that need to be regularly backed up, simple tape drives, even enterprise-class ones that have capacities of 100 GB or more, simply cannot do the job efficiently. That is where a tape library comes in. A tape library, also called a tape autoloader, is essentially a box with a robotic arm or other device that can store a number of tapes, select and load them, and unload and store them as needed. Tape libraries range from small boxes that can sit on a desk or be mounted in an equipment rack to large room-sized enclosures containing multiple tape drives and thousands of tapes.
Marketplace
One of the more popular vendors in the tape library market is Exabyte Corporation, whose tapes use a proprietary format called Mammoth (now Mammoth-2 or M2). IBM is another major player with its standards- based linear tape open (LTO) 3584 UltraScalable Tape Library, which can store as much as 240 TB of data. Other popular vendors of tape libraries include ADIC, Grau Data Storage, Spectra Logic Corporation, StorageTek, and several others.
See Also backup , disaster recovery ,
Stands for Telephony Application Programming Interface, a set of standard application programming interfaces (APIs) developed by Microsoft Corporation and Intel Corporation for accessing telephony services.
See Also Telephony Application Programming Interface (TAPI)
A family of telco specifications for digital trunking, also used for high-speed wide area network (WAN) connections.
Overview
The original trunking (long-haul communications) architecture of the Plain Old Telephone System (POTS) was analog in nature. This L-carrier system allowed multiple analog local loop connections to be aggregated into trunk lines using frequency-division multiplexing (FDM). The main advantage of trunking was that it saved carriers the cost of having to deploy multiple long-haul lines between different geographical locations.
L-carrier services, however, suffered from distance limitations due to noise and signal distortion. As a result, Bell Laboratories developed T-carrier technology in the early 1960s to replace long-haul analog trunking lines with digital lines. This improved performance and made digital data services available to companies that needed to connect remote branch offices with mainframe computing centers. The first level of T-carrier, the T1 service, was first commercially deployed in the mid-1980s.
Uses
T-carrier services such as T1 and T3 lines have a variety of uses in the enterprise:
To connect Private Branch Exchange (PBX) equipment at the customer premises with the telco central office (CO)
To economically provide enterprises with integrated voice/fax/data services
For building dedicated, high-speed WANs
For providing users on corporate networks with reliable high-speed Internet access
For connecting corporate Web servers to the Internet
Architecture
The T-carrier family of specifications basically outlines two considerations: physical media and signaling. As a physical media specification, various levels of the T-carrier hierarchy run over copper twisted-pair wiring, coaxial cabling, fiber-optic cabling, or wireless microwave transmission. For example, T1 lines employ two twisted pairs (four wires) to ensure efficient signaling (this is in contrast to traditional analog POTS lines, which employ only one pair of wires). By contrast, T3 uses fiber-optic cabling as its transmission medium. The table shows the various combinations allowed for physical media. Note that the exact specifications for T3 to run over fiber have never been standardized-as a result, different telcos have developed their own proprietary optical transmission schemes for T3.
T-Level | Media |
T1 | copper |
T1C | copper |
T2 | copper/microwave |
T3 | fiber/microwave |
T3C | fiber/microwave |
T4 | fiber/microwave |
As far as signaling is concerned, the T-carrier system is based on the DS-0 signaling standard defined by AT&T for digital voice transmission. A single digitized voice channel (DS-0 or Digital Signal Zero channel) carries binary data at 64 kilobits per second (Kbps) and forms the building block of the T-carrier service hierarchy. For example, the T1 service consists of 24 DS-0 channels multiplexed to provide a total data rate of 1.55 megabits per second (Mbps). Note that no carrier is defined for DS-0 itself-that is, you cannot use a single DS-0 channel for digital data transmission, only multiple DS-0 channels aggregated together.
In T-carrier services, these DS-0 channels are multiplexed using time-division multiplexing (TDM) instead of the FDM scheme used in the older L-carrier services. For more details of the T-carrier multiplexing process, see the article "T1" earlier in this chapter. T-carrier services form a hierarchy of standard digital transmission speeds, as shown in the table below. In real life, however, only T1 and T3 services are implemented; T2 is rarely used and no real standard exists for T4 transmission. For digital transmission speeds faster than T-carrier services offer, a newer technology called Synchronous Optical Network (SONET) was developed. SONET now forms the basis of most long-haul and backbone transmission networks for telcos, with T1 and T3 used mainly for provisioning high-speed data services to the customer premises.
The following table shows the different T-carrier services that have been defined (there are no levels defined beyond T4, as SONET has taken over in this domain). Note that in common parlance T1 and DS-1 mean the same thing, but in fact T1 defines the physical specification and DS-1 defines the signaling method. Note also that T1C and T3C refer to "concatenating" (joining together) two T1 or T3 circuits to double aggregate bandwidth. Despite the number of different T-carrier levels defined, only T1 and T3 are commonly used, and T4 has never been implemented, as SONET covers that range. There is also a variant service called fractional T1 offered by most telcos, which essentially means a full T1 circuit is provisioned to the customer but transmission is limited (and charged) for only a portion of the circuit-for example, 4, 8, or 12 DS-0 channels instead of the full 24 channels.
T-Level | DS-Level | Number of DS-0 Channels | Bandwidth |
T1 | DS-1 | 24 | 1.544 Mbps |
T1C | DS-1C | 48 | 3.152 |
T2 | DS-2 | 96 | 6.312 |
T3 | DS-3 | 672 | 44.736 |
T3C | DS-3C | 1344 | 91.053 |
T4 | DS-4 | 4032 | 274.176 |
Implementation
T-carrier is usually provisioned as a leased line service from Incumbent Local Exchange Carriers (ILECs). Costs for these services are high, but their high level of reliability makes them a staple of enterprise telecommunications networks. T-carrier can be provisioned in two basic formats for customers:
Channelized: Individual DS-0 channels can carry their own traffic. Channelized T1, for example, allows a single T1 line to carry 24 separate voice channels multiplexed together, which is cheaper for customers than ordering 24 separate local loop connections. Channelized T1 is often used to connect corporate Private Branch Exchanges (PBXs) to ILECs. A channelized T1 line is essentially a single line that is logically equivalent to 24 separate "virtual" telephone lines.
Unchannelized: The DS-0 channels are essentially merged into a single fat pipe, primarily for carrying data for high-speed WAN links. Note the difference: channelized T1 is used for voice, unchannelized for data.
Often what is referred to as T-carrier is really a different service running over the T-carrier physical interface. For example, a 1.544 Mbps frame relay link is really a frame relay running over two-pair twisted wiring using the T1 physical layer specification.
Notes
In Europe a different digital carrier hierarchy called the E-carrier system evolved. For example, the European equivalent of the T1 line is the E1 line, which carries data at 2.048 Mbps. T-carrier and E-carrier systems are incompatible but can interface with each other by using special multiplexing equipment.
Japan calls its digital hierarchy the J-carrier system, but it is essentially the same as the American T-carrier system.
See Also DS-0 , DS-1 ,DS-3 ,frame relay ,Incumbent Local Exchange Carrier (ILEC) ,Plain Old Telephone Service (POTS) ,Synchronous Optical Network (SONET) , wide area network (WAN)
Stands for Transmission Control Protocol, a transport layer protocol of the Transmission Control Protocol/Internet Protocol (TCP/IP) suite.
See Also Transmission Control Protocol (TCP)
Stands for Transmission Control Protocol/Internet Protocol, an industry-standard protocol suite forming the basis of the Internet.
See Also Transmission Control Protocol/Internet Protocol (TCP/IP)
The procedure used for establishing and terminating Transmission Control Protocol (TCP) sessions.
Overview
All TCP communications are connection-oriented in nature. In other words, a TCP session must be established before the hosts involved can engage in the exchange of data between them. The TCP three-way handshake does this by establishing a logical connection between the hosts to ensure reliable transmission can be achieved.
TCP three-way handshake. How the TCP three-way handshake procedure operates.
The three stages of a TCP three-way handshake are the following:
The initiating host sends a TCP packet requesting a new session. This packet contains the initiating host's sequence number for the connection. The packet includes information such as a set SYN (synchronization) flag and data about the size of the window buffer on the initiating host.
The target host sends a TCP packet with its own sequence number and an ACK (acknowledgment) of the initiating host's sequence number.
The initiating host sends an ACK containing the target sequence number that it received.
A similar three-way process is used to terminate a TCP session between two hosts. Using the same type of handshake to end the connection ensures that the hosts have completed their transactions and that all data is accounted for.
See Also ACK , connection-oriented protocol ,
Stands for time-division multiplexing, a method for sending several data streams over a single communication path.
See Also time-division multiplexing (TDM)
Stands for Time Division Multiple Access, a cellular communications technology based on time-division multiplexing (TDM).
See Also Time Division Multiple Access (TDMA)
Stands for time domain reflectometry, a cable testing technique for finding breaks or shorts in a cable.
See Also time domain reflectometry (TDR)
An information resource program developed by Microsoft Corporation for IT (information technology) professionals who work with Microsoft products.
Overview
Microsoft TechNet is important to those who plan, deploy, maintain, support, and evaluate Microsoft business products, such as Microsoft Windows 2000 and members of the Microsoft BackOffice suite. The TechNet program includes a monthly CD subscription, a Web site, electronic newsletters, regular technical briefings at locations around the world, and special offers. Each month a collection of CDs updates your TechNet binder to ensure that you have the latest and most accurate information on all Microsoft products and services. The subscription includes four categories of CDs:
Monthly issues of up-to-date technical information, which include the Technical Information, Supplementary Drivers and Patches, and full Microsoft Knowledge Base CDs. The Technical Information CD includes a large collection of manuals, resource kits, and other documentation on current versions of Microsoft products; you can either browse or search this CD to find the information you need. The Knowledge Base CD is a collection of thousands of articles written by Microsoft support professionals that answer technical questions, provide detailed how-to information, resolve bugs, list fixes, and document changes and corrections to Microsoft products.
The Server Utilities, Client Utilities, and Software Library Archive CDs, as well as CDs that contain utilities from various Microsoft product resource kits.
Service Pack CDs for all Microsoft products for which service packs are available. These service packs provide cumulative fixes and patches for bugs and known problems with Microsoft products and can include additional enhancements to the original product version.
Extras, such as CDs for Microsoft Seminars Online and time-limited evaluation versions of Microsoft software.
An enhanced version of TechNet, called TechNet Plus, includes beta evaluation software for various upcoming Microsoft products.
For More Information
Visit Microsoft TechNet online at www.microsoft.com/technet.
A local telephone company.
Overview
The term telco is generally used to refer to the local telephone company that owns the local loop connection between your customer premises and the telco central office (CO). The term is also sometimes used to mean any carrier or service provider that can provision voice or data services over your local loop, regardless of whether they actually own that loop.
In its strictest sense, your telco is your Incumbent Local Exchange Carrier (ILEC), typically one of the four Regional Bell Operating Companies (RBOCs), but sometimes an independent phone company, especially in rural areas. Other companies sometimes referred to as telcos include Competitive Local Exchange Carriers (CLECs), inter-exchange carriers (IXCs), and even metropolitan area network (MAN) service providers such as metropolitan Ethernet providers.
Architecture
From the point of view of business customers, who are usually located in dense urban areas, the most important aspect of a telco is how it implements and provides access to its MAN for provisioning high-speed data services. In a typical scenario, the MAN is a dual high-speed Synchronous Optical Network (SONET) ring running on fiber owned by the telco. OC-48 rings running at 2.5 gigabits per second (Gbps) are still common, but most telcos are upgrading to faster services such as OC-96 or OC-192. Business customers who need the highest performance can usually connect an Ethernet switch or router at their customer premises through a T3 line or trunked T1 lines to an Asynchronous Transfer Mode (ATM) switch residing at the periphery of the telco's network. The ATM switch then connects to the SONET ring to allow customers to establish wide area network (WAN) connections among different branch offices.
See Also Asynchronous Transfer Mode (ATM) , central office (CO) ,Competitive Local Exchange Carrier (CLEC) ,Ethernet switch ,Incumbent Local Exchange Carrier (ILEC) ,inter-exchange carrier (IXC) ,metropolitan Ethernet ,optical carrier (OC-x) level ,Regional Bell Operating Company (RBOC) ,router ,Synchronous Optical Network (SONET) ,
A national trade organization representing all aspects of the telecommunications industry in the United States.
Overview
Working in conjunction with its subsidiary, the MultiMedia Telecommunications Association (MMTA), and its industry peer organization, the Electronic Industries Alliance (EIA), the Telecommunications Industry Association (TIA) represents its members in activities such as establishing public policies and government regulatory issues, developing standards for communication and networking, and organizing trade shows and other events. The TIA's goal is to provide member companies, which are drawn mostly from service providers and hardware vendors in the communication industry, with a forum for discussing industry issues and a voice for representing members' interests on the national and international level.
Active in telecommunications standards development, the TIA is endorsed and accredited by the American National Standards Institute (ANSI). The Standards and Technology Department consists of five divisions organized in more than 70 groups responsible for formulating new standards. These five divisions are
Fiber Optics
User Premises Equipment
Network Equipment
Wireless Communications
Satellite Communications
For More Information
Visit the TIA at www.tiaonline.org
See Also American National Standards Institute (ANSI) ,EIA/TIA wiring standards ,Electronic Industries Alliance (EIA) ,standards organizations
Various services provided to customers by telcos.
Overview
In addition to standard voice services, telcos offer a wide variety of data transmission services. These services are provisioned to business customers by connecting switching and multiplexing devices located at the telco central office (CO) to customer premises equipment (CPE) such as access servers, routers, and Ethernet switches. These services may be provisioned over the ubiquitous copper local loop wiring, specially conditioned twisted-pair wiring, fiber-optic cabling, or even wirelessly, using microwave transmission or satellites. Using telco data services companies can deploy
Dedicated leased lines for permanent, dedicated point-to-point wide area network (WAN) connections
Packet-switched services such as frame relay and X.25 for building multipoint WANs
Dial-up remote access solutions for mobile users using either analog modems over the circuit- switched Public Switched Telephone Network (PSTN) or using Integrated Services Digital Network (ISDN) services
High-speed Internet access for corporate users using digital subscriber line (DSL) technologies such as Asymmetric Digital Subscriber Line (ADSL) and High-bit-rate Digital Subscriber Line (HDSL)
Other data communication technologies that telcos sometimes provide include
Asynchronous Transfer Mode (ATM)
digital data service (DDS)
Switched 56
Switched Multimegabit Data Services (SMDS)
T-carrier services such as T1, T3, and fractional T1
See Also analog modem , Asynchronous Transfer Mode (ATM) ,central office (CO) ,digital data service (DDS) ,Digital Subscriber Line (DSL) ,frame relay ,Integrated Services Digital Network (ISDN) ,Public Switched Telephone Network (PSTN) ,Switched 56 ,Switched Multimegabit Data Services (SMDS) , wide area network (WAN), X.25
Working from a location other than the office, usually from home.
Overview
Telecommuting has emerged as a phenomenon of the late 1990s as a new way of working. A knowledge worker who works from home may be called either a telecommuter or a teleworker. The empowering technology behind telecommuting is the Internet, a ubiquitous public network that provides cheap and easy network connectivity between home workers and the office. The negative side of this new technology is security-the Internet is a notoriously unsafe place with hackers, viruses, Trojan horses, and other threats to both home computers and office networks.
Some of the different ways companies can connect their teleworkers to their corporate networks include
Dial-up using 56K modems: This method is more secure than a bare dedicated connection because a new Internet Protocol (IP) address is typically assigned to the home user for each session initiated with the user's Internet service provider (ISP). Speed is slow but cost is low also-typically $50 per month, including both an ISP account and a second phone line. Dial-up users should connect to their corporate local area network (LAN) using an encrypted virtual private network (VPN).
Digital Subscriber Line (DSL): This is emerging as the technology of choice for teleworkers due to its high speed (at least 20 times that of dial-up), low cost (typically $50 per month over existing phone lines), and security (DSL is a dedicated point-to- point connection). Nevertheless, teleworkers using DSL need a firewall for full protection, and a VPN is required as well. The main problem with DSL is that it is usually available only in urban areas.
Cable modem: This is usually not recommended for teleworkers due to the security issues concerning having your home computer on a LAN segment with other cable modem subscribers in your neighborhood. A firewall and VPN are definitely needed in this scenario.
Other less common teleworking scenarios include
Fixed wireless: Availability is limited, coverage is usually in dense urban areas, and the service is costly.
Satellite: The main issue here is the latency introduced by the long distances the signal must traverse between the subscriber and the service provider. Initial equipment is expensive and monthly costs are moderate.
See Also cable modem ,Digital Subscriber Line (DSL) ,Internet ,Internet service provider (ISP) ,modem ,wireless networking
A set of standard application programming interfaces (APIs) developed by Microsoft Corporation and Intel Corporation for accessing telephony services.
Overview
Telephony Application Programming Interface (TAPI) receives programmatic telephony requests from applications and then forwards them to drivers for telephony devices such as telephones, modems, Integrated Services Digital Network (ISDN) equipment, and Private Branch Exchanges (PBXs). TAPI manages various telephony functions for these devices including
Signaling
Call hold and transfer
Call conferencing and call parking
Specialized PBX functions
See Also application programming interface (API) ,modem ,Integrated Services Digital Network (ISDN) ,Private Branch Exchange (PBX)
An Internet standard protocol for executing commands on remote hosts.
Overview
Telnet is an application-layer protocol that is part of the Transmission Control Protocol/Internet Protocol (TCP/IP) suite of protocols. Using Telnet, a user on one IP host can connect to and run text-based commands on a different IP host, provided the user can be authenticated and has suitable privileges. The term telnet is also commonly used to refer to software that implements this protocol on a particular platform or system. The Telnet protocol is defined in RFC 854.
Uses
Telnet is widely used for remote administration of routers, Ethernet switches, and UNIX mail and Web servers. For example, you can use Telnet to connect to a Web server on port 80 to issue Hypertext Transfer Protocol (HTTP) commands to troubleshoot the server or to an Internet mail forwarding host on port 25 to issue Simple Mail Transfer Protocol (SMTP) commands to do the same.
Telnet is one of five common methods for remotely administering Cisco routers and access servers, the other four methods being
Direct serial connection from a local terminal to the RS-232 console port
Remote serial connection through a modem to the RS-343 auxiliary port
Remote IP connection using HTTP from a Web browser
Remote IP connection using Simple Network Management Protocol (SNMP) from an SNMP management console
Implementation
Telnet is a client/server protocol in which a Telnet client on the user's machine issues commands to a Telnet server (for example, a UNIX machine running the telnet daemon or a Microsoft Windows 2000 Server running the Telnet Server service). The Telnet client runs within a command-line window on the client machine; in other words, the user opens a command prompt and types telnet to start the Telnet client service. The user specifies the remote host's name or IP address, enters her credentials for authentication, and then issues commands. Any application that can be run from the command line on the remote Telnet server can also be remotely executed from the Telnet client machine. When the program or command is executed, its output (if any) is returned to the client and displayed within the command-prompt window.
Windows 2000 includes both a Telnet client implemented as a command-line utility and Telnet server software that supports as many as 63 simultaneous client connections but is licensed to provide only up to two simultaneous client connections. If you require support for additional client connections, you should obtain the Microsoft Windows Services for UNIX 2 add-on pack for Windows 2000 Server.
See Also Hypertext Transfer Protocol (HTTP) , router ,Simple Mail Transfer Protocol (SMTP) , UNIX
Traditionally, a device that provides user access to a mainframe computer.
Overview
Terminals originated in the mainframe computing environment, where they were used as front-end devices to allow users to access the processing power of these mainframes in an interactive way. Users would type commands and data into a terminal, and the information they typed would be sent over serial links to the mainframe for processing. Once the mainframe had completed the processing, it would return the results to the user's terminal and display it in the appropriate format.
The earliest terminals were called teletypes (abbreviated TTY) and were essentially electric typewriters through which users would send commands and data to a mainframe and the mainframe would then type the output returned to the user. A terminal that supports only text output is sometimes called an ASCII terminal.
Over the years a number of standards called terminal protocols have been developed that govern their use. The VT-100 terminal developed by Digital Equipment Corporation was a popular ASCII text-based terminal standard that is still used in some places, such as library online catalog systems, which remote users typically access by running a Telnet client over a dial-up connection. IBM's TN3270 terminal protocol is still widely used in IBM mainframe environments, and their TN5250 terminal protocol is popular with their AS/400 midrange computing environments. Other common terminal protocol standards include ANSI (American National Standards Institute), VT52, and VT220.
Implementation
Terminals generally have little or no inherent data- processing power and rely entirely on the back-end system to do the processing. The terminal is responsible only for processing and queuing input from the keyboard (and other additional input devices, such as a mouse), transmitting this information in a format recognized by the back-end host (typically a mainframe, midframe, or PC-based terminal server). The information the user enters on the keyboard is typically transmitted to the mainframe over an RS-232 or RS-423 asynchronous serial connection, but sometimes it is transmitted instead over an Ethernet or a Token Ring local area network (LAN) connection. Once the processing is completed, the output is sent back to the terminal and typically presented on a "green screen" monitor, which is usually in ASCII format on older systems, or by providing a graphical desktop environment in newer terminal server computing platforms. In other words, the application runs in one location (the mainframe) and the user interface is in a different location (the terminal).
That the mainframe traditionally does all the processing explains the origin of the term dumb terminal , which means that a terminal by itself is generally useless unless it is connected to a back-end processing system. However, there are also "smart" or "intelligent" terminals that have various degrees of inherent processing capability.
Terminals can be either local terminals, which are directly connected to their back-end mainframe host through a dedicated serial or shared/switched LAN connection; or remote terminals, which are typically connected over a telephone line using modems at both ends of the connection.
Prospects
The popularity of terminals declined in the late 1980s and early 1990s with the advent of distributed client/server environments and the eclipse of mainframe computing environments. In a client/server environment, data processing is shared between the front-end client computer, usually a full-featured PC with a graphical user interface (GUI) such as Microsoft Windows, and the back-end server, which can be a Windows NT-based server, a Novell NetWare server, an AS/400, or some other system. In the late 1990s, however, the pendulum started to swing back toward terminals with the rising popularity of terminal emulators and PC-based terminal servers. A terminal emulator is hardware or software, or both, that runs on a stripped-down PC with no operating system and causes the PC to function as a terminal, and a terminal server is a back-end server that generates and delivers the user desktop environment to the terminals and performs all the processing. This arrangement allows for low-cost "thin clients" to be used and centralizes system administration at the back end, reducing deployment and management costs associated with a distributed client/server systems environment.
See Also RS-232 ,
A security protocol supported by Cisco routers.
Overview
Terminal Access Controller Access Control System (TACACS) is a family of security protocols used for Authentication, Authorization, and Accounting (AAA). TACACS is similar to the industry standard Remote Authentication Dial-in User Services (RADIUS) security protocol but is more flexible and powerful. In particular, TACACS separates the AAA components and allows them to be used independently of one another. For example, a common scenario employed by Internet service providers (ISPs) is to use RADIUS for authentication and TACACS for authorization and accounting.
The original version of TACACS was developed in the 1980s by the Defense Data Network for MILNET, the U.S. military portion of the Internet. A variation of this protocol called Extended TACACS (XTACACS) was developed in 1990 and standardized in RFC 1492. Cisco Systems then developed a third version called TACACS+ that is not compatible with earlier versions and has enhanced security features that make the earlier versions obsolete. The remainder of this article focuses on the Cisco version TACACS+ because it is the one in general use.
TACACS+ supports up to 16 different privilege levels and a variety of authentication methods, including standard logon, shell logon, Point-to-Point Protocol (PPP), Novell Asynchronous Services Interface (NASI), and AppleTalk Remote Access Protocol (ARAP).
Comparison
Although TACACS+ is similar to RADIUS, there are architectural differences in how the two protocols work. For example, RADIUS is a connectionless protocol that runs over User Datagram Protocol (UDP). In RADIUS the authentication and authorization features are integrated, and only passwords are encrypted. RADIUS also supports only Internet Protocol (IP) as a network transport and has no method for controlling access to which commands can be executed on a RADIUS- enabled router.
By contrast, TACACS+ is a connection-oriented protocol that runs over Transmission Control Protocol (TCP), separates the three components of AAA functionality, supports a wide variety of network transports, and uses packet encryption and router access lists for greater security. TACACS+ also includes more than 50 attribute/value pairs and supports secure virtual private networking (VPN). Despite these advantages, RADIUS is still the more widely deployed of the two protocols due to its being a vendor-independent industry standard, and TACACS+ is more commonly used in Cisco- only shops.
Implementation
In a typical ISP scenario using TACACS+, a dial-in user connects through the Public Switched Telephone Network (PSTN) to a Cisco access server (router) located at the ISP. The connection between the dial-in user and the router uses an authentication protocol such as Password Authentication Protocol (PAP), Challenge Handshake Authentication Protocol (CHAP), or Microsoft CHAP (MS-CHAP) for securely transmitting the user's credentials to the router.
During the authentication process, the access server forwards the user's credentials to a Cisco AAA server, which is also located at the ISP. The communication between the access server and the AAA server employs TACACS+ as its security protocol. Once the AAA server has authenticated the user, it informs the access server to allow the client connection attempt to be accepted and the user then accesses the Internet.
Terminal Access Controller Access Control System (TACACS). How a TACACS+-enabled AAA server is used to authenticate dial-in users by an ISP.
See Also AAA , Challenge Handshake Authentication Protocol (CHAP) ,Internet Protocol (IP) ,Internet service provider (ISP) ,Microsoft Challenge Handshake Authentication Protocol (MS-CHAP) ,Password Authentication Protocol (PAP) ,Point-to-Point Protocol (PPP) ,Public Switched Telephone Network (PSTN) ,Remote Authentication Dial-In User Service (RADIUS) , User Datagram Protocol (UDP), virtual private network (VPN)
Hardware or software, or both, that enables a PC to operate as a terminal.
Overview
Terminal emulators let you use a standard PC to connect to a back-end mainframe or terminal server. The rising popularity of terminal emulators led to the demise of older terminals with their chattery keyboards and green screens. Terminal emulators are typically software packages that run on standard PCs and may include accompanying interface cards to support different kinds of connections such as serial, Ethernet, or Token Ring.
Terminal emulators are often designed to emulate several terminal modes including American National Standards Institute (ANSI), VT52, VT100, VT220, TN3270, and TN5250. Terminal emulators also offer productivity features not supported by older terminals, such as keyboard remapping, support for using scripts and macros to automate tasks, hot-linking of emulator data with desktop applications such as Microsoft Excel, multiple session windows, Web browser interface, and so on.
Marketplace
Microsoft HyperTerminal is one popular terminal emulator that supports common terminal emulation modes and is included with 32-bit Microsoft Windows operating systems. Many other vendors offer terminal emulation products, including E-Term32 from DCSi, CRT from Van Dyke Technologies, HotVT from Datamission, and Softerm Modular TE from Softronics.
Notes
When running a terminal emulator, the emulation mode on the clients must match the terminal mode running on the back-end system for communication to work. If you are trying to connect to an unknown mainframe or other back-end system and your emulator cannot automatically detect the terminal mode needed, try using ANSI mode first. If that fails, try VT100 and other popular terminal modes.
See Also terminal ,TN3270
Generally, a server that provides the back-end support needed for terminals to function.
Overview
A terminal server can be a mainframe system, a UNIX host running X Windows, or a PC-based server running software such as the Terminal Services included with Microsoft Windows 2000 Server, Windows XP, and Windows .NET Server. The function of a terminal server is to generate the desktop environment that is presented to the user of the terminal and to perform all the processing of data submitted by the user.
The main advantages of terminal-based computing over a traditional client/server PC network are
Lower hardware costs: "Thin clients" (special devices or stripped-down PCs) can be used instead of full-featured desktop PCs. For example, Terminal Services of Windows 2000 Server can present a 32-bit Windows user environment on older PCs that lack the hardware requirements for running a local copy of the latest versions of Windows operating systems.
Lower management costs: Operating systems and applications are installed and run only on the back-end terminal servers, which simplifies deployment and troubleshooting and makes administration more centralized. For example, Terminal Services for Windows 2000 Server supports running applications such as Microsoft Office from centralized terminal servers instead of installing them on every desktop computer in the enterprise.
Multiplatform support: Allows the same applications and desktop environments to be presented on a variety of client platforms, including Windows- based PCs, Macintosh computers, UNIX workstations, and other devices.
Notes
Single-port terminal servers are sometimes used in mainframe environments to allow users connected to different controllers to communicate over the corporate local area network (LAN) without needing a dedicated point-to-point communication link. In a typical configuration, the controller is connected to a terminal server by an RS-232 serial connection, and the terminal server is linked to the LAN by an Ethernet interface.
Some vendors produce rack-mountable terminal server devices with 8 or 16 RJ-45 ports that can be used to connect asynchronous terminals to an Ethernet local area network (LAN) running Transmission Control Protocol/Internet Protocol (TCP/IP) or some other network protocol. Such devices can be used to provide terminals (or PCs running terminal emulation software) with access to network file servers or dial-up access to the Internet. Windows-based management software allows these devices to be remotely managed from a PC for viewing and configuring port information. Built-in support for Password Authentication Protocol (PAP), Challenge Handshake Authentication Protocol (CHAP), and Remote Authentication Dial-In User Service (RADIUS) are often included to control user access. Users can dial in to the device, be authenticated, and select a desired host on the LAN they want to communicate with.
See Also terminal ,terminal emulator
A component of Microsoft Windows 2000 Server, Windows XP, and Windows .NET Server that supports terminal-based computing.
Overview
Terminal Services enables users to access the Windows 2000, Windows XP, and Windows .NET Server desktop and run Microsoft Windows applications on remote computers and other terminal devices. Terminal Services enables each of these operating systems to function as a terminal server and provide terminal emulation for a wide range of client computers. By moving all processing to the server, Terminal Services reduces total cost of ownership by
Simplifying system administration by centralizing the installation and management of all applications on the server and supporting full remote administration from a single desktop
Extending the life of legacy hardware by enabling client computers with minimal processing power and memory to run standard Windows applications
Extending the life of legacy operating systems by allowing applications designed for Windows 2000, Windows XP, or Windows .NET Server to run on legacy versions of Windows
Increasing security by using encrypted sessions between clients and servers, by enabling administrators to fully monitor and control user operations by shadowing client sessions from another client computer, and by enabling administrators to input keyboard and mouse actions during client sessions for remote control purposes
Terminal Services. Underlying architecture of how Windows 2000 Terminal Services works.
Implementation
Three components are required for Terminal Services to work:
Terminal server: A Windows 2000, Windows XP, or Windows .NET Server running Terminal Services that provides each client computer with its own Windows desktop.
Terminal Services client: A "thin client" application that displays the Windows 2000, Windows XP, or Windows .NET Server desktop and running applications within a window on the client computer. Terminal Services clients are provided for all versions of Windows, including 32-bit clients that can run on computers running Windows .NET Server, Windows XP, Windows 2000, Windows NT 4, Windows NT 3.51, Windows 98, or Windows 95 on either Intel or Alpha platforms, and a 16-bit client for Windows for Workgroups 3.11. Special client software can also be embedded in devices such as Windows-based terminals and handheld PCs.
Remote Desktop Protocol (RDP): A protocol suite based on the T.120 standard from the International Telecommunication Union (ITU), which provides the basis for communication between the client and the terminal server. RDP takes all keystroke and mouse actions performed by the terminal client, transports them to the terminal server for processing, and returns the display output to the terminal client. RDP employs Transmission Control Protocol/Internet Protocol (TCP/IP) as its underlying network transport.
Notes
To use Terminal Services you must install both Terminal Services and Terminal Services Licensing, and you must specify the directory location of the licensing server database. You can install Terminal Services during setup or afterward using Add/Remove Programs in Control Panel (you should typically install Terminal Services on a member server instead of a domain controller because installation on a domain controller can affect the domain controller's performance as a result of the additional load that Terminal Services places on server processor, memory, and network interface). Once these services are installed, you can configure the terminal server's security to allow users to remotely run multiuser applications, configure user accounts to allow them to log on to the terminal server, create user profiles and home directories if desired, and install Terminal Services client software on client computers. You can install client software either by downloading it across the network or by creating client installation disks for manual installation.
By installing the Citrix MetaFrame add-on, non- Windows clients such as UNIX, Macintosh, and OS/2 Warp can also access a Windows 2000-, Windows XP-, or Windows .NET-based system running Terminal Services to run Windows 2000, Windows XP, or Windows .NET Server applications.
A good rule of thumb is that a terminal server needs an additional 4 to 8 MB of RAM for each additional client it supports. Also, do not run legacy MS-DOS or 16-bit Windows on the terminal server, because this can significantly reduce the number of concurrent users that the server can support and increase the memory requirements for each connected client.
See Also Remote Desktop Protocol (RDP) ,
A device connected to one end of a bus or cable that absorbs signals.
Overview
In a bus-based system, a single wire or series of wire segments connects network components in a chain formation. If the ends of the cable are not terminated, a signal placed on the wire by one component will bounce back and forth between the ends of the cable, hogging the cable and preventing other components from signaling. Terminators eliminate this signal bounce by absorbing the signal after each component has seen it once, allowing other components to place their signals on the cable. By supplying a load equal to the impedance of the cable, the terminator prevents reflections or standing waves from developing on the cable. Terminators also prevent interference caused by signal reflection, which can lead to signal loss. Most communication systems such as networks and computer buses require some form of termination at the ends of the data path, although this is often provided internally by the devices at the ends of the data path.
Types
Terminators can be passive (simple resistors) or active (more complex electronics), depending on the type of bus being terminated. Passive terminators use resistors to provide this impedance matching, while active terminators generally use voltage regulators.
Examples of different types of terminators include the following:
Coaxial cabling terminators: Passive terminators that come in various sizes and use Bayonet-Neill- Concelman (BNC) threading to terminate
RG-58 thinnet cabling for 10Base2 Ethernet networks with termination resistance of 50 ohms
RG-59 cable television terminators with resistance of 75 ohms
RG-62 ARCNET cabling terminators with resistance of 93 ohms
Small Computer System Interface (SCSI) terminators: The ends of a SCSI cable must always be terminated in a chain of SCSI devices. The internal termination is usually supplied by the SCSI adapter card, and the external termination is supplied by the last device in the chain. SCSI terminators can be passive, active, differential, or forced-perfect. Forced- perfect terminators compensate for the differences in impedance along the length of a SCSI bus. Diagnostic terminators analyze and display the condition of the data paths within a SCSI bus and are useful for high-availability uses such as clustering.
Free connectors: Connectors on the hubs at both ends of a series of stackable hubs. These terminators are specific to the type of hub sold by a vendor.
Terminator. Using an ohmmeter to test if thinnet cabling is properly terminated.
Notes
You can test the termination of a long 10Base2 network without having to hunt for the ends of the cable. Simply use an ohmmeter and test the resistance between the central conductor and the shield of any BNC T-connectors (after removing the cable from the network card to which it is attached). If the reading is about 25 ohms, the cable is properly terminated; if the reading is about 50 ohms, one of the terminators is loose or missing. If the cable appears to be properly terminated but network problems persist, remove one of the terminators and use the ohmmeter to test the connection to the T-connector that you exposed. If the result is less than 50 ohms, you probably have a short in the cable; if it is more than 56 ohms, there is probably a loose T-connector somewhere on the network.
See Also 10Base2 , coaxial cabling ,Small Computer System Interface (SCSI) ,
A European standard for digital mobile radio services.
Overview
Terrestrial Trunked Radio (Tetra) is an initiative from the European Telecommunications Standards Institute (ETSI) for a single standardized form of digital mobile communications. Tetra is defined in a memorandum of understanding among equipment vendors, service providers, testing bodies, and regulatory agencies that was laid out in 1994. Tetra consists of two complementary standards:
A standard Time Division Multiple Access (TDMA) cellular communication system for voice and data communication on 25-kilohertz (kHz) channels
A Packet Data Optimized (PDO) protocol for packet-switched data-only transmission at 36 kilobits per second (Kbps) on 25-kHz channels
Tetra includes support for security features such as multilevel authentication and encryption, allows voice and data communication to be combined using the same equipment, and supports multiplexing of up to four channels to provide data rates of up to 144 Kbps. Tetra complements the Global System for Mobile Communications (GSM) cellular communication standard: GSM itself can be considered an extension of the Integrated Services Digital Network (ISDN) to the wireless domain, and Tetra is an extension of ISDN Private Branch Exchange (PBX) systems to the same domain. Tetra thus provides additional communication functionality not built into GSM, such as direct mobile-to- mobile communication that bypasses the communication infrastructure, support for broadcast and group call features, fast call setup, priority call, and so on.
Notes
Because of the recent growth of the Internet and wide demand for high-speed wireless mobile data services, a new high-speed wireless mobile packet-switching system called the Digital Advanced Wireless System (DAWS) is currently being developed by the ETSI to supercede the Tetra PDO standard.
For More Information
Visit the Tetra home page at www.tetramou.com
See Also cellular communications , Digital Advanced Wireless System (DAWS) ,Global System for Mobile Communications (GSM) ,Integrated Services Digital Network (ISDN) ,Private Branch Exchange (PBX) ,standards organizations ,
A general name for equipment used to configure, diagnose, and troubleshoot networking and telecommunications systems.
Overview
Test equipment is invaluable to busy network administrators for troubleshooting local area network (LAN) or wide area network (WAN) connections, to system integrators who install networks and communication services at customer premises, and to wiring and cabling installation service people. You can buy test equipment for dedicated, single-use testing purposes, but multifunction test equipment is more cost effective.
Test equipment comes in all shapes and sizes, from rack-mounted equipment for enterprise troubleshooting, to handheld scanners and packet sniffers, to laptops that run special software and use special Personal Computer Memory Card International Association (PCMCIA)-attached probes. Here are some examples:
Copper cable testers: Typically handheld devices that can test installed copper cabling for compliance with Electronic Industries Association/Telecommunications Industry Association (EIA/TIA) standards for cabling system performance. These are usually multifunction devices that support both coaxial cabling and twisted-pair cabling. Two-way testers enable you to test a cable from both ends.
Fiber-optic cable testers: Usually a separate category from coax/twisted-pair cable testers. These devices might support testing of single-mode fiber- optic cabling, multimode fiber-optic cabling, or both, and provide detailed measurements in decibels for optical link budget (OLB) calculations to ensure that a fiber installation will support the intended equipment layout. A typical fiber tester consists of two modules: a light source for injecting signals into the system at 850 or 1300 nanometers (depending on the type of fiber) and a power meter to measure what comes out the other end. Some devices include both functions and can be used to test fiber that is still on the spool.
Token Ring testers: Test for shorts, opens, and grounds on shielded twisted-pair (STP) cabling in Token Ring installations.
LAN analyzers (sniffers): For troubleshooting problems with LAN protocols at all levels of the Open Systems Interconnection (OSI) reference model protocol stack, from lower-level protocols such as Data Link Control (DLC), Internetwork Packet Exchange/Sequenced Packet Exchange (IPX/SPX), NetBIOS Enhanced User Interface (NetBEUI), and Transmission Control Protocol/Internet Protocol (TCP/IP) to higher-level protocols such as File Transfer Protocol (FTP), Hypertext Transfer Protocol (HTTP), NetBIOS, Server Message Block (SMB), and Simple Mail Transfer Protocol (SMTP). These devices basically capture LAN traffic and allow you to analyze and filter packets that use specific protocols, that are transmitted and received from specific computers, that are portions of a specific communication session between two computers, and so on. Microsoft Network Monitor, which is included with Microsoft Systems Management Server (SMS), Microsoft Windows 2000, Windows .NET Server, and Windows NT, is a software-based sniffer that runs on any PC with a network card and can capture and analyze most forms of LAN traffic.
SCSI testers: Test Small Computer Systems Interface (SCSI) buses for shorts, opens, or improper termination. These are usually dedicated to a specific type of SCSI interface.
ISDN and T1 test equipment: Includes continuity testers, channel testers, and line-quality analyzers for testing Integrated Services Digital Network (ISDN) and T-carrier circuits. They can sample frames to check for jitter and lack of synchronization.
WAN analyzers: Test serial transmission protocols such as RS-232 and V.35, which are used to connect WAN devices such as routers and bridges to Channel Service Unit/Data Service Units (CSU/DSUs). They are typically used to troubleshoot frame relay, High-level Data Link Control (HDLC), Point-to- Point Protocol (PPP), Synchronous Data Link Control (SDLC), Serial Line Internet Protocol (SLIP), Systems Network Architecture (SNA), and X.25 connections. You can connect a WAN analyzer to a serial connection by using a Y-shaped connector called a data tap, which lets you monitor communication without interfering with the data being transmitted.
Notes
Cable testers can trace cables through walls, ceilings, and floors by measuring the length of a cable and telling you whether the cable is terminated, has an open end, is connected to a port on a hub, and so on. You can plug two- way cable testers into a wall plate and test the patch panel to determine which cable connects to the wall plate.
Use a cable tester on a new enhanced Category 5 (Cat5e) cabling installation before you install and configure your Fast Ethernet network equipment. Good- quality cable testers typically test all four pairs of wires in unshielded twisted-pair (UTP) cabling over frequencies of up to 100 megahertz (MHz) or higher, checking cable integrity for shorts and opens, measuring cable segment lengths using time domain reflectometry (TDR), and allowing measurement of attenuation, near-end crosstalk (NEXT), and PowerSum NEXT to an accuracy of 0.1 decibels or better.
See Also cabling , Channel Service Unit/Data Service Unit (CSU/DSU) ,crosstalk ,enhanced Category 5 (Cat5e) cabling ,fiber-optic cabling ,Integrated Services Digital Network (ISDN) ,network troubleshooting ,Open Systems Interconnection (OSI) reference model ,RS-232 ,shielded twisted-pair (STP) cabling ,Small Computer System Interface (SCSI) , unshielded twisted-pair (UTP) cabling, V.35
Stands for Terrestrial Trunked Radio, a European standard for digital mobile radio services.
See Also Terrestrial Trunked Radio (Tetra)
Another name for an ASCII file, a file that contains unformatted ASCII text.
See Also ASCII file
Stands for Trivial File Transfer Protocol, a simple file transfer protocol for Transmission Control Protocol/Internet Protocol (TCP/IP).
See Also Trivial File Transfer Protocol (TFTP)
Another name for thicknet, the thick coaxial cabling used in Standard Ethernet (10Base5) networks.
See Also thicknet
The thick coaxial cabling used in Standard Ethernet (10Base5) networks.
Overview
Thicknet coaxial cabling is usually 3/8 inch in diameter, is fairly rigid, and has an impedance of 50 ohms. It can carry signals up to 500 meters (1640 feet)-hence the designation 10Base5 for "10 megabits per second Base band transmission over 5 hundred meters." Thicknet was commonly used in the 1980s, mainly for Ethernet cabling, but it has been superceded by unshielded twisted-pair (UTP) cabling and fiber-optic cabling.
To connect a computer to a thicknet cable, you attach a vampire tap to the cable. The tap pierces the cable's insulation layers and makes contact with the signal- carrying copper core. The tap is connected to a transceiver, and a drop cable connects the transceiver to an attachment unit interface (AUI) connector on the computer's network interface card (NIC).
See Also 10Base5 ,coaxial cabling ,Ethernet ,fiber-optic cabling ,unshielded twisted-pair (UTP) cabling
A client used for terminal-based computing.
Overview
A thin client is a device or stripped-down PC that has only the hardware necessary to support terminal client software. In terminal-based computing, the terminal client sends keystrokes and mouse movements over the network to a terminal server where actual applications reside. The server processes the client input and returns display data to the client, which displays the results for the user. In true terminal-based computing the server does the processing-the client is basically a "dumb" terminal that supports keyboard/mouse input and video output only. The thin clients used in terminal-based computing contrast with traditional "fat clients" in the form of standard PCs, which have greater hardware requirements, consume more network bandwidth, are more complex to manage, and cost more than thin clients.
The main difference between the newer generation of thin clients and the older mainframe-based dumb terminals is that today's thin clients can use an Ethernet network running Internet Protocol (IP) as their underlying network transport, but legacy dumb terminals ran over dedicated serial connections instead. The main advantage of thin clients over traditional fat clients is manageability-applications and user profiles can be securely and centrally managed on terminal servers locked away in back rooms, and practically nothing can go wrong with the machines on users' desktops other than a loose network connection.
History
The first thin client developed for the Microsoft Windows platform was WinFrame from Citrix Systems, a multiuser client/server terminal application developed in 1995 for the Windows NT 3.51 Server platform. To distinguish Citrix WinFrame clients from legacy dumb terminals, the term Windows-Based Terminal (WBT) was coined. Microsoft Corporation included its own terminal server platform in an edition of its next version of Windows NT, namely Terminal Server Edition for Windows NT 4 Server. Citrix then developed an enhanced version of its own product called MetaFrame. Citrix and Microsoft continue to be the two market leaders in thin-client computing platforms, with Microsoft including its Terminal Services as part of its Windows 2000, Windows XP, and Windows .NET Server platforms and Citrix MetaFrame supporting the Windows, Macintosh, UNIX, and Linux platforms.
Architecture
Despite the similar approach of the Citrix and Microsoft platforms, there are some underlying architectural differences between them. Most significantly, although Microsoft uses its proprietary Remote Desktop Protocol (RDP) for transporting keyboard, mouse, and video information over the network, Citrix uses its own Independent Computing Architecture (ICA) protocol for this purpose. Both of these protocols can operate over any IP network, including local area network (LAN) and dial-up or dedicated wide area network (WAN) connections. Both platforms also include ActiveX controls that allow the client to be any machine running the Microsoft Internet Explorer Web browser (Citrix also includes a plug-in for the Netscape Navigator browser), enabling Windows-based computing within a Web browser interface.
Marketplace
WBTs for enterprise markets come in a variety of formats, including compact desktop units, handheld devices, and even mobile devices such as wireless Personal Digital Assistants (PDAs). A wide variety of vendors produce WBTs that are compatible with Microsoft's RDP and Citrix's ICA architectures, including Wyse Technology, IBM, and many others. Boundless Technologies offers its Capio II terminal, which runs Windows CE, supports Super Video Graphics Array (SVGA) graphics, and has two universal serial bus (USB) ports and built-in 10/100BaseT Ethernet. Network Computing Devices has a similar entry called ThinSTAR. Wyse, which was the first to market with a WBT product, offers the Winterm 3200LE and many other models.
Web-based WBTs are increasingly popular also, because they allow thin clients to run within standard Web browsers. Popular products in this category include Nfuse from Citrix, HobLink from Hob Software, and Microsoft's Windows 2000 Terminal Services Advanced Client (TSAC).
Prospects
Although thin clients reduce the management costs associated with managing desktop PCs, thin clients themselves require licensing and hence have only a small impact on software costs. Some vendors are trying to work around these licensing costs by providing innovative forms of terminal-based access to Windows 2000, Windows XP, and Windows .NET Server terminal servers. For example, Tarantella has a terminal server product called Tarantella that runs on UNIX platforms, emulates multiple WBT clients to a Windows 2000 terminal server, and supports both native 32-bit Windows applications and Java-based clients running within Web browsers.
Notes
Although the term thin client is usually used nowadays in the context of Windows terminal-based computing, other platforms have at times fallen under the banner of thin client, specifically:
Legacy-free PC: Formerly known as the NetPC, this is a revised initiative from Microsoft and PC manufacturers whose aim is to simplify the PC architecture by removal of legacy features such as the Industry Standard Architecture (ISA) bus and to make systems more secure by removing expansion capabilities. Examples of legacy-free PCs in the marketplace include the Compaq iPaq, the IBM NetVista, and the Hewlett-Packard e-pc. Legacy- free PCs tend to be easier to manage than traditional PCs but cost about the same. Legacy-free PCs are not thin clients, however, because they run applications locally and hence need processor, memory, and storage to accomplish this.
Network computer (NC): The NC architecture was created by IBM, Oracle Corporation, and Sun Microsystems as a terminal-based computing platform using a Java-based thin client. The NC architecture has recently re-emerged and scored some success with its new SunRay platform. The NC is a true "thin client" architecture, but it lags in popularity compared to the Windows-based Citrix and Windows 2000 Terminal Services platforms.
Note that the term thin server refers to a rack-mount server having a 1U or 2U format, not a terminal server for a thin client. See the article "rack," elsewhere in this book, for more information about thin servers.
See Also Independent Computing Architecture (ICA) , Internet Protocol (IP) ,rack ,Remote Desktop Protocol (RDP) ,
Also called thinnet, the thin coaxial cabling used for 10Base2 Ethernet networks.
See Also thinnet
The thin coaxial cabling used for 10Base2 Ethernet networks.
Overview
Thinnet cabling is RG-58 coaxial cabling that is 3/16 inch in diameter, is relatively flexible, and has an impedance of 50 ohms. Thinnet uses Bayonet-Neill- Concelman (BNC) connectors to connect cable segments, computers, and concentrators (hubs) together in bus-style networks. Many older hubs, bridges, routers, and other networking devices still have at least one thinnet port for connecting to 10Base2 networks.
Thinnet was often used in the 1980s for workgroup or departmental local area networks (LANs). Thinnet has been superceded by the more popular unshielded twisted-pair (UTP) cabling used in structured wiring deployments for premise cabling. One place where thinnet is still used occasionally, however, is in electrically noisy environments such as shop floors in factories, where electromagnetic interference (EMI) caused by motors, generators, and other heavy equipment can disrupt communication over UTP. Coaxial cabling, with its internal shielding, can easily withstand this kind of noise.
Notes
Thinnet cables must be terminated at both ends. If communication on a thinnet network is down, check the termination points first, then check for loose BNC T-connectors attached to the computers on the network. Note that thinnet cabling can become damaged if it is sharply bent or twisted, so handle it carefully. (It is not nearly as fragile as fiber-optic cabling, however.)
See Also 10Base2 , BNC connector ,bus topology ,coaxial cabling ,electromagnetic interference (EMI) , unshielded twisted-pair (UTP) cabling
Stands for Telecommunications Industry Association, a national trade organization representing all aspects of the telecommunications industry in the United States.
See Also Telecommunications Industry Association (TIA)
A cellular communications technology based on time-division multiplexing (TDM).
Overview
Time Division Multiple Access (TDMA) is used to refer to two systems:
Any digital cellular communications system that employs TDM to enable a single channel to carry multiple conversations simultaneously.
A specific cellular phone system in the United States that is operated by AT&T and is commonly called North American TDMA. More properly speaking, TDMA is the air interface for AT&T Wireless Services.
The first definition is the engineer's, and the second is the popular one from a consumer's point of view. This article will take the broader view of TDMA as a technology and consider the various popular implementations it has achieved.
History
The original analog cellular phone system developed by Bell Laboratories in the late 1970s and widely deployed in the United States was called Advanced Mobile Phone System (AMPS). This technology was based on Frequency Division Multiple Access (FDMA), which assigned one conversation to each channel. The main problems with this system were that FDMA systems supported only a limited number of concurrent users, and, because the channels were narrowly spaced with respect to each other, interference sometimes occurred. To overcome these problems, TDMA was used as the underlying technology for the first all-digital U.S. cellular communications system, called Digital Advanced Mobile Phone System (D-AMPS). The D-AMPS system operated in the same 800 megahertz (MHz) band of the frequency spectrum as AMPS but was able to handle greater numbers of simultaneous conversations and was more immune to interference between channels. D-AMPS is based on the IS-54 standard and is still used in parts of the United States, but newer Code Division Multiple Access (CDMA) cellular systems have become more widely deployed.
Other TDMA-based cellular systems eventually followed, including
Global System for Mobile Communications (GSM): This is the world's most popular cellular communications system, but although it is prevalent in Europe and parts of Asia, it is still behind CDMA systems in the United States. GSM operates at 900 and 1800 MHz in Europe and at 1900 MHz in the United States.
Personal Digital Cellular (PDC): This is the world's second most popular cellular system and is used exclusively in Japan.
Personal Communications Services (PCS): This is an umbrella term for a variety of cellular communication systems that operate at 1900 MHz in the United States. One of the technologies used by PCS is TDMA.
Integrated Digital Enhanced Network (IDEN): This TDMA-based system was developed by Motorola and is used in some parts of the United States.
Implementation
TDMA works by dividing a radio channel in time to create a series of short slots or time intervals, each a small fraction of a second. Signals from different subscribers are then assigned to specific slots, and the whole series of slots is repeated many times per second. The result is that small delays are introduced into conversations, but this happens so quickly that it cannot be noticed by the unaided ear.
Different TDMA-based cellular systems use different slot and cycle times. For example, North American TDMA (that is, D-AMPS) uses 30 kilohertz (KHz)-wide channels segmented into three time slots each 6.67 microseconds long and capable of carrying 320 bits of data per slot. The whole eight-slot frame repeats itself 50 times per second. By contrast, GSM uses eight slots each 0.577 microseconds long and carrying 156 bits of data, cycled at 217 times per second. The result is that a single D-AMPS channel can carry three simultaneous conversations and a GSM channel can carry eight conversations, and the quality of GSM communications is smoother than D-AMPS.
Prospects
Although TDMA systems based on the IS-54 are considered second generation (2G) cellular communication systems, a newer standard IS-136 has been developed as a 2.5G system capable of higher data transfer rates of 43.2 kilobits per second (Kbps).
See Also 2G , 2.5G ,Advanced Mobile Phone Service (AMPS) ,cellular communications ,Code Division Multiple Access (CDMA) ,Digital Advanced Mobile Phone Service (D-AMPS) ,Frequency Division Multiple Access (FDMA) ,Global System for Mobile Communications (GSM) ,Personal Communications Services (PCS) ,
A method for sending several data streams over a single communication path.
Overview
In time-division multiplexing (TDM), data from different input channels is apportioned into fixed-length segments and then multiplexed in round-robin fashion into a single output data stream, which can then be transmitted over a single channel transmission system and then demultiplexed at the destination location. TDM segments can be created by the multiplexer itself or can be inherent in the input channel signals, such as fixed- length frames. For example, if input streams A, B, and C are divided into segments as shown here
A = A1, A2, A3,... B = B1, B2, B3,... C = C1, C2, C3,...
the time-division multiplexed output stream will look like this:
MUX(ABC) = A1, B1, C1, A2, B2, C2, A3, B3, C3,...
One weakness in the TDM approach is that if an input channel does not have anything important to carry for a time, empty segments are inserted into the output stream regardless. For example, if channel A is not transmitting data, one-third of the output channel contains null data and is not being used. You can overcome this weakness by using a more sophisticated multiplexing technique called statistical multiplexing.
Uses
TDM is used in a variety of different networking and telecommunications technologies. In T-carrier transmission, for example, TDM enables a single T1 line to simultaneously carry 24 data channels by interleaving data into portions of a single 193-bit frame. For example, bits 1 through 8 represent channel 1, bits 9 through 16 represent channel 2, and so on to bits 185 through 192 for channel 24, plus one final bit number 193, which is used for link synchronization. This TDM-based framing process occurs 8000 times per second, producing a total throughput for T1 of 1.544 megabits per second (Mbps).
See Also multiplexing , statistical multiplexing (STM) ,
A cable testing technique for finding breaks or shorts in a cable.
Overview
A time domain reflectometer is a device that sends a pulse onto a cable and measures the time that it takes for the reflection to return from a short or break in the cable. (This is analogous to the use of sonar to determine ocean depth.) The time interval between transmission and reception of the signal is called the signal delay; this delay can be used to determine the location of the short or break, typically within a few centimeters, even though the break might be hidden within the cable's jacket and not be visible. You can also use the reflectometer to determine the length of an undamaged cable and identify cables running through walls and false ceilings in a cabling installation.
Most high-quality cable testers can perform time domain reflectometry (TDR) tests in addition to their other functions. Time domain reflectometers are available for testing both copper cabling and fiber-optic cabling.
See Also cabling ,network troubleshooting
Stands for top-level domain, a domain that is directly beneath the root domain in the hierarchical Domain Name System (DNS).
See Also top-level domain (TLD)
Stands for Transport Layer Security, a security protocol based on Secure Sockets Layer (SSL).
See Also Transport Layer Security (TLS)
A form of Telnet used for accessing mainframe hosts over an Internet Protocol (IP) network.
Overview
TN3270 was developed as an alternate to the regular Telnet service for accessing mainframe computers. TN3270 provides a better look and feel than standard Telnet, but its numeric field handling and keyboard interface are somewhat clumsy. TN3270 provides keyboard emulation and block-mode service at the client level, thus freeing the mainframe from translation functions. TN3270 supports workstation emulation only and does not include file-transfer or printer-emulation services. TN3270 originally stood for Telnet 3270 but is never referred to this way anymore.
By using Microsoft Host Integration Server, users running a TN3270 client can connect to mainframe computers using the TN3270 service included with Host Integration Server. TN3270 can also be used to connect clients to AS/400 systems, but the AS/400 systems must translate the 3270 data stream into 5250 format and provide keyboard mapping between the 3270 and 5250 key sequences, a process that consumes additional CPU resources on the AS/400.
See Also Telnet ,TN5250
A form of Telnet used for accessing AS/400 systems over an Internet Protocol (IP) network.
Overview
TN5250 is to the AS/400 computing environment what TN3270 is to the mainframe world. TN5250 offers full 5250 terminal emulation, including hot backup and security features similar to those included with the TN3270 service. TN5250 provides workstation emulation only and does not include file-transfer or printer- emulation services. TN5250 originally stood for Telnet 5250 but is never referred to this way anymore.
A TN5250 service included with Microsoft Host Integration Server lets TN5250 clients connect to AS/400 systems without installing Transmission Control Protocol/Internet Protocol (TCP/IP) on the AS/400. Using Host Integration Server, TN5250 provides workstation emulation that supports almost all the field attributes and keyboard sequences of a "real" SNA 5250 except text assist.
See Also Telnet ,TN3270
A local area network (LAN) technology developed by IBM.
Overview
Token Ring was first developed by IBM in 1984 as an alternative to Ethernet. Token Ring originally operated at 4 megabits per second (Mbps). This speed was later extended to 16 Mbps, which enabled Token Ring to compete favorably for a while with the 10 Mbps speed of standard Ethernet. Over the years, the evolution of Token Ring, however, has not matched that of Ethernet. Fast Ethernet brought speeds of 100 Mbps, and an initiative called High-Speed Token Ring (HSTR) was undertaken jointly by Token Ring vendors IBM, Madge Networks, and Olicom to do the same. But in 1998, in the face of emerging Gigabit Ethernet (GbE) standards, IBM abandoned its HSTR efforts, which spelled the death knell for Token Ring and relegated it to the realm of a legacy technology. Despite this occurrence, there is still a large installed base in some shops, but it seems inevitable that they will have to consider migrating to Ethernet technologies in the near future.
Token Ring. The physical and logical topologies of a Token Ring network.
Token Ring was standardized in the Institute of Electrical and Electronics Engineers (IEEE) 802.5 specifications, which describe a token-passing ring network configured as a physical star topology using structured wiring implemented with twisted-pair cabling and active hubs.
Implementation
In a Token Ring network, stations (computers) are wired in a star formation to a central wiring concentrating unit called a Multistation Access Unit (MAU). Note that the term Multistation Access Unit is sometimes abbreviated as MSAU instead of MAU to distinguish it from media attachment unit , a term used in older Ethernet networking technologies.
The MAU unit concentrates wiring in a star topology but internally forms a logical ring topology over which network traffic can travel. Lobes connect the individual stations to the MAU. The maximum cable length for a lobe is 74 feet (22.5 meters) or 328 feet (100 meters), depending on the cable type, but you can extend this distance up to 1.5 miles (2.4 kilometers) using repeaters designed for Token Ring networks. Note that distances between MAUs and attached stations are usually specified as lobe lengths, which refer to round-trip signal paths. Thus, a station with a lobe length of 655 feet (200 meters) actually uses a cable 328 feet (100 meters) long.
MAUs typically support 8 or 16 connections for attaching lobes. You can extend a Token Ring network by connecting the ring-out port of one MAU to the ring-in port of another MAU to form larger rings that can support larger numbers of stations (stackable MAUs simplify this interconnection process). The maximum number of MAUs that can be interconnected in this way is 33. Some MAUs also support interconnection using fiber-optic cabling to create networks that span a building or even a campus. Most MAUs also support in-band management by using Simple Network Management Protocol (SNMP) plus out-of-band management by using a serial interface.
Token Ring networks come in two types, both of which can operate at 4 or 16 Mbps:
Type 1 Token Ring: Generally uses shielded twisted-pair (STP) cabling with a special data connector developed by IBM for Token Ring installations. However, 16-Mbps MAUs generally have ports for RJ-45 or DB9 connectors.
Type 3 Token Ring: This type uses standard unshielded twisted-pair (UTP) cabling with RJ-45 connectors.
Type 1 Token Ring is often considered more reliable than Type 3, but the larger installed base of UTP cabling made Type 3 an attractive option for many Token Ring installations. Type 1 configurations support as many as 260 stations per ring, while Type 3 can support up to 72 stations per ring. Most MAUs and NICs are dual-speed and can run at either 4 or 16 Mbps, but not both. However, you can use bridges or routers to connect 4-Mbps Token Ring networks to 16-Mbps Token Ring networks.
STP cabling for Type 1 Token Ring comes in nine types, only two of which are common now:
Type 1 cable: Uses two-pair 22-gauge shielded, grounded solid copper wire. Use this type for longer cable runs such as those between wiring closets and work areas. The maximum lobe length is 655 feet (200 meters).
Type 6 cable: Uses two-pair 26-gauge stranded, shielded copper wire and is more flexible (and looks nicer!) than Type 1 cable. Use this type for work areas in which cables will be visible or where equipment will be moved frequently and especially for patch cables. The maximum lobe length is 148 feet (45 meters).
Token Ring stations pass a single data packet called a token from one computer to the next rather than let each node transmit independently, as in a contention-based network such as Ethernet. Only one token can be on the network at a time, so collisions do not occur in Token Ring networks as they do in Ethernet networks. This process is analogous to sending messages to a group of people by passing a hat. In order to pass a token in a Token Ring network, each station must know who its neighbors are and must perform a check to make sure that the circuit is unbroken. Messages containing this information are continually sent around the ring. The token circulates so long as this message is received. To generate the required information, the first station online in the ring assumes the role of Active Monitor Station. It creates the token and is responsible for taking action if the token is lost or damaged. The Active Monitor Station sends out an Active Monitor Present frame every seven seconds to the next node down the line. Each node in turn informs its downstream neighbor that it is its Nearest Active Upstream Neighbor. An error- detection process called beaconing occurs if the ring breaks and the token fails to circulate. If the Active Monitor Station fails, another station assumes its role of monitoring the status of the network and generating a new token if the existing one is lost.
If a station wants to transmit data over the network, it waits until the token comes by; if the token has not been claimed by another station, it claims the token and inverts the monitor setting bit to mark it "busy" so that no other station can claim the token for a predefined but variable amount of time. The originating station then removes the last byte from the token (called the delimiter byte), appends data to the token, and appends the delimiter byte to the end to form a frame of variable length (up to 8000 bytes). The token with data circulates around the ring in one direction from station to station. (Each station acts as a repeater to regenerate and forward the token.) When it returns to the originating station, the token and the data are removed and a new token is generated and placed onto the network.
Notes
You can get both types of cable in an adapter cable version (terminated at one end with an IBM data connector and at the other end with a DB9 male connector) or a patch-panel version (terminated at both ends with data connectors). Use patch panel cables to connect MAUs, and use adapter cables to connect stations to MAUs. You can also get baluns, which can convert Type 1 IBM cabling to UTP cabling to connect different Token Ring types, and you can get special adapters that allow data connectors to be connected to RJ-45 ports so that you can use installed UTP cabling with Type 1 MAUs.
Some network interface cards (NICs) for Token Ring networking support software-configurable physical layer addressing, but note that all Token Ring NICs must have unique MAC addresses for communications to work properly on a Token Ring network.
The following table provides suggestions for troubleshooting common Token Ring network problems.
Problem | Suggestion |
Mismatched ring speed | Be sure that all connected stations use 4 Mbps or that all use 16 Mbps. Do not mix stations of different speeds. |
Stations cannot receive | Check cables and reset the MAU. |
Conflicting MAC addresses | Use NIC configuration software to change the MAC address on one of the conflicting computers. |
Traffic congestion on the network | Segment the network by using a bridge or a router. |
See Also Ethernet ,Fast Ethernet ,Gigabit Ethernet (GbE) ,local area network (LAN) ,MAC address ,Multistation Access Unit (MAU or MSAU) ,network interface card (NIC) ,shielded twisted-pair (STP) cabling ,unshielded twisted-pair (UTP) cabling
A domain that is directly beneath the root domain in the hierarchical Domain Name System (DNS).
Overview
Top-level domains (TLDs) are relatively few in number and are used to identify broad classes of Internet services. The number of TLDs is controlled by the Internet Corporation for Assigned Names and Numbers (ICANN), which keeps this number small to maintain the efficiency of the hierarchical DNS naming system. Name resolution for TLDs is provided by the Internet's 13 root name servers and 10 top-level domain servers.
The various TLDs are listed in the following table. Several additional TLDs, such as .name, .pro, .museum, .aero, and .coop, have been approved by ICANN. The first three TLDs are managed commercially by domain name registrars, and their use varies widely. For example, although .net was originally intended for networking companies only, even some personal home pages use this domain.
Domain | Description |
.com | Commercial businesses and miscellaneous other uses |
.net | Networking and telecommunications companies |
.org | Nonprofit organizations |
.edu | Four-year degree-granting universities and colleges in North America |
.gov | U.S. federal government |
.mil | U.S. military use only |
.int | Organizations established by international treaty |
.biz | Businesses |
.info | General purpose |
In addition to the domains listed in the table, countries as well as states and provinces within countries are identified by two-letter country codes. For example, .uk is the top-level domain for the United Kingdom, .ca is the top-level domain for Canada, and mb.ca is the top-level domain for the province of Manitoba in Canada. Although the .com domain is by far the most popular one today due to the way it is marketed, many businesses are forced to use other domains such as .net, .biz, or their country domain because of the shortage of commercial top-level domains.
Notes
A special domain called in-addr.arpa is used for reverse DNS name lookups (resolving a host name given the host's Internet Protocol [IP] address).
See Also country code ,Domain Name System (DNS) ,in-addr.arpa ,Internet ,Internet Corporation for Assigned Names and Numbers (ICANN) ,root name server
The physical layout of computers, cables, switches, routers, and other components of a network.
Overview
The term topology can refer to either a network's physical topology, which is the actual physical layout or pattern of the cabling, or its logical topology, which is the path that signals actually take around the network. This difference is most evident in Token Ring networks, whose cabling is physically arranged in a star but whose signal flows in a ring from one component to the next. The term topology without any further description is usually assumed to mean the physical layout. The term comes from topos, the Greek word for "place."
When you design a network, your choice of topology will be determined by the network's size, architecture, cost, and management. Basic network topologies include the following:
Bus topology: The stations are connected in a linear fashion. An example is the 10Base2 form of Ethernet.
Star topology: The stations are connected to a single concentrating device called a hub (Ethernet) or a Multistation Access Unit, or MAU (Token Ring physical topology).
Ring topology: The stations are connected in a ring. Examples are Fiber Distributed Data Interface, or FDDI (logical and physical ring), and Token Ring (logical ring and physical star).
Mesh topology: The stations are connected in a complex, redundant pattern. This topology is generally used only in wide area networks (WANs) in which different networks are connected using routers.
Variations of these basic topologies include the following:
Star bus topology: Consists of many star networks whose concentrators (hubs) are connected in a linear bus fashion
Star-wired topology or cascaded-star topology: Consists of star networks whose hubs are joined in star formation to other hubs, forming a kind of tree-shaped network with the main hub at the top
See Also bus topology ,mesh topology ,ring topology ,star topology
On Microsoft Windows platforms, a utility used for troubleshooting communication on routed Internet Protocol (IP) networks such as the Internet. The corresponding utility on UNIX platforms is known as traceroute.
Overview
Tracert (or traceroute) is used to "trace the route" across an IP internetwork from a local host to a remote one. Tracert uses Internet Control Message Protocol (ICMP) echo packets similar to the way ping operates. When an attempt is made to use tracert to trace the route to a remote IP host, a series of ICMP echo packets are assigned a steadily increasing Time to Live (TTL) to test network connectivity with routers and IP hosts that are farther away along the route. This continues until either connectivity fails or the target host is finally contacted and successfully responds.
Examples
If you run
tracert research.microsoft.com
from Winnipeg through your local Internet service provider (ISP), you might get a display similar to the following, depending on the route the packets take at that moment:
Tracing route to research.microsoft.com [131.107.65.14] over a maximum of 30 hops:
1 100 ms 100 ms 110 ms wnpgas04.mts.net [205.200.55.1]
2 100 ms 90 ms 100 ms 205.200.55.6
3 90 ms 100 ms 110 ms wnpgbr01-g11-102.mts.net [205.200.28.82]
4 110 ms 100 ms 100 ms dis4-winnipeg32-pos11-0.in.bellnexxia.net [206.108.110.5]
5 120 ms 100 ms 100 ms core2-winnipeg32-pos6-2.in.bellnexxia.net [206.108.102.129]
6 120 ms 130 ms 120 ms core2-toronto12-pos10-1.in.bellnexxia.net [206.108.97.29]
7 120 ms 130 ms 120 ms core3-toronto12-pos6-0.in.bellnexxia.net [64.230.242.201]
8 180 ms 180 ms 181 ms core2-vancouver-pos10-2.in.bellnexxia.net [206.108.101.182]
9 191 ms 180 ms 190 ms core2-seattle-pos12-0.in.bellnexxia.net [206.108.102.209]
10 180 ms 190 ms 190 ms bx3-seattle-pos5-0.in.bellnexxia.net [206.108.102.202]
11 180 ms 190 ms 190 ms microsoft-gw.core1-seattle-pos6-2.in. bellnexxia.net [206.108.108.134]
12 180 ms 190 ms 190 ms 207.46.190.161
13 180 ms 1042 ms 180 ms iuscixtukc1202-ge-5-0.msft.net [207.46.129.48]
14 191 ms 190 ms 190 ms 207.46.168.122
15 181 ms 190 ms 190 ms 131.107.33.50
16 1142 ms 1021 ms 191 ms iusdinetdc7507-fe-0-1-0.msft.net [131.107.34.135]
17 190 ms 181 ms 190 ms 131.107.40.70
18 190 ms 191 ms 190 ms research.microsoft.com [131.107.65.14]
Trace complete.
Note that the destination host was finally reached after a distance of 18 hops, and note the gradually increasing response times.
See Also network troubleshooting ,ping
A method of coordinating a series of changes to a set of resources distributed over the network.
Overview
Transactions are units of work that must succeed or fail as a whole-a transaction can never partially succeed. If a transaction fails while only partially completed, the transaction is rolled back to the beginning. An example is a credit card purchase: The store requests the purchase amount from the credit card company, the company distributes the funds to the store, and the company bills the purchaser. If any part of the transaction fails, the entire transaction must fail to prevent money from being lost.
Component Services on Microsoft Windows 2000 (or Microsoft Transaction Server on Windows NT), a tool that provides the underlying support, or "plumbing," for creating scalable, distributed, transactional Web applications, provides failure isolations and mechanisms for recovering failed transactions and can run components of transactions as isolated processes for greater crash protection. Component Services uses the Distributed Component Object Model (DCOM) programming architecture for communication between components on Microsoft Windows networks.
See Also Distributed Component Object Model (DCOM)
A technology that provides fault tolerance and crash recovery for critical database files.
Overview
Transaction logs are used in products such as the Microsoft Exchange Server directory services database and information store and Microsoft SQL Server. Using Exchange Server as an example, data is written to transaction log files before it is applied to the directory or information store databases. This improves the performance of write operations to the Exchange databases. In Exchange, you might have several transaction logs in your database directory. When a database is backed up, the transaction logs are then purged.
Transaction logs also play an important role in providing fault tolerance and recoverability for databases. If a system crash corrupts the database files, you can use the transaction logs (if they are intact) to restore all changes to the database since the last backup. Transaction logs make online incremental and differential backups possible. Without transaction logs, you would be able to perform full backups only when backing up databases online.
Microsoft Corporation's version of Structured Query Language (SQL) used by Microsoft SQL Server.
Overview
Transact-SQL (sometimes called T-SQL) is a superset of the SQL-92 standard developed by the American National Standards Institute (ANSI) and the International Organization for Standards (ISO). Transact-SQL includes all the features of standard SQL plus several enhancements, including
Conditional programming constructs such as IF and WHILE
System stored procedures
Transact-SQL has continued to evolve with each new version of SQL Server released by Microsoft and is a powerful data manipulation language for relational database management systems (RDBMS).
See Also American National Standards Institute (ANSI) ,database ,International Organization for Standardization (ISO) ,SQL Server ,Structured Query Language (SQL)
An electronic device for connecting a computer to a baseband transmission network so that the computer can transmit and receive signals on the network.
Overview
In the 1980s transceivers were often separate devices attached to thicknet cabling using vampire taps, but today most network interface cards (NICs) have onboard transceivers built into them. Some Fast Ethernet NICs also have a media independent interface (MII) to which an external transceiver can be connected to provide different kinds of 100-megabits per second (Mbps) networking. This allows greater flexibility in your networking options. For example, 100BaseTX transceivers have an RJ-45 port for connecting unshielded twisted-pair (UTP) cabling, and 100BaseFX transceivers have an SC-type port for connecting fiber-optic cabling.
See Also baseband transmission , network interface card (NIC) ,
Also called a drop cable, a cable connecting a computer's network interface card (NIC) to a transceiver attached to a thicknet cable in Standard Ethernet.
See Also drop cable
A transport layer protocol of the Transmission Control Protocol/Internet Protocol (TCP/IP) suite.
Overview
Transmission Control Protocol (TCP) is one of two transport layer protocols used by TCP/IP, the other being User Datagram Protocol (UDP). Although UDP supports only unreliable, connectionless network communications, TCP provides support for reliable, connection-oriented delivery of Internet Protocol (IP) packets. TCP supports only point-to-point communications between two hosts and does not support multipoint communications as UDP does.
Some of the features of TCP communications include
Byte stream: TCP accepts a stream of bytes from application level protocols and apportions it into TCP packets without regard to application-level message boundaries within the stream.
Connection-oriented: Before transferring packets, TCP negotiates a connection between sending and receiving hosts using a process called a TCP Three- Way Handshake. TCP connections are also closed using the same process, and connections are maintained using a keep-alive process to ensure that they do not unnecessarily time out. These procedures enable TCP to guarantee that transmitted data will be delivered to its targeted destination.
Full-duplex: A TCP connection consists of two logical pipes for transmitting packets in opposite directions.
Reliable: All TCP packets within a particular byte stream (part of a specific communication session) are sequenced to ensure that the byte stream can be properly reconstructed at the destination. Packets that successfully arrive at their destination cause acknowledgements (ACKs) to be generated so the sending host will know that delivery has been successful. Packets that arrive out of order are buffered, and missing packets are retransmitted after a period of time when the sending host determines that no acknowledgements have been received for these packets. Sender-side and receiver-side flow control are implemented to prevent loss of packets when buffers are full and to eliminate subsequent unnecessary retransmissions. In addition, TCP checksums are included to enable the receiving host to verify the bit-level integrity of the transmission.
Notes
Microsoft Corporation's implementation of TCP on its Microsoft Windows 2000, Windows XP, and Windows .NET Server platforms include support for advanced features such as self-tuning to ensure that data is sent at a speed optimal for the receiving host, dead gateway detection to ensure that inoperative gateways do not hinder packet delivery, and checksums for ensuring error-free delivery.
See Also ACK , connectionless protocol ,connection-oriented protocol ,Internet Protocol (IP) , User Datagram Protocol (UDP)
An industry-standard protocol suite forming the basis of the Internet.
Overview
Transmission Control Protocol/Internet Protocol (TCP/IP) was developed in the 1970s and 1980s as a standard protocol for linking hosts and networks into wide area networks (WANs). TCP/IP is an open networking standard that is independent from underlying physical network transport mechanisms. It uses a simple addressing scheme called IP addresses that allow billions of individual hosts to communicate with one another on the Internet. TCP/IP is also a routable protocol that is suitable for connecting dissimilar systems (such as Microsoft Windows and UNIX hosts) in heterogeneous networks and is the most common network transport in use today.
TCP/IP is a constantly evolving protocol suite whose development is steered by such bodies as the Internet Society (ISOC), the Internet Architecture Board (IAB), and the Internet Engineering Task Force (IETF). The various protocols, addressing schemes, and concepts of TCP/IP are defined in a series of documents called Requests for Comments (RFCs) issued by the IETF under an open standards process.
The foundation of the TCP/IP protocol suite is the Internet Protocol (IP), which provides the addressing scheme and supports routing of traffic between networks. The current version of IP is called IPv4 (Internet Protocol version 4) and uses a 32-bit addressing scheme. Due to the explosion of popularity of the Internet in recent years, this addressing scheme is viewed as inadequate to handle the Internet's future growth. As a result, a new version called IPv6 is likely to be widely implemented over the next several years.
Architecture
As shown in the diagram, TCP/IP has a layered architecture consisting of four distinct operational layers. These four layers map loosely to the seven layers of the Open Systems Interconnection (OSI) reference model. The four-layer TCP/IP architecture is sometimes referred to as the DoD Model because TCP/IP was developed in connection with the ARPANET project of the U.S. Department of Defense (DoD). Each layer of the TCP/IP protocol suite has its associated component protocols, the most important of which are listed here:
Application layer protocols: These are responsible for application-level access to TCP/IP networking services. These include Dynamic Host Configuration Protocol (DHCP), Domain Name System (DNS), Hypertext Transfer Protocol (HTTP), File Transfer Protocol (FTP), Telnet, Simple Mail Transfer Protocol (SMTP), Simple Network Management Protocol (SNMP), and numerous others. In the Microsoft implementation of TCP/IP, application layer protocols interact with transport layer protocols by using either Windows Sockets or NetBIOS over TCP/IP (NetBT).
Transport layer protocols: These establish communication through connection-oriented sessions and connectionless broadcasts. Protocols at this layer include Transmission Control Protocol (TCP) and User Datagram Protocol (UDP).
Internet layer protocols: These are responsible for routing and encapsulation into IP packets. Protocols at this layer include Internet Protocol (IP), Address Resolution Protocol (ARP), Internet Control Message Protocol (ICMP), and Internet Group Management Protocol (IGMP).
Network layer protocols: These place frames on the network. TCP/IP can operate over a wide variety of network transports include the various local area network (LAN) architectures (such as Ethernet and Token Ring) and WAN telecommunication service technologies, including dial-up modem connections over the Public Switched Telephone Network (PSTN), Integrated Services Digital Network (ISDN), and Asynchronous Transfer Mode (ATM) networks.
TCP/IP employs two naming schemes to identify hosts and networks on an internetwork:
IP addresses: These are logical 32-bit (4-byte) numeric addresses usually written in the form w.x.y.z . Using an associated subnet mask, IP addresses are split into two portions, a network ID that uniquely identifies the local network on the internetwork and a host ID that uniquely identifies the host on the local network. For example, the IP address 205.116.8.44 is partitioned using the subnet mask 255.255.255.0 into the network ID 25.116.8.0 and the host ID 44. IP addresses are the basic or primary way of identifying hosts and networks on an internetwork; they can either be assigned to hosts manually as static addresses or automatically using DHCP as dynamic addresses.
Fully qualified domain names (FQDNs): These are alphanumeric names generally expressed in the form <host_name>.<domain_name> where <domain_name> identifies the particular network to which the host belongs and <host_name> uniquely identifies the host on the specific network. FQDNs are based on a hierarchical worldwide naming system called the Domain Name System (DNS). As an example, the FQDN server12. microsoft.com represents a host named server12 that belongs to a network whose domain name is microsoft.com. This microsoft.com domain is a second-level domain that belongs to the top-level domain named .com, which itself belongs to the root DNS domain named "." (dot). FQDNs are essentially "friendly" names that are easier to remember than IP addresses. For TCP/IP communications to take place, however, FQDNs must first be resolved into their associated IP addresses by using either a DNS server called a name server or using a hosts file stored on the local machine.
Transmission Control Protocol/Internet Protocol (TCP/IP). How the four layers of the DoD TCP/IP model map to the seven-layer OSI reference model.
See Also Address Resolution Protocol (ARP) , Asynchronous Transfer Mode (ATM) ,Domain Name System (DNS) ,Dynamic Host Configuration Protocol (DHCP) ,Ethernet ,File Transfer Protocol (FTP) ,fully qualified domain name (FQDN) ,hosts file ,Hypertext Transfer Protocol (HTTP) ,Integrated Services Digital Network (ISDN) ,Internet ,Internet Architecture Board (IAB) ,Internet Control Message Protocol (ICMP) ,Internet Engineering Task Force (IETF) ,Internet Group Management Protocol (IGMP) ,Internet Protocol (IP) ,Internet Society (ISOC) ,IP address ,NetBIOS over TCP/IP (NetBT) ,Open Systems Interconnection (OSI) reference model ,Public Switched Telephone Network (PSTN) ,Request for Comments (RFC) ,Simple Mail Transfer Protocol (SMTP) ,Simple Network Management Protocol (SNMP) ,subnet mask , User Datagram Protocol (UDP), Windows Sockets
Layer 4 of the Open Systems Interconnection (OSI) reference model.
Overview
The transport layer is responsible for providing reliable transport services to the upper-layer protocols. These services include:
Flow control to ensure that the transmitting device does not send more data than the receiving device can handle
Packet sequencing for segmentation of data packets and remote reassembly
Error handling and acknowledgments to ensure that data is retransmitted when required
Multiplexing for combining data from several sources for transmission over one data path
Virtual circuits for establishing sessions between communicating stations
Notes
Transmission Control Protocol (TCP) resides at the equivalent of the OSI transport layer in the Transmission Control Protocol/Internet Protocol (TCP/IP) suite of protocols.
See Also Open Systems Interconnection (OSI) reference model ,
A security protocol based on Secure Sockets Layer (SSL).
Overview
Transport Layer Security (TLS) is based on SSL 3 and is very similar in architecture and operation to that protocol. Netscape Communications originally developed SSL in 1993 to provide secure communications over the Internet for Hypertext Transfer Protocol (HTTP) traffic. SSL included support for public and symmetric key cryptography, two-way encrypted authentication, support for anonymous connections, client/server negotiation of the encryption algorithm to be used, and message integrity using digital certificates.
TLS supports all these features of SSL and provides services for secure authentication, data integrity, and confidentiality. TLS is used to secure HTTP, Simple Mail Transfer Protocol (SMTP), and other forms of Internet traffic.
TLS is defined in RFC 2246. A variant of TLS called EAP-TLS that uses the Extensible Authentication Protocol (EAP) extension to Point-to-Point Protocol (PPP) is defined in RFC 2716.
See Also Extensible Authentication Protocol (EAP) ,Hypertext Transfer Protocol (HTTP) ,Point-to-Point Protocol (PPP) ,public key cryptography ,Secure Sockets Layer (SSL) ,Simple Mail Transfer Protocol (SMTP)
Also called a domain tree, a hierarchical grouping of Microsoft Windows 2000 or Windows .NET Server domains.
See Also domain tree
A simple file transfer protocol for Transmission Control Protocol/Internet Protocol (TCP/IP).
Overview
Trivial File Transfer Protocol (TFTP) is a simple file transfer protocol that differs from the more popular File Transfer Protocol (FTP) mainly in that it does not support any form of authentication. TFTP copies files to and from remote hosts by using the User Datagram Protocol (UDP). The remote host must be running the TFTP service or daemon for the TFTP client to be able to communicate with it. TFTP is defined in RFC 1350.
Uses
One place where TFTP is sometimes used is in UNIX environments where the bootstrap protocol (BOOTP) is used for booting diskless workstations. In this scenario, TFTP is used to download the boot disk image from the BOOTP server to the workstation. Another use for TFTP is in Cisco router networking where TFTP can be used to upload or download router configuration information or even perform a flash install of a new version of Cisco Systems' Internetwork Operating System (IOS).
Notes
The Microsoft Windows 2000 and Windows .NET Server platforms include both a command-line TFTP client and an optional TFTP service called the Trivial File Transfer Protocol Daemon (TFTPD) that is installed when the Remote Installation Services component is enabled.
See Also bootstrap protocol (BOOTP) ,File Transfer Protocol (FTP) ,Internetwork Operating System (IOS) ,router ,User Datagram Protocol (UDP)
Any method for aggregating multiple physical network links into a single logical link.
Overview
Trunking provides a way of overcoming the bandwidth limitations of a single physical network link. Trunking is generally employed in three contexts:
In switched Ethernet networking, trunking can be used in either switch-switch or switch-server connections to relieve traffic congestion by providing increased bandwidth.
In remote access and wide area networking, trunking is often used to aggregate multiple wide area network (WAN) links into a single fat pipe.
In telecommunications, telcos sometimes use trunking to aggregate multiple Digital Subscriber Line (DSL) connections for transmission over T1 lines using Asynchronous Transfer Mode (ATM).
The Institute of Electrical and Electronics Engineers (IEEE) 802.3ad standard ensures interoperability among Fast Ethernet and Gigabit Ethernet (GbE) switches that support trunking.
Implementation
Looking specifically at trunking in switched Ethernet networks, trunking is essentially a form of inverse multiplexing that can be either hardware-based or software- based in its implementation. Trunking was originally developed to reduce congestion in switch-switch connections in switched local area network (LAN) environments. By aggregating several 100-megabit-per-second (Mbps) links between Fast Ethernet switches, for example, you can achieve data rates of 300 or 400 Mbps between the switches to accommodate network backbone traffic. In a full-duplex configuration, this means rates of 600 or 800 Mbps, which rivals the more expensive GbE technology and gives new life to old switches. Not only is it often more economical to trunk Fast Ethernet lines than to upgrade to GbE, but trunked Fast Ethernet cable runs can go farther than GbE cable runs. However, in certain situations trunking does not improve matters. For example, trunking cannot speed up server-to-server backups. GbE switches can be similarly joined for increased backbone capacity in congested enterprise networks. Note, however, that although the theoretical speed for quadruple-trunked full-duplex Fast Ethernet connections is 800 Mbps, in practice the maximum achievable rate is about 560 Mbps because of traffic overhead. Note that switches must be intelligent if they are to support trunked connections properly, so check your switch documentation before you attempt to implement trunking on your network.
Trunking can also be implemented in switch-server connections so that multiple connections to a single server can be aggregated. This form of trunking can be purely software based or can be implemented as a combination of both hardware and software. For example, trunking software installed on multiple network interface cards (NICs) in the server automatically handles load balancing across the various server interfaces and can remove an interface from the trunking group if the interface goes down. This provides increased bandwidth between the server and the switch and ensures fault-tolerant operation. Note that software-based trunking adds an overhead of up to 5 percent to the server's CPU, depending on the software and the NIC used. Look for special NICs from trunking software vendors with on-board processors that can run the trunking software and thus reduce the load on the CPU. Also, do not mix and match trunking software or hardware from different vendors in a single trunking group.
Trunking. Two forms of trunking used in switched Ethernet networks.
There are two basic approaches to how trunking can be implemented:
Symmetrical trunking: Allows any port in a trunking group to transmit packets to any other port. Full-duplex connections are thus supported over all links in the group. For example, a server can both transmit and receive data at 400 Mbps in a trunked group of four interfaces and one switch.
Asymmetrical trunking: Allows any port in a trunking group to transmit packets but allows only one port (the port on the switch) to receive packets. The server can transmit data at 400 Mbps but can receive data at only 100 Mbps.
Notes
Trunking by itself is limited to point-to-point connections between two switches or between a switch and a server. However, you can use the Multipoint Link Aggregation (MPLA) technology developed by 3Com Corporation to aggregate physical links connected to different switches into a single logical link. MPLA thus supports multipath trunking between multiple switches and servers, giving network administrators flexibility in configuring their hardware for optimal traffic servicing. Other vendors are working on similar multipath trunking technologies, but standards are still developing in this arena.
See Also 802.3ad , Digital Subscriber Line (DSL) ,Ethernet ,Ethernet switch ,Fast Ethernet ,Gigabit Ethernet (GbE) ,network interface card (NIC) ,
A secure communication channel between two domains in Microsoft Windows NT, Windows 2000, or Windows .NET Server.
Overview
Trust relationships allow users in one domain to access resources in another domain. Trusts work by having one domain trust the authority of the other domain to authenticate its user accounts.
Windows NT trusts, which are based on Windows NT Challenge/Response Authentication, are managed by the Windows NT Directory Services (NTDS). In Windows NT, trusts are one-way-the trusting domain (or resource domain) trusts the trusted domain (or accounts domain). This means that global users in the trusted domain can be authenticated for accessing resources in the trusting domain. Global users from the trusted domain can log on to any computer in either domain and can access resources in either domain if they have the appropriate permissions. Windows NT trusts are also nontransitive. In other words, if domain A trusts domain B and domain B trusts domain C, it is not true that domain A trusts domain C. If you want to establish a two-way trust between two Windows NT domains, you must create two trusts, one in each direction.
Administrators can set up trust relationships between domains by using the Policies menu in User Manager for Domains. The administrator on the accounts domain should permit the trust first, and then the administrator on the resource domain should complete the trust. Only global accounts (global users and global groups) can cross trusts. By using trusts, you can join Windows NT domains into a variety of domain models, including the complete trust model, the master domain model, and the multiple master domain model. You can join domains to support 100,000 or more users for enterprise-level networks.
Trust . How trust relationships work in Windows NT and Windows 2000.
In Windows 2000 and Windows .NET Server, trusts are managed by Active Directory directory service and are based on the Kerberos v5 security protocol. These trusts are always two-way-in other words, if domain A trusts domain B, users in either domain can access resources in the other domain if they have the appropriate permissions. These trusts are also transitive-in other words, if domain A trusts domain B and domain B trusts domain C, domain A also trusts domain C. Trusts are much easier to manage on these platforms than earlier Windows NT trusts, primarily because there are far fewer trusts to manage. This is because Windows 2000 and Windows .NET Server domains are combined into hierarchical structures called domain trees. All users in a domain tree can access resources in any domain of the tree if they have suitable permissions. In Windows 2000 and Windows .NET Server, you can also use another type of trust called an explicit trust, which is a one-way trust similar to that implemented in Windows NT, to form a trust relationship between two domain forests.
See Also Active Directory , domain (DNS) ,domain tree ,Kerberos ,
The original name for Remote Desktop Protocol, a protocol for terminal-based computing.
See Also Remote Desktop Protocol (RDP)
A method for transporting packets of one network protocol over a different network protocol.
Overview
Tunneling is a way of using one network infrastructure (called the transit network) for carrying traffic for a different network. This is done by encapsulating the packets of the sending node in frames of the transit network and adding a suitable header to route the frame across the transit network to the receiving node. When the encapsulated frame arrives at the receiving node, it is de-encapsulated so the node can read it. The two nodes (sending and receiving) are called the tunnel endpoints, and the path over which encapsulated frames are routed across the transit network is called the tunnel. In addition to encapsulating traffic, most tunneling technologies also encrypt traffic for greater security as it travels over the transit network, usually an intermediate public network such as the Internet.
Types
Tunneling is widely used as a wide-area networking (WAN) technology for connecting networks using an intermediate public network such as the Internet. Some common examples of tunneling technologies include the following:
IPX over IP: Here Internetwork Packet Exchange (IPX) packets are encapsulated in Internet Protocol (IP) datagrams to enable them to be routed over an IP internetwork such as the Internet. This process allows legacy NetWare 3.x networks to communicate over IP.
SNA over IP: Here Systems Network Architecture (SNA) traffic is encapsulated in IP using User Datagram Protocol (UDP) headers, a process described in RFC 1795 and also known as Data Link Switching (DLSw).
Point-to-Point Tunneling Protocol (PPTP): This is a Microsoft Corporation protocol for tunneling IP, IPX, and NetBIOS Enhanced User Interface (NetBEUI) traffic over the Internet.
Layer-2 Tunneling Protocol (L2TP): This protocol supports tunneling of IP, IPX, and NetBEUI traffic over any point-to-point datagram delivery service including X.25, frame relay, Asynchronous Transfer Mode (ATM), and IP.
IP Security (IPsec): This protocol has a tunnel mode that allows IP traffic to be encapsulated and encrypted for transmission over a public IP network such as the Internet.
See Also Internet ,Internet Protocol (IP) ,Internet Protocol Security (IPsec) ,Internetwork Packet Exchange (IPX) ,Layer 2 Tunneling Protocol (L2TP) ,Point-to-Point Tunneling Protocol (PPTP) ,Systems Network Architecture (SNA) ,wide area network (WAN)
A form of coaxial cabling with twin central conducting cores.
Overview
Twinax cabling typically uses 20 AWG stranded copper conductors, has an outside diameter of 1/3 inch, and comes with either a polyvinyl chloride or plenum jacket. Twinax cabling typically has an impedance of 80 to 100 ohms. Twinax cabling is used primarily for connecting IBM System 3X or AS/400 systems to 5250 terminals.
Notes
To extend a twinax connection over long distances, use a repeater. Twinax repeaters can typically transmit signals up to 1 mile (1.6 kilometers) over unshielded twisted-pair (UTP) cabling and over longer distances using duplex fiber-optic cabling. One repeater is required at both ends of the connection.
Use a multiport repeater (hub) to connect several terminals over a single connection to an AS/400 or System 3X host. You can use twinax-to-RJ-45 baluns to connect the terminals and mainframe host to the hub by using UTP cabling. Some repeaters have RJ-11 ports for extending twinax connections over standard telephone cabling. Twinax cabling is traditionally used in a daisy-chained topology, but if you use a multiport repeater, you can also use a star topology configuration.
See Also cabling ,coaxial cabling ,repeater ,unshielded twisted-pair (UTP) cabling
Copper wire cabling consisting of multiple wires twisted together.
Overview
In computer networking and telecommunications, twisted-pair cabling may consist of from one to four pairs of color-coded insulated stranded copper wires that are twisted together in pairs and enclosed in a protective outer sheath. The twists in twisted-pair cabling help reduce frequency loss and improve signal transmission by reducing the effects of crosstalk. This is because twisting the wires together makes the cabling more resistant to electromagnetic interference (EMI), which helps maintain a high signal-to-noise ratio for reliable network communication to take place.
The earliest uses for twisted-pair cabling was for the Plain Old Telephone System (POTS), where the cabling was used for local loop wiring and was terminated with RJ-11 connectors. Twisted-pair cabling was developed in both shielded and unshielded configurations, with shielded cabling having better performance but costing more. Twisted-pair cabling is today the cabling medium of choice for building computer networks of all sizes from departmental local area networks (LANs) to structured wiring systems for office towers and campuses. Such twisted-pair cabling used for networking purposes employs RJ-45 connectors instead of the RJ-11 connectors used for telephony applications.
Twisted-pair cabling used in Ethernet networking is usually unshielded twisted-pair (UTP) cabling, but shielded twisted-pair (STP) cabling is typically used in Token Ring networks. UTP cabling comes in different grades for different purposes, the most common of which is Category 5 (Cat5) cabling.
Notes
In a telephone environment, one pair of wires is sufficient for ordinary telephone communication to take place. Most customer premises wiring established by telcos uses two-pair wiring in case a second line is later needed for fax or modem use.
See Also Category 5 (Cat5) cabling , crosstalk ,electromagnetic interference (EMI) ,Ethernet ,Plain Old Telephone Service (POTS) ,RJ connectors ,shielded twisted-pair (STP) cabling , unshielded twisted-pair (UTP) cabling
A trust relationship between two domains in Microsoft Windows 2000 and Windows .NET Server.
Overview
By default, all Windows 2000 and Windows .NET Server trusts are two-way, meaning that each domain trusts the authority of the other domain for authentication. A Windows 2000 trust is also transitive-if domain A trusts domain B and domain B trusts domain C, domain A trusts domain C. Windows 2000 two-way transitive trusts are based on the Kerberos v5 security protocol.
Because of the two-way transitive nature of Windows 2000 and Windows .NET Server trusts, all domains in a domain tree implicitly trust one another. This means that resources of one domain are available to users in all other domains in the domain tree if they have suitable permissions.
Notes
You can also create one-way nontransitive trusts for Windows 2000- and Windows .NET Server-based networks. These one-way trusts are similar to the trust relationships formed by Windows NT domain controllers. A one-way trust between a domain and a domain tree provides users of the domain with access only to the domain in the tree to which it is joined. One-way trusts can be useful when domains require a less permanent relationship-for example, when two companies take part in a joint venture. Only the resources needed by the other company are made available to the trusted domain; the entire domain tree is not exposed.
See Also Active Directory , domain (DNS) ,Kerberos ,