5-4-3 rule. This rule restricts bus-style Ethernet networks to five total segments, four repeaters, and three populated segments.
An obsolete 1 megabit per second (Mbps) local area network (LAN) networking technology.
Overview
1Base5 was actually developed after the 10Base5 Standard Ethernet specification was ratified but before the ratification of the 10BaseT standard. It was developed as an initial attempt to deploy computer networking over existing unshielded twisted-pair (UTP) cabling at a time when Ethernet was still based on coaxial cabling. The popular name for 1Base5 networking technologies was StarLAN because it was wired in a hierarchical star topology. When 10BaseT was ratified as an Institute of Electrical and Electronics Engineers (IEEE) standard in 1990, however, StarLAN and similar technologies quickly became obsolete and fell into disuse.
See Also 10BaseT ,Ethernet
Basic Rate Integrated Services Digital Network (ISDN) service, which uses two B channels and one D channel for signaling and communications.
See Also Integrated Services Digital Network (ISDN)
Stands for Two Binary One Quaternary, a physical layer encoding mechanism for translating digital information into electrical signals, standardized by the American National Standards Institute (ANSI) and used in Integrated Services Digital Network (ISDN) networks. In 2B1Q, four electrical amplitude and polarity values are used to represent binary information.
See Also line coding
Stands for second generation and refers to widely deployed cellular communications systems such as Digital Advanced Mobile Phone Service (D-AMPS), Time Division Multiple Access (TDMA), Code Division Multiple Access (CDMA), Global System for Mobile Communications (GSM), and Personal Communication Services (PCS) systems.
See Also 3G
A label some vendors use to refer to several interim cellular communications systems currently being deployed as a prelude to full third-generation (3G) systems.
Overview
For many cellular systems, upgrading immediately from existing second-generation (2G) systems to full 3G functionality is costly and complex. Furthermore, some vendors question whether there is a need for the high (2 megabits per second) data rates promised by 3G and suggest that lower data rates in the hundreds of kilobits per second (Kbps) are more than sufficient for the majority of current applications. They cannot justify the enormous cost of an upgrade considering the insufficient demand for the type of interactive multimedia communications that 3G will support. Even 2.5G would be a significant leap forward from today's current 2G systems, where typical data rates are a mere 9.6, 14.4, or 19.2 Kbps.
As a result, some cellular service providers have opted to migrate from 2G to 3G in stages by first upgrading their existing Time Division Multiple Access (TDMA) and Code Division Multiple Access (CDMA) networks to interim technologies which offer much higher data transmission rates. These interim 2.5G systems include Enhanced Data Rates for Global Evolution (EDGE), developed by Ericsson, and High Data Rate (HDR), developed by Qualcomm.
For example, some TDMA networks that can carry data at only 19.2 Kbps are being upgraded first to General Packet Radio System (GPRS) overlaid on Global System for Mobile Communications (GSM) to support data transmission at 144 Kbps, and from there to EDGE which supports transmission rates of 384 Kbps. In fact, GPRS is a natural upgrade path for GSM-based systems to support Wireless Application Protocol (WAP)-based data communications easily and inexpensively at speeds from 10s to 100s of Kbps.
Similarly, many CDMA systems carrying data at 14.4 Kbps are being upgraded to intermediate technologies such as HDR before being fully upgraded to Wideband CDMA (W-CDMA) with its 2 megabits-per-second (Mbps) data rate.
In addition to the increased data rates of 2.5G systems compared to 2G (about an order of magnitude faster), another advantage is that 2.5G systems utilize packet-switching technologies that are more efficient in the transmission of IP data over cellular systems as opposed to the circuit-switching technologies in use by existing 2G systems.
Notes
Most 2.5G systems are asymmetric in their transmission speeds, supporting faster downloads than uploads. In this way they bear some similarity to cable modems and Asymmetric Digital Subscriber Line (ADSL).
Some vendors market their 2.5G systems as 3G, but this is a misnomer because they fall short of complying with the International Telecommunication Union (ITU)'s IMT-2000 standard that defines 3G technologies and systems.
See Also 3G
Stands for third generation, an umbrella name for the latest generation of cellular phone networks capable of high-speed wireless data transmission rates of up to 2 megabits per second (Mbps). 3G technologies are occasionally referred to by the name broadband wireless.
Overview
The history of wireless cellular communications has had three distinct generations since its beginnings in the early 1980s. These generations are typically identified as follows:
1G (first generation of cellular communications): This involved analog transmission of voice only (did not support data) over circuit-switched networks. 1G is represented by legacy systems such as the Advanced Mobile Phone Service (AMPS) introduced by AT&T in the early 1980s and still in use in some parts of North and South America. Other 1G systems included the Extended Total Access Communications System (E-TACS) used in Europe and Asia, the Nordic Mobile Telephone system used in Scandinavia, and the Total Access Communication System (TACS) used in Great Britain, Hong Kong, and Japan. 1G systems transmitted their traffic in clear (unencrypted) form.
2G (second generation): This saw the development of digital cellular systems for transmitting digitally encoded voice and data (though data transmission speeds are limited to about 19.2 kilobits per second [Kbps]). Examples of 2G systems include Digital AMPS (D-AMPS), Time Division Multiple Access (TDMA), Code Division Multiple Access (CDMA), Global System for Mobile Communications (GSM), and Personal Communication Services (PCS) systems. 2G systems are circuit switched and support a variety of extended features such as encryption, Short Message Service (SMS), and fax transmission.
3G (third generation): This was first envisioned by the International Telecommunication Union (ITU) in 1992 and had two aims. The first was to be a truly global digital packet-switched cellular system (as opposed to the circuit-switched 2G systems) for voice and data communications, where users could roam between continents without requiring new phones. The second was to provide consistent high-speed data transmission rates of up to 2 Mbps. Unfortunately, neither of these goals has been completely achieved.
In 1992 the ITU proposed a system called IMT-2000 as the single global 3G- based standard for cellular communications. In 1999 conflicts between carriers, regulators, and vendors in different countries and regions led to a compromise version of the IMT-2000 standard. The agreed-upon standard incorporated a mix of three different proposed versions, two of which use CDMA technologies and the third using TDMA. Existing CDMA providers will find it easier to upgrade to 3G technologies than TDMA providers will.
Types
Qualcomm has proposed a 3G upgrade for its popular 2G system cdmaOne, used throughout much of Asia and North America. The upgrade is called cdma2000 3XMC. It has recently been ratified by the ITU and is expected to provide real data rates of 1.117 Mbps. Other CDMA providers have proposed that an initial upgrade called cdma2000 1XMC be deployed first as an interim "2.5G" cellular system before full deployment of cdma2000 3XMC is achieved. The cdma2000 1XMC technology only supports real data rates of 144 Kbps, although vendor-proprietary extensions have been proposed to extend this.
Another 3G CDMA system similar to cdma2000 3XMC and currently being rolled out in some countries and regions is called Universal Mobile Telecommunications System (UMTS) or Wideband CDMA (W-CDMA) (carriers in Europe prefer the UMTS designation, but those in North America prefer W-CDMA or WCDMA). UMTS is designed to interoperate with existing GSM networks and will provide a natural upgrade path for these networks from 2G to 3G. The first country widely rolling out W-CDMA is Japan, which is using it to replace its Personal Digital Cellular (PDC) system, whose capacity is saturated. Unfortunately, UMTS/W-CDMA cannot be provisioned in the United States because the frequencies this technology requires have already been allocated by the Federal Communications Commission (FCC) for other uses. UTMS/W-CDMA supports wireless data transmission at 2 Mbps. Note, however, that this speed will be achievable for stationary users only-for mobile environments this slows to 384 Kbps and for fast- moving vehicles such as a car on the highway, it is down to 144 Kbps.
Another proposed 3G system based on TDMA instead of CDMA is called EDGE (Enhanced Data Rates for Global Evolution). EDGE is an upgrade to existing General Packet Radio Services (GPRS) systems and supports data rates of 384 Kbps.
Prospects
3G was originally envisioned to become widespread by 2000 (hence the name of the original proposed IMT-2000 system), but these systems are only beginning to be deployed (now pushed back to between 2003 and 2005, depending on the country or region). They provide data rates less than those anticipated, and have developed interoperability problems that may not be easily resolved.
3G technology is also expensive to develop and deploy. In this regard, Europe has a head start on the United States; however, the total cost of deploying 3G throughout the European Union (EU) has been estimated to be as high as $1 trillion. The long-term viability of many European 3G carriers may be affected by the high licensing fees imposed by governments for use of suitable radio spectrum. Most industry watchers expect a shakeout in the next few years, leaving only giants such as Deutsche Telekom, France Telecom, Vodafone, and a few others.
Many countries and regions, including the United Kingdom, Belgium, Germany, and Sweden, hold auctions to parcel off portions of 3G spectrum to the highest bidders. Carriers must quickly recoup these up-front costs to remain viable. Some countries and regions, such as Norway, Finland, France, and Spain, allocate spectrum on a fixed-fee basis to first-comers, but this practice is being challenged in the EU courts as a type of subsidy that is unfair and anticompetitive.
In the United States, AT&T laid out a road map in December 2000 for its deployment of 3G wireless that involves several intermediate steps. First, its existing TDMA network, which carries data at only 19.2 Kbps, will be upgraded to GSM with an overlay of GPRS to support data transmission at 144 Kbps. Then GSM/GPRS will be upgraded to EDGE to support data transmission rates of 384 Kbps. Finally, W-CDMA will be deployed to support data at 2 Mbps. Meanwhile, CDMA carriers such as Sprint Corporation and Verizon Communications are upgrading their systems to cdma2000 and beyond. This split between W-CDMA, supported by AT&T, and cdma2000, supported by Verizon and Sprint, presents some interoperability problems for users in North America, but efforts are underway to resolve these issues.
Ultimately, when 3G systems are fully operational, they will provide mobile users with the equivalent of broadband Internet access, opening up a new world of wireless Web-based services and applications.
Notes
Although 3G wireless systems such as UMTS will be capable of supporting data transmission speeds of 2 Mbps, this will be for stationary clients only, which includes laptop computers equipped such as 3G PC cards. Transmission speeds for mobile clients such as 3G cell phones and handsets will likely top out at 384 Kbps or lower, depending on the environment's reception characteristics.
Because of the high costs and technical challenges associated with rolling out 3G systems, some cellular communications providers are rolling out unofficial "2.5G" systems as a kind of intermediate step along the way to the high-speed data rates of 3G. For more information, see the article "2.5 G" elsewhere in this chapter.
For More Information
An alliance of regulators and vendors committed to remaining neutral in the 3G standards war is the Third Generation Partnership Project, whose site can be found at www.3gpp.org.
Details of the ITU's IMT-2000 standard can be found at www.itu.int/imt2000.
See Also 2.5G ,cellular communications
A service provided by some telcos for transmitting data over voice trunk lines.
Overview
When supporting 3.1 kHz bearer service communications, telco switches must have trunk-line echo cancellation turned off because echo cancellation will corrupt data transmissions sent over voice lines. This service is a legacy telecommunications technology that is no longer widely implemented.
Notes
Some carriers call this service data-over-voice (DOV).
An envisioned fourth-generation wireless cellular system to supersede 3G (third-generation) cellular.
Overview
4G is a term sometimes applied to proposed wireless networking systems using Orthogonal Frequency Division Multiplexing (OFDM). OFDM is a modulation scheme theoretically capable of transmitting data at rates of 622 megabits per second (Mbps), which is much higher than the 2 Mbps of the 3G wireless systems that are only now beginning to be deployed. OFDM is currently defined in two implementations, one supported by the Institute of Electrical and Electronics Engineers (IEEE) and the other by the European Telecommunications Standards Institute-Broadband Radio Access Networks (ETSI-BRAN). These two implementations are different in some respects, and efforts are underway by the OFDM Forum to harmonize these systems to reduce the incompatibilities that could result once 4G systems begin to materialize. Apart from these difficulties, 4G wireless networking systems are more of a dream than a reality.
See Also 3G
A specification describing limitations on constructing certain kinds of Ethernet networks. The 5-4-3 rule applies specifically to Ethernet networks based on either the thinnet or thicknet cabling option.
5-4-3 rule. This rule restricts bus-style Ethernet networks to five total segments, four repeaters, and three populated segments.
Overview
According to the Ethernet specifications, thinnet (or thicknet) Ethernet network segments can be joined using repeaters to form larger networks, but there are limitations on how you can do this. The maximum number of segments you can join is five. To join these segments, you need to use four repeaters because Ethernet typically uses a bus topology in which all segments are joined linearly. However, in this configuration, no more than three of the segments can actually have computers attached to them, leaving two segments that are used only for extending distances rather than hosting computers. These two unpopulated segments are called inter-repeater links. You should not violate this rule when implementing Ethernet networks; otherwise, unreliable network communications might result.
See Also Ethernet
A testbed for development of the next generation of the Internet Protocol (IP) protocol, IPv6.
Overview
The 6bone grew out of the IP Next Generation (IPng) project of the Internet Engineering Task Force (IETF) that led to the development of IPv6. It is an informal collaborative project among a number of groups around the world, operating under the oversight of the IETF's Next Generation Transition (NGtrans) Working Group. The 6bone began operation as a virtual network in which IPv6 packets were encapsulated in and tunneled over IPv4, but migration to native IPv6 is underway and should soon be completed.
The 6bone has been playing an important role in testing IPv6 before widespread migration from the current IPv4 (standard IP) networking protocol is recommended for public use. Projects such as FreeNet6, developed by Viag nie, and vendors such as Nortel Networks and 3Com Corporation have provided IPv6 researchers with gateways for tunneling from their IPv6 testbed deployments through the IPv4 Internet and over the 6bone to test and become familiar with the new IPv6 protocol.
For More Information
The official site for the 6bone can be found at www.6bone.net You can also find information there about how to join your network to the 6bone.
If you want to connect an individual workstation to the 6bone, you can use Freenet6, which can be found at www.freenet6.net. This site enables IPv6 end stations to connect to the 6bone through IPv4 tunneling over the Internet. You must have a dual IPv4/6 stack to connect. You can download a preliminary IPv6 stack from Microsoft Research at www.research.Microsoft.com/msripv6.
See Also IPv6
A mechanism for connecting networks running IPv6 by tunneling through the IPv4 Internet.
Overview
6to4 was developed as a transitional scheme to support the operation of IPv6 within an IPv4 world until the backbone of the Internet is completely migrated to IPv6. 6to4 works by encapsulating IPv6 packets inside IPv4 packets and transmitting them over the existing IPv4 Internet. In this way, different islands of IPv6 connectivity can be linked across the IPv4 ocean that constitutes the Internet.
A special IPv6 prefix is used to identify an encapsulated 6to4 packet. This header is represented in IPv6 address notation as 2002::/16 and identifies the IPv6 address space specially reserved by IANA for 6to4 tunneling.
For More Information
In order to use 6to4, you must install a dual IPv4/6 protocol stack on your machine. Microsoft is integrating basic support for IPv6 into its Windows XP and Windows .NET Server operating system platforms. You can also download a pre-liminary IPv6 stack from Microsoft Research at www.research.microsoft.com/msripv6.
See Also IPv6
A format for tape backup developed by Exabyte Corporation.
Overview
8mm is the standard tape format for videotape, and Exabyte developed this format into a tape backup solution to take advantage of this medium's low cost. Typical 8mm tape drives can store up to 40 gigabytes (GB) of data on a tape, and data can be written to tape at speeds in excess of 6 megabytes per second (MBps) using compression or 3 MBps without compression. Exabyte 8mm tape drives are inexpensive, have high performance, and make an attractive tape backup solution for companies developing a disaster recovery plan.
For More Information
See Exabyte online at www.exabyte.com
See Also tape format
An Institute of Electrical and Electronic Engineers (IEEE) standard for implementing 10 megabits per second (Mbps) Ethernet over thin coaxial cabling.
Overview
10Base2 is based on the 802.3 specifications of Project 802 developed by the IEEE. It was ratified as an IEEE standard in 1985 and quickly found its way into corporate networks for small local area networks (LANs) connected to larger 10Base5 backbones.
10Base2 is sometimes referred to as thinnet or thin coax because it uses thin coaxial cabling for connecting stations to form a network (as compared to 10Base5, which uses a thicker form of cabling and hence is called thicknet). The designation 10Base2 is derived from the network's speed (10 Mbps), the signal transmission method (baseband transmission), and the maximum segment length (185 meters, rounded off to 200 with the zeros removed). Another popular nickname for this technology was Cheapernet because thinnet cabling was considerably less costly than thicknet cabling.
10Base2. A typical 10Base2 network.
Implementation
10Base2 networks are wired together using a bus topology, in which individual stations (computers) are connected directly to one long cable. The maximum length of any particular segment of a 10Base2 network is 607 feet (185 meters). If distances longer than this are required, two or more segments must be connected using repeaters. Altogether, a total of five segments can be connected using four repeaters, as long as only three of the segments have stations (devices) attached to them. This is referred to as the 5-4-3 rule.
A 10Base2 segment should have no more than 30 stations wired to it. The minimum distance between these stations is 1.6 feet (0.5 meters). Stations are attached to the cable using BNC (British Naval Connector or Bayonet-Neill-Concelman) connectors, and the ends of the cabling have BNC cable connectors soldered or crimped to them.
10Base2 supports a maximum theoretical bandwidth of 10 Mbps , but in actuality the presence of collisions reduces this to more like 4 to 6 Mbps.
Notes
10Base2 networks are not deployed much anymore for two reasons. First, because their speed is limited to 10 Mbps, the networks perform poorly in today's bandwidth-hungry, Internet-connected world. Second, 10Base2 networks have a single point of failure-the long, linear bus cable used to connect the stations. A single break or loose connection brings down the entire network, thus every cable segment and station connection must be checked to determine the problem. If you are wiring an office for a small LAN with low bandwidth requirements, use 10BaseT instead, which is easier to manage and troubleshoot. For moderate to high bandwidth requirements, try using Fast Ethernet instead.
The two ends of a 10Base2 bus must be properly terminated. If they are not, signals will bounce and network communication will come to a halt.
See Also 5-4-3 rule ,10Base5 ,10BaseF Ethernet
An Institute of Electrical and Electronic Engineers (IEEE) standard for implementing 10 megabits per second (Mbps) Ethernet over coaxial cabling.
Overview
10Base5 is based on the 802.3 specifications of Project 802 developed by the IEEE. It was developed as a standard in the early 1980s and became hugely popular in corporate and campus networks. An earlier form of 10Mbps Ethernet developed by the DIX Consortium was superseded by 10Base5 when the IEEE 802.3 standard was created in 1983.
10Base5 is sometimes referred to as thicknet because it uses thick coaxial cabling for connecting stations to form a network (compared to 10Base2, which uses a thinner form of cable and is hence called thinnet). Another name for 10Base5 is Standard Ethernet because it was the first type of Ethernet to be implemented (it is also sometimes referred to as Original Ethernet, for obvious reasons). The designation 10Base5 is derived from the network's speed (10 Mbps), the signal transmission method (baseband transmission), and the maximum segment length (500 meters).
Implementation
10Base5 networks are wired together in a bus topology-that is, in a linear fashion using one long cable. The maximum length of any particular segment of a 10Base5 network is 1640 feet (500 meters). If distances longer than this are required, two or more segments must be connected using repeaters. Altogether, there can be a total of five segments connected using four repeaters, as long as only three of the segments have stations (computers) attached to them. This is referred to as the 5-4-3 rule.
A 10Base5 segment should have no more than 100 stations wired to it. These stations are not connected directly to the cable as in 10Base2 networks. Instead, a transceiver is attached to the cable, usually using a cable-piercing connector called a vampire tap. From the transceiver, a drop cable is attached, which then connects to the network interface card (NIC) in the computer. The minimum distance between transceivers attached to the cable is 8 feet (2.5 meters), and the maximum length for a drop cable is 164 feet (50 meters). Thicknet cable ends can have N-series connectors soldered or crimped on them for connecting segments together.
10Base5 was often used for backbones for large networks. In a typical configuration, transceivers on the thicknet backbone would attach to repeaters, which would join smaller thinnet segments to the thicknet backbone. In this way a combination of 10Base5 and 10Base2 standards could support sufficient numbers of stations for the needs of a moderately large company.
10Base5 supports a maximum bandwidth of 10 Mbps, but in actual networks, the presence of collisions reduces this to more like 4 to 6 Mbps.
Notes
10Base5 networks are legacy networks that are no longer being deployed, although some companies might choose to maintain existing ones for cost reasons. The complexity and bandwidth limitations of 10Base5 networks render them largely obsolete. If you are wiring an office for a small local area network (LAN) with low bandwidth requirements, use 10BaseT instead. For moderate to high bandwidth requirements, try using Fast Ethernet. If you are implementing a backbone for today's high-speed enterprise networks, Gigabit Ethernet (GbE) is now the preferred technology.
The two ends of a 10Base5 bus must be properly terminated. If they are not, signals will bounce and network communication will come to a halt.
See Also 5-4-3 rule ,10Base2 ,10BaseF ,10BaseT Ethernet, Fast Ethernet
An Institute of Electrical and Electronic Engineers (IEEE) standard for implementing 10 megabits per second (Mbps) Ethernet over fiber-optic cabling.
Overview
10BaseF is based on the 802.3 specifications of Project 802 developed by the IEEE and differs from other forms of 10-Mbps Ethernet by using fiber-optic cabling instead of copper unshielded twisted-pair (UTP) cabling. The designation 10BaseF is derived from the network's speed (10 Mbps), the signal transmission method (baseband transmission), and the physical media used (fiber-optic cabling).
The 10BaseF standard actually consists of three separate standards describing different media specifications:
10BaseFB: Defines how the synchronous data transmission occurs over the fiber-optic cabling. Using 10BaseFB segments, you can cascade or link synchronous fiber-optic hubs in configurations that are longer than traditional 10BaseT Ethernet networks and contain up to 1024 stations. This standard is more expensive and is not as widely implemented as 10BaseFL.
10BaseFL: Defines the characteristics of the fiber-optic link between the nodes and the hub or concentrator. 10BaseFL replaces the older standard for fiber-optic link segments, Fiber-Optic Inter-Repeater Link (FOIRL) segments, which was developed in the 1980s. 10BaseFL is the most commonly implemented version of 10BaseF.
10BaseFP: Defines the implementation of a star topology that does not use repeaters. 10BaseFP stands for Fiber Passive, and its segments can be only 500 meters (1640 feet) in length with a maximum of 33 stations connected. This standard is rarely used today.
Implementation
10BaseF is similar to 10BaseT in that each station is wired into a hub in a star topology to form the network. The maximum length of any segment of 10BaseF fiber-optic cabling is 6600 feet (2000 meters), compared to the 328 feet (100 meters) supported by 10BaseT, making 10BaseF suitable for long-haul interconnects.
The recommended cabling type for 10BaseF networks is 62.5-micron diameter fiber-optic cabling. This cable can be terminated with either ST connectors or SMA connectors, depending on the vendor and the hub configuration. Two-strand multimode fiber-optic cabling is used, with one strand allotted for transmitting data and the other for receiving data.
Marketplace
Nowadays, 10BaseF is only supported in legacy 10 Mbps Ethernet equipment and has been superseded by 100 Mbps Fast Ethernet in most circumstances for network backbones and interconnects.
Notes
10BaseF is preferable to 10BaseT in environments that are electrically noisy, such as in industrial areas, near elevator shafts, or around other motors or generators.
Fiber-optic cabling is often used for running cables between buildings. Differences in ground potential between the ends of copper cabling can induce voltages that can damage networking equipment if the ends are not grounded properly. Fiber-optic cabling also supports faster speeds than copper UTP cabling and provides a more suitable upgrade option to Fast Ethernet and beyond.
The maximum signal loss or attenuation on a given segment should be no more than 12.5 decibels. Using too many connectors in a segment of fiber-optic cabling can cause the attenuation to exceed this figure, which can lead to signal loss.
See Also 10Base2 ,10Base5 ,10BaseT Ethernet
An Institute of Electrical and Electronic Engineers (IEEE) standard for implementing 10 megabits per second (Mbps) Ethernet over twisted-pair cabling.
Overview
10BaseT is based on the 802.3 specifications of Project 802 developed by the IEEE and is the most popular form of 10-Mbps Ethernet. 10BaseT is deployed over structured cabling systems consisting of unshielded twisted-pair (UTP) cabling used for connecting end stations to centralized hubs to form a network. (Shielded twisted-pair [STP] cabling can also be used, but it never is.) The designation 10BaseT comes from the network's speed (10 Mbps), the signal transmission method (baseband transmission), and the physical medium used for transmission (twisted-pair cabling).
10BaseT became widely popular because of the earlier success of the Public Switched Telephone Network (PSTN), a hierarchical structured-wiring system to which 10BaseT bears many similarities. An advantage of 10BaseT over earlier 10 Mbps Ethernet systems such as 10Base5 and 10Base2 is that it is easier to manage because of the centralization of network traffic in hubs.
Implementation
In10BaseT networks, end stations such as workstations and servers are wired together in a star topology to a central hub. The UTP cabling used for wiring should be Category 3 (Cat3) cabling, Category 4 (Cat4) cabling, or Category 5 (Cat5) cabling, terminated with RJ-45 connectors. Patch panels can be used to organize wiring and provide termination points for cables running to wall plates in work areas. Patch cables then connect each port on the patch panel to the hub. Usually most of the wiring is hidden in a wiring cabinet and arranged on a rack for easy access.
The maximum length of any particular segment of a 10BaseT network is 328 feet (100 meters). In practice this is not a limitation because a survey by AT&T indicated that about 99 percent of desktops in commercial buildings are located within 328 feet (100 meters) of a wiring closet. If distances longer than that are required, two or more segments may be connected using repeaters. The minimum length of any given segment is restricted to 8 feet (2.5 meters).
By using stackable hubs or by cascading regular hubs into a cascaded star topology, you can network large numbers of computers using 10BaseT cable. Although 10BaseT can support up to 1024 nodes, networks with no more than 200 or 300 nodes will yield the best performance by keeping collision domains small. Hubs can be hierarchically arranged to a depth of up to three levels in order to accommodate much larger networks, but performance declines significantly as the number of stations exceeds several hundred.
Notes
Although 10BaseT theoretically supports a maximum bandwidth of 10 Mbps, in actual networks the presence of collisions reduces throughput to about 4 to 6 Mbps.
The maximum length of a 10BaseT cable segment is not a result of the specifications for round-trip communications on an Ethernet network but rather a limitation caused by the relatively low signal strength of 10BaseT systems. With enhanced Category 5 (Cat5e) cabling, you might be able to sustain network communications effectively with cable lengths up to about 490 feet (150 meters), although this is not normally recommended.
10BaseT. A typical 10BaseT network.
When wiring a new 10BaseT network, always use Cat5e cabling. This will make it unnecessary to rewire your network should you decide to upgrade later to Fast Ethernet and beyond.
See Also 10Base2 ,10Base5 ,10BaseF Ethernet
An abbreviation for 10G Ethernet.
See Also 10G Ethernet
Also known as 10GbE, an emerging form of networking technology based on Ethernet and operating at a speed of 10 gigabits per second (Gbps). 10GbE, which stands for 10 Gigabit Ethernet, is the successor to Gigabit Ethernet (GbE) and will be standardized under the Institute of Electrical and Electronic Engineers (IEEE) 802.3ae working group.
Overview
10GbE is a switched-based technology that has similarities to and differences from earlier versions of Ethernet. It abandons the Carrier-Sense Multiple-Access with Collision Detection (CSMA/CD) Media Access Control (MAC) method of earlier versions, using instead separate send and receive channels. In other words, 10GbE operates only in full-duplex communications mode, similar to GbE and Full-Duplex Fast Ethernet. This does not eliminate contention, however, because 10GbE is designed to interoperate with slower forms of Ethernet where collisions may occur.
10GbE uses the same 48- to 1518-byte frame format as standard 10 megabits per second (Mbps) Ethernet and 100 Mbps Fast Ethernet, abandoning the jumbo frames supported by GbE. And unlike earlier versions of Ethernet, 10GbE operates exclusively over fiber-optic cabling, although efforts are being made to develop a form of 10GbE over copper cabling that can be used over the short distances needed within wiring closets. The maximum range for 10GbE will be 328 to 984 feet (100 to 300 meters) over multimode fiber and 25 miles (40 kilometers) over single-mode fiber. Laser wavelengths of 1310 and 1550 nanometers are used for medium- and long-haul transmission respectively.
Types
The driving force behind the development of 10GbE is twofold:
Enterprise-level local area networks (LANs) that have deployed core switches with GbE interconnects (and even multiplexed GbE interconnects) are approaching saturation and require the higher backbone speed promised by 10GbE. Many server farms and clusters are being deployed today using 1000BaseT connections to network backbones. As a result, these backbones will soon face saturation unless a technology such as 10GbE can provide the necessary bandwidth to support these GbE farms (the Link Aggregation Control Protocol of 802.3ad only supports up to 8 Gbps aggregated links between switches). The impending arrival of 64-bit PC architectures such as Intel's Itanium also promises to increase the rate at which servers will be able to process data, and new servers are appearing that are able to just saturate a 1 Gbps 1000BaseT NIC. The promise here is that 10GbE will enable network architects to build Ethernet wide area networks (WANs) as easily as they build Ethernet LANs today, without having to learn more complex technologies such as Asynchronous Transfer Mode (ATM) or Synchronous Optical Network (SONET).
WAN service providers (carriers) see 10GbE as a simpler, less costly replacement for SONET/Synchronous Digital Hierarchy (SDH), and envisage a world of end-to-end Ethernet incorporating both LAN and WAN technologies. Although Ethernet gear only costs about a fifth of what SONET gear costs, incumbent carriers have invested heavily in SONET technologies and ATM, which may affect how quickly they adopt 10GbE. Startup carriers are therefore more likely to jump on the 10GbE bandwagon than existing carriers.
As a result of this twofold need of the carriers (for SONET-compatibility) and enterprise networks (for easy upgrading of core switches from GbE to 10GbE), two different versions of 10GbE are envisioned under the emerging IEEE 802.3ae standard. These versions, listed below, operate differently at the physical layer (PHY):
LAN version: This is essentially just speeded-up GbE and will operate at exactly 10.0 Gbps and use standard Ethernet frames. The primary difference with GbE is that only full-duplex communications will be supported. This version is designed to be an easy upgrade for backbone switches in enterprise LANs.
WAN version: This provides a method of transporting Ethernet frames over SONET/SDH at a speed of 9.58464 Gbps (identical to the current OC-192). This version is designed to maximize compatibility with carrier's existing SONET deployments in their long-haul lines.
10G Ethernet. Some scenarios where 10GbE could be deployed.
Implementation
Some envisioned uses for 10GbE include
High-speed interconnects between GbE backbone switches in enterprise LANs
Aggregation of multiple short-haul GbE links and high-speed campus-wide interconnects
Switched connections to server farms and high- performance clusters
Some telecom carriers plan to replace their SONET/ATM rings with 10GbE in their Metropolitan Area Networks (MANs) to provide high-speed connections in the local loop between Central Offices (COs) and customer premises. These carriers will then be able to provision corporate users with 10GbE Ethernet WAN links to provide end-to-end Ethernet services across the WAN. Because this eliminates the need for costly packet-conversion equipment such as frame relay assembler/disassemblers (FRADs) or ATM access routers at the customer premises, 10GbE WAN services are likely to be much less costly than current T1 and frame relay services. Furthermore, 10GbE has almost no protocol overhead compared to ATM/SONET, which can have as much as 56 percent protocol overhead when implementing Automatic Protection Switching (APS) routines. The desire of carriers to implement 10GbE to save cost is a major force driving vendors to quickly develop 10GbE switching technologies. Due to its distance limitations, however, 10GbE is not expected to replace SONET/SDH for long-haul telecommunications links between cities or continents.
Issues
Most servers are not yet capable of utilizing the full potential of 10GbE because of limitations in processing power and bus speed, though advances in bus, memory, and storage technologies are closing the gap. With standard 1500-byte Ethernet frames being employed, servers would need to be able to process 833,000 frames per second, which is beyond their current limitations. One solution would be to employ Jumbo Frames similar to those used in GbE, but the 802.3ae standard does not cover this feature. The result is that some vendors are starting to implement their own proprietary Jumbo Frame technologies, which raises concerns for future prospects of interoperability between 10GbE equipment from different vendors.
Prospects
Many users are looking forward to the day when carriers can provision their corporate networks with 10GbE. Carriers will likely supply these services using a bandwidth-on-demand model, allowing customers to replace their costly T1 and T3 lines with blocks of purchased bandwidth that can scale in real time to accommodate bursts of traffic. Application Service Providers (ASPs) are also waiting to take advantage of 10GbE to offer outsourced email and workflow applications that will perform as well as LAN-based implementations.
The IEEE published a draft standard (802.3ae) for 10GbE in September 2000, and the final standard is expected to be ratified in 2002. Until then, 10GbE equipment provided by vendors may involve proprietary technologies, and network architects should consider interoperability issues before committing budgets to this emerging technology. Another cost issue to consider is that existing copper and multimode fiber infrastructures would have to be replaced with more costly single-mode fiber if anything more than backbone switch interconnects is considered in a proposed 10GbE deployment.
Some of the vendors developing 10GbE plug-in backplane modules for their carrier-class backbone switches include Nortel Networks, Extreme Networks, and Foundry Networks. Because there is already demand in the corporate WAN market, service providers such as Yipes (www.yipes.com) are rapidly rolling out infrastructure to provide 10GbE carrier services in a number of cities across the United States and compete with local telco carriers selling T-carrier and frame relay services.
Notes
10GbE will support Simple Network Management Protocol (SNMP) and (Remote Network Monitoring) RMON at the PHY layer. This will enable carriers to remotely manage provisioned 10GbE services and should be a boon to enterprise network managers as well.
For More Information
You can find the 10 Gigabit Ethernet Alliance (10GEA) at www.10gea.org.
Also find out about the IEEE 802.3ae working group at www.ieee.org.
See Also Ethernet ,Fast Ethernet ,Gigabit Ethernet (GbE)
A term implying the uninterrupted running of network services. A 24 x 7 network is a network whose services and resources are available 24 hours a day, seven days a week, with virtually no downtime. A similar term is 24 x 7 x 365 , which implies virtually no downtime during the entire year.
Overview
In today's emerging e-business economy, availability of the network and its resources and applications can make the difference between business success and failure. A variety of technologies can be used to approach the goal of 24 x 7 availability. One of the most successful is clustering, a technology in which an application runs redundantly on multiple servers. When one of the nodes in a cluster fails, the other nodes take on the workload of the failed node so that there's no interruption in service while the failed node is being repaired or replaced. Windows 2000 and .NET Server Advanced Server editions offer clustering services for high-availability line-of-business applications.
Other technologies important to the 24 x 7 goal include fault-tolerant hardware technologies. Examples are hot-swappable components, such as power supplies and hard disks that can be removed and replaced without the system having to power down or reboot. Fault-tolerant storage technologies such as RAID 1 or RAID 5 guard against data loss from disk failure. Cellular phones and pagers also play a role in attaining the 24 x 7 goal, as they allow businesses to keep in touch with administrators and technical support staff around the clock.
Notes
If you are an MIS (Manager of Information Services) or CIO (Chief Information Officer) running an e-business that requires 24 x 7 uptime, it is wise to make sure your network administrators and support staff are adequately compensated-they must work hard to maintain a 24 x 7 networking environment. Also, such people can be difficult to replace, so it is advisable to do what you can to minimize turnover.
See Also network management
Any computing hardware based on a processor capable of manipulating 64 bits of information at a time and directly accessing up to 264 bytes (16 exabytes) of different physical memory addresses.
Overview
The PC revolution of the 1990s was based to a large degree upon the x86 processor platform, a series of 32-bit processors developed by Intel Corporation. Compared to their 16-bit predecessors, these new processors could manipulate twice as many bits simultaneously and access up to 4 gigabytes (GB) of directly addressable memory, which provided more than enough processing power and memory space for most PC-based applications.
As the PC platform began to push its way into the high-end server platform previously dominated by Reduced Instruction Set Computing (RISC) processors, the demand for greater power and memory addressability grew. This need has been felt most in the area of enterprise-level database applications, which are processor and memory intensive in their requirements. To address these needs, two leading processor vendors, Intel and AMD, began development of 64-bit processors in the late 1990s.
Comparison
AMD's x86-64 architecture can run today's 32-bit operating systems and applications, but only when running in Legacy mode. Unfortunately, this mode limits addressable memory to 4 GB and provides no processing gain over the present Pentium III platform. In order to run true 64-bit operating systems and applications, the x86-64 architecture must be running in Long mode, which does not support backward compatibility with existing 32-bit applications and operating systems.
In order for Intel's IA-64 architecture to run legacy 32-bit software, an x86 hardware emulation mode is provided. When 32-bit software is run using emulation mode, however, performance is generally slower than that of a typical Pentium III processor.
In order to properly run legacy software on Itanium (or on Sledgehammer running in Long mode) and enable such software to address the full 64-bit address space, the software needs to be recompiled to produce a binary bit-image appropriate to the selected 64-bit platform. Unfortunately, these binary images are different for the two architectures, so either a single architecture will win out or vendors will need to compile two different versions for the two platforms. Until existing software is recompiled, or until 64-bit operating systems and applications are developed from scratch, the usefulness of these two new architectures will be limited as far as the business enterprise is concerned.
Architecture
Intel's new 64-bit Itanium chip uses a radically new architecture called IA-64 that differs in fundamental ways from the x38 architecture used in all of its earlier 32-bit processors, from the 386 through the Pentium III Xeon. Itanium uses a Very Long Instruction Word (VLIW) algorithm that allows it to read strings containing multiple instructions. Using a technique called Explicitly Parallel Instruction Computing (EPIC), Itanium can then execute up to six instructions per clock cycle, which greatly enhances performance over 32-bit processors (not just Intel x86 processors, but also 32-bit RISC and complex instruction set computing [CISC] processors) and eliminates the need for complicated algorithms such as the Out-of-Order Processing implemented in Pentium chips.
Itanium processors are capable of performing 6 billion floating-point operations per second (6 gigaflops) and support up to 512-way symmetric multiprocessing (SMP) for the most transaction-intensive e-commerce and database applications. Itanium uses a new motherboard interface and requires a bus speed of 366 MHz.
AMD's new 64-bit Sledgehammer chip uses an architecture called x86-64 that is a natural evolution from the x86 standard, and AMD's 32-bit Athlon processor. Sledgehammer operates in two modes: long mode, which runs native 64-bit operating systems and applications, and legacy mode, which supports existing 16-bit and 32-bit operating systems and applications. Sledgehammer uses a motherboard with a Plastic Pin Grid Array (PPGA) processor slot and requires a 266 MHz system bus.
Prospects
Both architectures have their camps supporting them. Intel's IA-64 architecture is supported by Microsoft Corporation, Hewlett-Packard, IBM, Compaq Computer Corporation, and others, and is likely to be the platform of choice for users of Microsoft Windows once Microsoft releases its 64-bit version of Windows currently under development. AMD's x86-64 architecture is supported by Sun Microsystems, which has plans to port Solaris (a version of UNIX) to x86-64, and by the Linux community, which is working on a similar port.
Notes
Some software vendors have developed ways of going beyond the hardware limitations of 32-bit processors that limit addressable physical memory to 4 GB. Examples include Microsoft Windows 2000 Advanced Server and Microsoft Windows 2000 Datacenter Server, both of which employ a technique called Physical Address Extension (PAE) to enable them to address 8 GB and 64 GB of memory respectively.
For More Information
More information on these two platforms can be found at www.intel.com/ebusiness/products/ia64 and www.amd.com/products/cpg/64bit.
See Also 64-bit Windows
A version of Microsoft Windows 2000, Windows XP, and Windows .NET Server designed to operate on computers using Intel's new IA-64 (64-bit) processor architecture.
Overview
64-bit Windows is designed to support today's most processor- and memory-intensive business applications. These applications include e-commerce, data mining, online transaction-processing, high-end graphics, and high-performance multimedia applications. Based on Intel's 64-bit processor architecture called IA-64, which is first realized in the Itanium chip, 64-bit Windows is fully compatible with the Win32 API, allowing developers to develop and compile code for both the 32-bit and 64-bit platforms with equal ease. The main advantage of using a 64-bit processor architecture is that the amount of directly addressable physical memory increases from 64 GB for 32-bit Windows 2000 and Windows .NET Server Datacenter Server to a staggering 16 terabytes (TB) for the new platform. (A terabyte equals 1024 gigabytes.) This means that large databases can be preloaded in-to memory to provide vastly improved data access times, improving the performance and efficiency of enterprise-level applications.
For More Information
For more information on the current state of the 64-bit Windows platform, see www.microsoft.com/hwdev/64bitwindows/ and www.microsoft.com/windowsxp/ pro/techinfo/howitworks/64bit/default.asp.
See Also 64-bit architecture ,Microsoft Windows
A rule of thumb that says that 80 percent of network traffic is local and only 20 percent is destined for remote networks.
Overview
The 80/20 rule was developed in the era of traditional routed Ethernet networks and is now considered a classical description of local area network (LAN) traffic patterns that no longer applies. The main consideration here is the Internet, and corporate use of Internet access means that ever-increasing amounts of remote traffic are being routed through LANs. The 80/20 rule was used in the 1980s to help enterprise network architects plan and design large-scale routed networks.
Notes
Other types of 80/20 rules exist. For example, in programming, the 80/20 rule says that 80 percent of the benefit in coding an application comes from 20 percent of the work involved. In other words, prioritize your development tasks to achieve the most in the time available. The original 80/20 rule seems to have originated with Vilfredo Pareto, an early 20th-century economist who stated that 80 percent of the volume of product produced comes from only 20 percent of the producers who make it. Another modern variant says that 80 percent of your business comes from only 20 percent of your customers, so identify these customers and give them premier service.
See Also local area network (LAN)
An Institute of Electrical and Electronics Engineers (IEEE) standard for implementing 100 megabits per second (Mbps) Ethernet over fiber-optic cabling.
Overview
100BaseFX is based on the 802.3u standard, which is an extension of the 802.3 standard of Project 802 developed by the IEEE. It's a type of Fast Ethernet that is often used for wiring campus backbones using fiber-optic cabling. The designation 100BaseFX is derived from the network's speed (100 Mbps), the signal transmission method (baseband transmission), and the physical medium used for transmission (fiber-optic cabling).
Implementation
100BaseFX networks are wired together in a star topology using fiber-optic cabling and 100-Mbps fiber-optic hubs or Ethernet switches. 100BaseFX systems may be interconnected with 100BaseTX, 100BaseT4, and 10BaseT systems using auto negotiating hubs and switches with suitable ports. Two-strand fiber-optic cabling is required, and ST, SC, and MIC connectors are all supported. Signaling is at 125 megahertz (MHz), which when combined with the 80 percent efficiency of the 4B5B line coding mechanism used results in an overall transmission speed of 100 Mbps.
The maximum length of any segment of fiber-optic cabling connecting a station (computer) to a hub in 100BaseFX is 1350 feet (412 meters), and not 1480 feet (450 meters) as some sources indicate. The grade of fiber-optic cabling used is usually two-strand multimode fiber-optic cabling, with one strand carrying transmitted data and the other strand receiving data. However, you can also use two-strand single-mode fiber-optic cabling. If multimode fiber-optic cabling is used, the variety used is typically a grade with a 62.5-micron core diameter.
Repeaters can be used to extend the length of cabling and for interfacing between 100BaseFX/TX and 100BaseT4 segments. The maximum allowable distances with repeaters are 2 kilometers using multimode fiber-optic cabling and 10 kilometers using single- mode fiber-optic cabling. Only one or two repeaters can be used per collision domain, depending on whether Class I or Class II repeaters are used.
Notes
100BaseFX and a related standard, 100BaseTX, are sometimes collectively referred to as 100BaseX.
When using 100BaseFX with repeaters for backbone cabling runs, Ethernet switches cannot be more than 1350 feet (412 meters) apart when running in half- duplex mode and 6600 feet (2000 meters) apart when running in full-duplex mode.
See Also 100BaseT4 ,100BaseTX Fast Ethernet
Another name for Fast Ethernet and thus for all 100 megabits per second (Mbps) Ethernet varieties. Also sometimes used interchangeably with 100BaseT4.
See Also Fast Ethernet
An Institute of Electrical and Electronics Engineers (IEEE) standard for implementing 100 megabits per second (Mbps) Ethernet over twisted-pair cabling.
Overview
100BaseT4 is based on 802.3u, which is an extension of the 802.3 specifications of Project 802 developed by the IEEE. 100BaseT4 is the most commonly used implementation of Fast Ethernet today. The designation 100BaseT4 is derived from the network's speed (100 Mbps), the signal transmission method (baseband transmission), and the physical medium used for transmission (all four pairs of wires in standard twisted-pair cabling). 100BaseT4 is now considered legacy technology and has been largely superseded by 100BaseTX.
100BaseT4. A typical 100BaseT4 network.
Implementation
100BaseT4 networks are wired together in a star topology using unshielded twisted-pair (UTP) cabling and 100-Mbps hubs or Ethernet switches. The UTP cabling involved may be Category 3 (Cat3), Category 4 (Cat4), or Category 5 (Cat5) cabling-with Cat5 cabling and enhanced Category 5 (Cat5e) cabling being the most commonly used solutions nowadays.
100BaseT4 uses all four pairs of wire in standard UTP cabling, for signaling with signaling rates of 25 megahertz (MHz) and an 8B6T line coding mechanism. One pair is used exclusively for transmission and a second pair for reception. The other two pairs are bidirectional and can be used either to transmit or to receive data as required. In this way, three of the four wire pairs are used at any given time to provide half-duplex transmission or reception of signals. Sharing three pairs of wires for data transfer allows 100BaseT4 to make use of lower-grade Cat3 cabling already installed in many older buildings.
The maximum length of any segment of cabling connecting a station (computer) to a hub is 328 feet (100 meters). This ensures that round-trip signaling specifications are met, because violating these specifications can produce late collisions that disrupt network communications. The Electronic Industries Alliance/Telecommunications Industry Association (EIA/TIA)- recommended length of cabling between the station and the wiring closet is only 295 feet (90 meters), allowing up to 32 feet (10 meters) more of cabling for patch cables used to connect patch panels to hubs or switches. The pinning of the RJ-45 connectors used for 100BaseT4 wiring is the same as for 10BaseT wiring.
Notes
Make sure all your cabling, connectors, and patch panels are fully Cat5-compliant. For example, ensure that when UTP cabling is connected to patch panels, wall plates, or connectors, the wires are not untwisted more than half an inch at the termination point.
100BaseT4 hubs and switches are typically available in an autosensing 10/100-Mbps variety for interoperability with older 10BaseT networks and to facilitate an easy upgrade from 10BaseT to 100BaseT.
See Also 100BaseFX ,100BaseTX Fast Ethernet
An Institute of Electrical and Electronics Engineers (IEEE) standard for implementing 100 megabits per second (Mbps) Ethernet over twisted-pair cabling.
Overview
100BaseTX is based on 802.3u, which is an extension of the 802.3 specifications of Project 802 developed by the IEEE. The designation 100BaseTX is derived from the network's speed (100 Mbps), the signal transmission method (baseband transmission), and the physical medium used for transmission (the same two pairs of wires in standard four-wire twisted-pair cabling that are used for 10BaseT Ethernet).
Implementation
100BaseTX networks are wired together in a star topology using either unshielded twisted-pair (UTP) cabling or data-grade shielded twisted-pair (STP) cabling. If UTP cabling is used (which is the most common scenario), it must be Category 5 (Cat5) cabling or enhanced Category 5 (Cat5e) cabling. Stations are connected together using hubs or switches. Unlike 10BaseT hubs, which can be hierarchically connected up to three levels deep, 100BaseTX hubs can only be connected two layers deep, which imposes additional distance limitations and may necessitate rewiring of existing cabling before upgrading from 10BaseT to 100BaseTX.
100BaseTX. A typical 100BaseTX network.
100BaseTX uses the two pairs of wires in twisted- pair cabling that are used by 10BaseT networks, with one pair of wires used for transmission and the other used for reception. With the appropriate equipment, 100BaseTX is capable of supporting both the fa- miliar half-duplex Ethernet used by 10BaseT and newer full-duplex Ethernet signaling technologies. 100BaseTX employs a signaling rate of 125 megahertz (MHz) for each pair of wires, which, because the 4B5B line coding algorithm used is only 80 percent efficient, translates into a data transmission speed of 100 Mbps.
The maximum length of any segment of cabling connecting a station to a hub is 328 feet (100 meters). This ensures that round-trip signaling specifications are met, because violating these specifications can produce late collisions that disrupt network communications. The pinning of the RJ-45 connectors used for 100BaseTX wiring is the same as for 10BaseT wiring with wires 1 and 2 used for transmission and wires 3 and 6 used for reception. This enables 100BaseTX autosensing hubs and switches to operate in mixed 10/100 Mbps Ethernet networks.
Notes
The maximum distance between 100BaseTX hubs and bridges or switches is 738 feet (225 meters), further than the maximum hub-station distance of only 328 feet (100 meters).
100BaseTX and a related standard, 100BaseFX, are sometimes collectively referred to as 100BaseX.
Although the maximum length of segments joining stations to hubs is 328 feet (100 meters), the Electronic Industries Alliance/Telecommunications Industry Alliance (EIA/TIA) recommends only 295 feet (90 meters) of cabling between the station (computer) and the wiring closet, allowing up to 32 feet (10 meters) more of cabling for patch cables used to connect patch panels to hubs or switches.
Make sure all your cabling, connectors, and patch panels are fully Cat5-compliant. Make sure that when UTP cabling is connected to patch panels, wall plates, or connectors, the wires are not untwisted more than half an inch at the termination point.
See Also 100BaseFX ,100BaseTX Fast Ethernet
Another name for 100VG-AnyLAN.
See Also 100VG-AnyLAN
An umbrella term that includes 100BaseFX and 100BaseTX Fast Ethernet technologies.
See Also 100BaseFX ,100BaseTX Fast Ethernet
Also called 100BaseVG, a 100 megabits per second (Mbps) networking technology that at one time was a serious alternative to Fast Ethernet, but which now is considered legacy technology.
Overview
100VG-AnyLAN is defined by the 802.12 standard of the Institute of Electrical and Electronics Engineers (IEEE) Project 802. It is a local area network (LAN) communications technology developed by Hewlett- Packard and AT&T around 1994. 100VG-AnyLAN is similar to Fast Ethernet in its speed (100 Mbps) and ability to be deployed in both shared-media (hub-based) and switched implementations. It differs, however, in its media access method, because it uses demand priority instead of the Carrier-Sense Multiple-Access with Collision Detection (CSMA/CD) method used by Ethernet.
Implementation
100VG-AnyLAN networks are wired together in a star topology using unshielded twisted-pair (UTP) cabling, shielded twisted-pair (STP) cabling, or fiber-optic cabling. If UTP cabling is used, it can be Category 3 (Cat3), Category 4 (Cat4), or Category 5 (Cat5) cabling-with Cat5 cabling or enhanced Category 5 (Cat5e) cabling preferred. When using UTP Cat3 or Cat4 cabling, the maximum length of a segment is 100 meters (328 feet). With UTP Cat5 cabling or STP cabling, the maximum length of a segment is 656 feet (200 meters). When using multimode fiber-optic cabling, the maximum length is 6600 feet (2000 meters).
100VG-AnyLAN uses all four pairs of wires in UTP cabling with 25 megahertz (MHz) signaling for each pair. 100VG-AnyLAN uses a special coding scheme called quartet signaling, which makes it possible to transmit data over all four pairs of wires in a UTP cable simultaneously. This means that special 100VG-AnyLan hubs are required to support demand priority media access. Otherwise, the frame format, topology, and other specifications of 100VG-AnyLAN are the same as for Ethernet.
100VG-AnyLAN. A typical 100VG-AnyLAN network.
100VG-AnyLAN employs the demand priority Media Access Control (MAC) method specified by the IEEE 802.12 standard. Demand priority means that the central hub controls which station is allowed to transmit data (only one station is allowed to transmit at any given moment). This eliminates the overhead of collisions and means that 100VG-AnyLAN networks have an efficiency of 100 percent compared to the 60-70 percent efficiency of contention-based Fast Ethernet networks. Demand priority works by having the hub poll each station successively in round-robin fashion until a station indicates it has data to transmit, whereupon control is passed temporarily to the transmitting station.
If multiple hubs are used, these are connected hierarchically with the top hub, called the root hub, being the master controller for the network. If a station is connected to a local hub other than the root hub in a large 100VG-AnyLAN network, the local hub detects the station's request to transmit data during regular polling and then forwards this request to the next highest hub in its hierarchy until it reaches the root hub, which communicates with the local hub, telling it when it is allowed to service the station's request. Thus even in a large 100VG-AnyLAN network with a hierarchy of hubs, only one station is allowed to transmit data at any given time. 100VG-AnyLAN supports five levels of hubs in a hierarchy, but performance is best if cascading is limited to only three levels.
Because demand priority is a deterministic rather than contention-based media access method, stations cannot join the network on their own and begin transmitting data. Instead, when a station is first connected, it waits to be polled by the hub and then transmits a training sequence of special frames that enables the station to determine the frame type in use and other network restrictions. This training sequence temporarily suspends polling by the hub, but takes only 5 msec to be completed and so does not seriously disrupt network communications.
Advantages and Disadvantages
100 VG-AnyLAN offers greater efficiency due to elimination of collisions, but another advantage that it has over Fast Ethernet is that it supports both 802.3 Ethernet frames and 802.5 Token Ring frames, although a single hub cannot support both at the same time (all hubs on any given 100VG-AnyLAN network must be configured to support the same frame type, either 802.3 or 802.5).
Yet another advantage 100VG-AnyLAN has over Ethernet is that it has built-in support for traffic pri- oritization. Two levels of traffic are recognized by 100VG-AnyLAN hubs: normal priority and high priority. High-priority traffic requests are serviced immediately by the local hub without the need for that hub to first communicate with the network's root hub. All network traffic is then suspended until the high-priority transmission is completed.
Marketplace
Although 100VG-AnyLAN is in some ways a superior technology to Fast Ethernet due to its greater efficiency and inherent support for traffic prioritization, it has not been widely deployed. This is mainly due to limited vendor support, which has kept prices high and made Fast Ethernet more affordable. Furthermore, Fast Ethernet is simpler and is completely compatible with widely deployed 10BaseT Ethernet networks and forms the natural upgrade path for these networks.
See Also demand priority ,Fast Ethernet
A wiring standard for twisted-pair cabling defined by the Telecommunications Industry Association (TIA).
Overview
The 568-A standard was originally intended for analog voice applications but is generally used in data networks as well. The main difference between 568-A and the competing 568-B standard is in the pin layouts for RJ45 jacks, which generally terminate unshielded twisted-pair (UTP) cabling. The pin layout for 568-A wiring is technically referred to as T568A. For a comparison of these two wiring schemes, see the next article, "568-B."
Notes
The T568A wiring scheme is preferred in Canada, although this has been changing recently.
See Also 568-B ,UTP cabling
A wiring standard for twisted-pair cabling defined by the Telecommunications Industry Association (TIA).
Overview
The 568-B standard was designed for data transmission over building wires used in computer network and telecommunications systems. It differs from the competing 568-A standard mainly in the pin layouts for the RJ45 jacks that are normally used to terminate unshielded twisted-pair (UTP) cabling. The pin layout for 568-B wiring is technically referred to as T568B, and that for 568-A cabling systems is called T568A. The drawing illustrates these differences between the two wiring systems.
568-B. Comparison of pin layouts in RJ45 connectors between 568-A and 568-B.
Notes
For 10BaseT and 100BaseT Ethernet networks, only two pairs of wires are used for data transmission, but 1000BaseT Gigabit Ethernet (GbE) uses all four pairs. As a result, running GbE over a cabling infrastructure that contains a mixture of 568-A and 568-B terminations will cause signaling problems that can be hard to troubleshoot.
See Also 568-A ,UTP cabling
The Institute of Electrical and Electronics Engineers (IEEE) working group dealing with general local area network/wide area network (LAN/WAN) networking architectures and protocols and their standards.
Overview
This group is responsible for developing standards and recommending practices for LAN architectures, LAN/metropolitan area network (MAN)/WAN internetworking, network management, and all protocol layers above the media access control (MAC) and Logical Link Control (LLC) layers. The three main standards ratified by this group include
802.1D: MAC bridging
802.1G: Remote MAC bridging
802.1Q: Virtual LANs (VLANs)
Some of the projects this group is currently involved in include
802.1p: Prioritization scheme for 802.1Q VLAN traffic
802.1s and 802.1w: Issues relating to the Spanning Tree Algorithm
802.1t: Further MAC bridging issues
802.1u and 802.1v: Issues relating to VLANs and their classification.
802.1x: Port-based network access control
See Also media access control method ,Project 802 ,virtual LAN (VLAN)
An Institute of Electrical and Electronics Engineers (IEEE) standard specifying a prioritization scheme for Ethernet networks that support the 802.1Q virtual local area network (VLAN) standard. The 802.1p standard allows a rudimentary form of quality of service/class of service (QoS/CoS) on Layer-2 switched networks.
Overview
One of Ethernet's biggest drawbacks is its lack of an inherent mechanism for prioritizing network traffic. The new 802.1p standard tries to resolve this by making use of three reserved bits in the tags appended to packets by devices that support the 802.1Q VLAN standard. Employing these three bits enables 23 =8 different priority levels to be assigned to each packet of data, allowing switches and routers to handle these different types of traffic according to preconfigured schemes. The table shows the eight different traffic classes that correspond to priorities 0 (binary 000) through 7 (binary 111) according to IEEE recommendations.
Priority | Type of Traffic |
0 | Standard best-effort LAN traffic (default priority setting) |
1 | Lowest priority, used primarily for traffic that should not be allowed to affect other forms of traffic, for example, bulk file transfers |
2 | Unallocated at present |
3 | Excellent effort-in other words, best-effort delivery for important data |
4 | Controlled load-supports admission control for mission-critical business traffic |
5 | Low-priority multimedia traffic with jitter and latency less than 100 ms |
6 | High-priority multimedia traffic with jitter and latency less than 10 ms |
7 | Highest priority for traffic that controls the network infrastructure |
Marketplace
An example of an Ethernet switch that supports the 802.1p standard is the ProCurve 8000M from Hewlett-Packard. These switches also support 802.1Q, 802.3ad, QoS, and port security.
Notes
Ethernet networks that use Layer 3 switches can implement more sophisticated CoS/QoS schemes such as Differentiated Services (DiffServ) and Integrated Services with resource Reservation Protocol (IntServ/RSVP). For mixed Layer 2/3 networks, the 802.1p standard provides standard mappings between these Layer-3 CoS/QoS mechanisms and the Layer-2 802.1p prioritization scheme.
See Also 802.1Q ,virtual LAN (VLAN)
An Institute of Electrical and Electronics Engineers (IEEE) standard defining the architecture and operation of virtual local area networks (VLANs).
Overview
The 802.1Q VLAN standard ratified in 1999 allows a tag of 4 bytes to be appended to packets by network hosts and Layer 2/3 switches in Ethernet networks. This tag identifies the VLAN to which the given packet belongs. For a more detailed explanation, see the article "virtual LAN (VLAN)" elsewhere in this book. The 4 bytes of tag information are embedded into the standard Ethernet header by dividing them between the 2-byte tag protocol identifier (TPID) field and the 2-byte tag control information (TCI) field. The TCI field consists of
User priority field: This three-bit field forms the basis of 802.1p traffic prioritization schemes in Level-2 switched networks.
Canonical format indicator: This is a one-bit field that is used to toggle VLAN formats.
VLAN Identifier (VID): This is a 12-bit field that identifies the VLAN to which the packet belongs. Up to 212 (4096) VLANs can be defined.
Marketplace
Many vendors produce Ethernet switches compatible with the 802.1Q standard, although mostly at the high end of their switching gear offerings. At the workgroup level, fewer vendors support this standard in their switches. For example, the low-end Catalyst 2900 switches from Cisco Systems do not support 802.1Q, but Hewlett-Packard's new ProCurve 10/100 stackable switches with Gigabit Ethernet (GbE) backbone transceivers do.
See Also 802.1p ,virtual LAN (VLAN)
The Institute of Electrical and Electronics Engineers (IEEE) standards for the Logical Link Control (LLC), defining its operation for all network architectures that follow the Open Systems Interconnection (OSI) model. The physical layer (PHY) and media access control (MAC) layer specifications are specified for a particular technology under the 802 committee that deals with them (for example, 802.3 deals with Ethernet and 802.5 with Token Ring). The IEEE 802.2 working group is no longer active.
See Also logical link control (LLC) layer ,Project 802
The Institute of Electrical and Electronics Engineers (IEEE) working group concerned with Ethernet in its many forms. Although this set of standards is generically referred to as Ethernet, this term is actually a trademarked version of 802.3.
Overview
The 802.3 standards cover those relating to local area networks (LANs) based on Carrier-Sense Multiple- Access with Collision Detection (CSMA/CD) as their Media Access Control (MAC) method, in other words with
10 Mbps Ethernet (original 802.3 standard)
100 Mbps Fast Ethernet (802.3u)
1 Gigabit Ethernet (GbE) (802.3z and 802.3ab)
10 Gigabit Ethernet (under development)
Some recently ratified standards of the 802.3 working group include
802.3z: Gigabit Ethernet (1998)
802.3ab: 1000Base-T (GbE over copper) (1999)
802.3ac: Tag switching in virtual local area networks (VLANs) (1998)
802.3ad: Link aggregation (2000)
Some of the projects the 802.3 working group is currently involved with include
802.3ae: 10 Gigabit Ethernet (10GbE)
802.3af: Transmission of electrical power over unshielded twisted pair (UTP) cabling to power Ethernet devices
Notes
802.3 is also used to refer to the frame format used in Ethernet frames. See the article "frame type" elsewhere in this book for more information.
See Also Ethernet ,frame type ,Project 802
The Institute of Electrical and Electronics Engineers (IEEE) standard for Gigabit Ethernet (GbE) over copper, also known as 1000BaseT or 1000BaseTX. This standard was ratified in June 1999.
See Also 10BaseT ,Gigabit Ethernet (GbE)
An Institute of Electrical and Electronics Engineers (IEEE) standard for link aggregation of Ethernet switch ports and network interface cards (NICs) to combine multiple data streams into a single large stream.
Overview
Prior to 802.3ad, many vendors of Fast Ethernet and Gigabit Ethernet (GbE) gear developed their own proprietary trunking protocols for aggregating multiple links together into a single higher speed link. Examples include Adaptec's Duralink technology and Cisco Systems' Inter-Switch Link Trunking (ISL) protocol. The 802.3ad standard was developed to promote interoperability between switches from different vendors. The need for such technology is increasingly apparent in a world of streaming media and outsourcing Line of Business (LOB) functions to Application Service Providers (ASPs).
The 802.3ad standard covers all versions of Ethernet (10, 100, and 1000 megabits per second [Mbps]) and enables multiple links of the same or different speeds to be aggregated into a single logical link. For example, two 1-gigabit-per-second (Gbps) GbE ports can be combined to make a single 2-Gbps link, or a 1-Gbps GbE port and a 100-Mbps Fast Ethernet port can be combined to create a single 1100-Mbps link. The maximum bandwidth supported by aggregated 802.3ad links is 8 Gbps-in other words, up to eight full duplex 1-Gbps GbE ports can be combined into a single fat data pipe. The protocol used to accomplish this is called the Link Aggregation Control Protocol (LACP).
Advantages and Disadvantages
The main advantage of 802.3ad is that additional bandwidth for network backbones and connections to server farms can be added incrementally as needed, thus providing greater scalability at minimum cost. The 802.3ad standard includes a mechanism for load balancing traffic across aggregated links. It also supports fault tolerance to handle the failure of any link within an aggregated set of links by automatically rerouting traffic through a different link. This fault-tolerant failover feature makes 802.3ad ideal for mission-critical point-to-point links.
Marketplace
The 802.3ad standard was ratified in July 1999, and 802.3ad-compliant products are just beginning to appear, primarily GbE backbone switches and GbE NICs for high-end server clusters.
See Also Gigabit Ethernet (GbE)
The proposed Institute of Electrical and Electronics Engineers (IEEE) standard for 10 Gigabit Ethernet (10GbE).
See Also 10G Ethernet
An emerging Institute of Electrical and Electronics Engineers (IEEE) standard that allows data-carrying unshielded twisted pair (UTP) cabling to carry not only data but also electricity for powering Ethernet networking devices.
Overview
Traditionally, in networking devices the wires carrying power to run the device and the wires carrying data to and from the device are kept separate. This is done to isolate the sensitive data-handling electronics from the high power levels required by the device. Many newer network-capable devices such as hubs, Internet Protocol (IP) telephones, and Personal Digital Assistants (PDAs) require only a few watts to operate, and 802.3af promises to simplify the operation of these devices. Instead of separate cables for carrying data and power, 802.3af devices will require only data connections and will receive their power over these connections. The advantage is that fewer cables need to be deployed and used for compliant networking devices, which promises to make Voice over IP (VoIP) telephone equipment as easy to deploy and use as traditional Plain Old Telephone Service (POTS) telephones, which communicate through and are powered by the same RJ-11 outlet found in private homes.
802.3af allows up to 10 watts of power to be carried over Category 3 (Cat3) or Category 5 (Cat5) cabling when running 10BaseT or 100BaseTX Ethernet on the network. The standard provides for the transmis- sion of both AC and DC power as required. To protect non-802.3af devices from damaging power levels that could fry their electronics, the 802.3af standard includes a handshaking protocol to enable 802.3af-compliant devices to be recognized before transmitting power to them.
Marketplace
PowerDsine, Siemens, and Lucent Technologies are vendors that have developed early versions of 802.3af-compliant switches, hubs, and IP telephony equipment. Refer to these vendors' Web sites for the latest information on these products.
Notes
Cisco Systems already has its own proprietary scheme called Inline Power for power transmission over data lines. The Cisco version supports DC power transmission over UTP cabling.
See Also Category 5 (Cat5) cabling ,Ethernet ,UTP cabling
The Institute of Electrical and Electronics Engineers (IEEE) standard for Fast Ethernet that was ratified in 1995. This standard defines the physical layer (PHY) and media access control (MAC) layer specifications for this technology.
See Also Fast Ethernet
The Institute of Electrical and Electronics Engineers (IEEE) standard for Gigabit Ethernet (GbE) over fiber, which was ratified in June 1998. This standard defines the physical layer (PHY) and media access control (MAC) layer specifications for this technology.
See Also Gigabit Ethernet (GbE)
The Institute of Electrical and Electronics Engineers (IEEE) standard for the Token Bus networking architecture. This legacy architecture is now found only in some industrial settings. The IEEE 802.4 working group is no longer active.
See Also Project 802
The Institute of Electrical and Electronics Engineers (IEEE) standard for the Token Ring networking architecture.
Notes
802.5 is also used to refer to the frame format used in Token Ring frames. See the article "frame type" elsewhere in this book for more information.
For More Information
The Institute of Electrical and Electronics Engineers (IEEE) 802.5 Web site is actually not found at the IEEE Web site (www.ieee.org). Instead, it can be found at www.8025.org
See Also frame type ,Token Ring
The Institute of Electrical and Electronics Engineers (IEEE) standard for the Metropolitan Area Networks (MANs). The IEEE 802.6 working group is no longer active.
See Also Project 802
The Institute of Electrical and Electronics Engineers (IEEE) Broadband Technical Advisory Group (TAG). The IEEE 802.7 working group is no longer active.
See Also Project 802
The Institute of Electrical and Electronics Engineers (IEEE) Technical Advisory Group (TAG) for developing standards and recommending practices for fiber-optic networking. The IEEE 802.8 working group is no longer active.
See Also Project 802
The Institute of Electrical and Electronics Engineers (IEEE) standards for isochronous local area networks (LANs) for simultaneous transmission of data and voice/video. The only developed form of this standard is the 802.9a isoEthernet specification.
Overview
Isochronous means "at a precise time," which indicates that 802.9 technologies operate using clocked signals in order to deliver data according to a rigid time schedule. In this fashion 802.9 resembles the Public Switched Telephone Network (PSTN), a circuit-switched network operating at 8 kilohertz (kHz) clocking rate and transmitting one byte every 125 microseconds to produce an overall channel (called a B channel) whose bandwidth is 64 kilobits per second (Kbps).
The isoEthernet standard developed in conjunction with National Semiconductor combines traditional 10BaseT Ethernet data transmission at 10 megabits per second (Mbps) with a separate 6.144 Mbps isochronous data channel for transporting time-sensitive traffic such as voice and video. The isochronous data channel (called the Circuit channel) consists of 96 B-channels of 64 Kbps each. IsoEthernet's total bandwidth, therefore, is 16 Mbps, but this is split between the two types of transmission: data and voice/video. There are also some additional channels (D-, M-, and P-channels) that are used for control and maintenance, but these use only about 200 Kbps of bandwidth. The 10 Mbps of 10BaseT data and 6.144 Mbps of isochronous transmission are carried on the same unshielded twisted pair (UTP) cabling system by incorporating time-division multiplexing (TDM) techniques.
Notes
The IEEE 802.9 working group is no longer active and isoEthernet is now considered a legacy technology.
See Also Project 802 ,time-division multiplexing (TDM)
The Institute of Electrical and Electronics Engineers (IEEE) working group for local area network (LAN) security standards. The IEEE 802.10 working group is no longer active.
See Also Project 802
A set of Institute of Electrical and Electronics Engineers (IEEE) standards for wireless networking designed to do for wireless networking what the 802.3 standards have done for Ethernet: provide clear guidelines for vendors to develop technologies that are standards-based, interoperable, and can operate at different speeds and frequencies and also can use different encoding mechanisms.
Overview
Three standards are currently ratified under 802.11:
802.11: The original wireless networking standard supporting data rates of up to 2 megabits per second (Mbps) and now superseded by 802.11b. This standard was approved in June 1997 and provided for both direct sequence spread spectrum (DSSS) and frequency hopping spread spectrum (FHSS) physical-layer (PHY) transmission methods. 802.11 also supported transmission over infrared, but no products were developed for this method. 802.11 allows data to be transmitted at either 1 Mbps or 2 Mbps, depending on which is the highest rate negotiable in a given configuration of client and access point (the further the distance between them, the lower the possible data rate).
802.11b: A standard for wireless networking supporting data rates of up to 11 Mbps and operating in the unlicensed 2.4 GHz Industrial, Scientific, and Medical (ISM) band. 802.11b transmission is standardized on DSSS.
802.11a: A standard supporting higher data rates of up to 54 Mbps and operating in the 5 GHz UNII (Unlicensed National Information Infrastructure) band. 802.11a transmission is similarly standardized on DSSS.
Of these standards, 802.11b is currently the most widely deployed, with support from over 20 vendors, and 802.11 is considered legacy and is being phased out. Part of the success of 802.11b is due to the successful lobbying of the Wireless Ethernet Compatibility Alliance (WECA), which has helped ensure adherence to standards and interoperability between equipment from different vendors. The 802.11a standard is newer and is seen as the next generation for wireless networking and the natural upgrade path for companies needing its higher data transmission rates.
All these standards use the same Media Access Control (MAC) method, Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA), to allow multiple users to share a single communications channel. Note that this is not the MAC method used by Ethernet (it uses CSMA/CD), so it is a misnomer to call 802.11 technology wireless Ethernet.
Notes
If you upgrade your 802.11 base stations to 802.11b, they will still support legacy 802.11 clients-but only at their maximum data rate of 2 Mbps, which in actual practice is more like 1 Mbps.
For secure wireless transmission, data needs to be encrypted. The most popular solution is Wired Equivalent Privacy (WEP), which now is included as a standard feature with most wireless network interface cards (NICs), PC cards, and access points.
For More Information
You can find information about the Wireless LAN Alliance at www.alana.com and the Wireless Ethernet Compatibility Alliance (WECA) at www.wirelessethernet.org.
See Also 802.11a ,802.11b wireless networking
An Institute of Electrical and Electronics Engineers (IEEE) wireless networking communications standard that supports data transmission rates of up to 54 megabits per second (Mbps). 802.11a is widely seen as the successor to 802.11b.
Overview
The 802.11a standard uses the same Media Access Control (MAC) layer mechanism as the more commonly deployed 802.11b, namely Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA), which allows multiple users to share a single channel at the same time. The encoding scheme used at the physical layer differs, however, because 802.11b uses the same physical-layer (PHY) encoding scheme as Ethernet, but 802.11a uses a different scheme: Coded Orthogonal Frequency Division Multiplexing (COFDM or Coded OFDM).
The frequency bands used by 802.11a also differ from 802.11b, because 802.11b uses the 2.4 GHz Industrial, Scientific, and Medical (ISM) band for its operation, and 802.11a employs portions of the 5 GHz Unlicensed National Information Infrastructure (UNII) band recently allocated by the Federal Communications Commission (FCC).
A single 802.11a communications channel consists of a 20 megahertz (MHz)-wide block of frequencies divided up into 52 subchannels each 300 kilohertz (KHz) wide, 48 of which are used for carrying data, with the remaining 4 subchannels reserved for error correction. By encoding data using 64-level quadrature amplitude modulation (64QAM), a data rate of 1.125 Mbps per subchannel can be achieved, which translates into a maximum overall transmission speed of 48 x 1.125 = 54 Mbps. Due to the limitations of the MAC layer mechanism, however, actual data rates are only about 70 percent of theoretical, so that sustained data rates for 802.11a are rarely above 35 Mbps. Nevertheless, these high attainable speeds are sufficient to enable 802.11a systems to support some types of broadband applications, including Web browsing and multimedia.
The frequencies used by 802.11a cover three 200 MHz blocks of the 5 gigahertz (GHz) band:
5.15 to 5.25 GHz: This band supports five independent channels with a maximum power output of 50 megawatts (mW).
5.25 to 5.35 GHz: This band also supports five independent channels but with a maximum power output of 250 mW for greater distance and penetration of building materials.
5.725 to 5.825 GHz: This upper band is designed for outdoor communications, supports five independent channels, and a maximum power output of 1 watt.
Advantages and Disadvantages
The main advantage of 802.11a over 802.11b is the increased data rate. By way of analogy, 802.11a is to 802.11b as Fast Ethernet is to standard 10 Mbps Ethernet. In fact, both of these standards are sometimes referred to as wireless Ethernet, but this is a misnomer because they use CSMA/CA instead of Ethernet's Carrier-Sense Multiple-Access with Collision Detection (CSMA/CD) mechanism at the MAC layer.
Besides the higher data rates supported by 802.11a as compared to 802.11b, another advantage of 802.11a is that the COFDM encoding mechanism includes greater error correction, which means better recovery from multipath reflection.
Implementation
As a result of these differences in signal encoding mechanisms, 802.11b and 802.11a are incompatible standards. Devices using one standard cannot communicate with devices supporting the other standard. Yet because these two technologies operate at different frequency bands, they can be deployed together and overlap without interference. 802.11b is already widely deployed in many corporate enterprises; the logical migration path to 802.11a is to deploy 802.11a in small pockets where it is most required and gradually extend this as needed.
Marketplace
Two vendors pioneering in 802.11a hardware include Radiata Communications (now owned by Cisco Systems) and Atheros Communications. Both vendors offer devices supporting 6, 12, 24, 36, 48, and 54 Mbps transmission speeds, with different rates being achieved by using different encoding mechanisms. Atheros also supports bonding two channels to achieve a theoretical maximum transmission rate of 108 Mbps, and other vendors are developing proprietary extensions to 802.11a that may boost speeds up to 155 Mbps.
Issues
Because of the higher frequencies used, 802.11a devices are not susceptible to the kind of electromagnetic interference, including interference from microwave ovens and cell phones, that plagues 802.11b devices. However, because of different governmental policies regarding allocation of frequencies in the 5 GHz band, interoperability issues exist for 802.11a technologies used in different parts of the world.
For example, in Europe the upper 100 MHz portion (5.725 to 5.825) is already allocated for other uses, limiting operation of 802.11a gear to only 10 clear channels instead of 15. The 802.11a standard is also incompatible with the competing HiperLAN/2 standard being developed in Europe, which also uses frequencies from the 5 GHz band. In Japan only the lower 100 MHz portion is unallocated, which allows only five channels for 802.11a to work with. Other countries and regions use portions of the 5 GHz spectrum for military or satellite communications purposes, limiting the capability of 802.11a devices within their borders.
See Also 802.11 ,802.11b wireless networking
An Institute of Electrical and Electronics Engineers (IEEE) wireless networking communications standard that supports data transmission rates of up to 1 megabit per second (Mbps). 802.11b is sometimes called wireless Ethernet, but this is a misnomer because the two technologies use different Media Access Control (MAC) methods.
Overview
The 802.11b standard was developed over a number of years by Lucent Technologies, Intersil Corporation, and other industry partners and ratified as an IEEE standard in 1999. It is the most popular and widely deployed wireless networking standard currently in use, though it may be superseded over the next few years by 802.11a, which supports higher data transmission rates.
802.11b is capable of supporting data transmission rates of 1, 2, 5.5, and 11 Mbps. The actual transmission rate used will depend on the distance between the client adapter (PC card or network interface card [NIC]) and access point (base transmitting station) and upon interference due to obstacles in the transmission path. In general, the further the distance from the base station, the slower the data rate. For directional antennae operating with clear line-of-sight in point-to-point transmission, the top speed of 11 Mbps can be sustained for distances up to about 10 miles (16 km). Within a building using omnidirectional antennae, however, devices may need to be within 100 feet (15 meters) in order to maintain their highest transmission rate. If a device exceeds this range, the transmission rate drops to the progressively lower levels until a sustained transmission rate can be successfully negotiated. Each data rate uses its own characteristic modulation scheme, such as Binary Phase Shift Keying (BPSK) for 1 Mbps transmission, Quadrature Phase Shift Keying (QPSK) for 2 Mbps, and so on.
Although the theoretical top transmission speed for 802.11b is 11 Mbps, in practice the media access control (MAC) layer mechanism (Carrier Sense Multiple Access with Collision Avoidance [CSMA/CA]) and protocol overhead reduce efficiency to 60-70 percent. The result is that data rates of more than 6 Mbps are rarely achieved and 4-5 Mbps is more common. Note that this is the total shared bandwidth for all 802.11b clients within the same service area (serviced by the same set of access points or base stations), so actual throughput may be much less when many clients are active. The 802.11b standard allows multiple users to share a single communications channel at the same time by using the CSMA/CA method as its MAC layer protocol. The 802.11b standard supports Request to Send/Clear to Send (RTS/CTS) as a MAC enhancement for transmission in environments where coverage is distorted by obstacles and interference, but this mechanism adds significant overhead to the protocol, especially during the transmission of small packets.
The drawing shows a typical 802.11b packet. This consists of a 30-byte header containing information for addressing, sequencing, and frame control. This is followed by a data or payload section that may range from zero to 2312 bytes in length, followed by a 4-byte checksum trailer.
802.11b. Typical structure of an 802.11b packet. 
As far as frequency and transmission method are concerned, 802.11b employs direct-sequence spread spectrum (DSSS) transmission over an 83-MHz-wide chunk of the unlicensed 2.4 GHz Industrial, Scientific, and Medical (ISM) band of the electromagnetic spectrum, specifically the frequency band from 2.400 to 2.483 GHz. Because this frequency band overlaps with emissions from microwave ovens and some cell phones, proximity of these devices can affect the transmission speed achieved.
Under Federal Communications Commission (FCC) regulations, 802.11b provides 11 separate channels for communications. Theoretically, this means that the total maximum bandwidth for an 802.11b local area network (LAN) utilizing all channels would be 11 x 11 Mbps = 121 Mbps. In practice, due to interference between channels, a given service area can only support three simultaneous channels, which limits total available bandwidth to 3 x 11 = 33 Mbps in theory, or about 14 Mbps in practice once protocol overhead is factored in.
A typical 802.11b wireless LAN segment may support up to 65 different clients, but the contention involved in the CSMA/CA media access method means that one busy client downloading a large file can utilize the segment's entire available bandwidth, leaving the other stations unable to communicate. Adding additional access points to the service area provides fault tolerance and load balancing but does not increase the total available bandwidth.
Implementation
When deciding whether you should immediately deploy 802.11b equipment in your enterprise or wait for the faster 802.11a to gain greater vendor support, you have several issues to consider. Small- and mid-sized companies might find 802.11b a cheap, easy, turnkey solution to deploy and might not need the higher data rates of 802.11a. And although both systems can be deployed in parallel, 802.11b equipment cannot be upgraded to 802.11a. Dual-radio access points are under development, but their deployment is complicated by the two standards' different transmission characteristics. Finally, enterprise architects may want to wait for 802.11a products that are certified by the Wireless Ethernet Compatibility Alliance (WECA) before committing large portions of their budget to a massive rollout of 802.11a.
Marketplace
More than 20 different vendors sell 802.11b-compliant hardware, and the standard`s success is evident in the high degree of interoperability between equipment supplied by different vendors. A client network adapter (NIC or PC card) from one vendor is likely to work well with an access point (base station hub) from a different vendor without needing any modifications. Access points from different vendors are less likely to work together, particularly if they need to be configured for roaming users that move from one service area (cell) to another and require seamless networking during the transition. So the best solution is still to purchase all of your 802.11b devices from the same vendor in order to achieve the highest level of interoperability, particularly in large, geographically dispersed deployments with multiple access points spanning multiple subnets.
As a result of increased competition, prices for 802.11b equipment have fallen drastically of late. Client adapters are now available for $99 or less, and access points can be had for under $299. Some popular vendor products include
3Com Corporation: AirConnect PC card and companion AirConnect Wireless Access Point designed for the enterprise, which supports Mobile IP and Microsoft Point-to-Point Encryption
Apple Computer: AirPort client adapter and companion Airport access point, a bargain system designed for Small Office Home Office (SOHO) environments and including Ethernet and V.90 modem support
Intel Corporation: PRO/Wireless LAN PC card and companion PRO/Wireless LAN Access Point, with integrated Web server and 128-bit Wired Equivalent Privacy (WEP) support. Intel also recently acquired Xircom, which makes a wireless Ethernet adapter that is popular with Notebook users and supports 40- and 120-bit WEP.
Issues
While the maximum allowed power output for an 802.11b-compliant device is 1 watt, most 802.11b-compliant NICs and PC cards emit only about 30 milliwatts (mW) to conserve battery life in laptops and avoid excessive heat dissipation. This is generally considered too low to constitute any health hazard to individuals using this equipment-for comparison, a typical cell phone may generate an electromagnetic power output of half a watt or more.
Notes
For an explanation of the different components needed in the deployment of a wireless network and how they operate, see the article "wireless networking" elsewhere in this book.
Do not operate 802.11b devices in an aircraft during flight-turn your device off when boarding. The signals transmitted by your device could adversely affect the aircraft's electronic systems.
Look for the "WiFi" seal of approval on vendor packaging of 802.11b equipment. This designation indicates that the vendor's equipment is approved by WECA as being fully compliant with 802.11b standards.
Wireless networks based on 802.11b can be susceptible to electromagnetic interference produced by microwave ovens, particularly if the clients are located far from the base station. If this occurs, move the oven, move the client adapter, or increase the transmission strength of the access point until you can achieve reliable communications.
To conserve power, especially for 802.11b PC cards in laptops, these client adapters should be set to run in Polled Access Mode (PAM). Clients configured for PAM go to sleep when not communicating with the base station and wake up according to a predefined schedule to check for an inbound Traffic Information Map (TIM) packet from the base station indicating that it has data ready to send the client. PAM typically uses about 10 percent of the power consumed when running in Constant Access Mode (CAM), where the client is always listening.
If your buildings have metal floors, wireless networking using 802.11b systems may not operate beyond 100 feet (30 meters) due to signal absorption. In such buildings you may also require an access point on each floor.
802.11b and Bluetooth devices operate on the same frequencies, so using both systems in the same service area may cause dropouts and transmission errors to occur.
See Also 802.11 ,802.11b Bluetooth, wireless networking
The Institute of Electrical and Electronics Engineers (IEEE) standards outlining the architecture and operation of the Demand Priority Media Access Control (MAC) method. This access method is employed by the legacy 100VG-AnyLAN technology developed in 1994 by Hewlett-Packard and AT&T. The IEEE 802.12 working group is presently in hibernation and is no longer active.
See Also 100VG-AnyLAN ,demand priority ,Project 802
The Institute of Electrical and Electronics Engineers' (IEEE's) Project 802 includes a number of standards and working groups relating to networking. These standards and working groups currently range from 802.1 through 802.17, but for some reason 802.13 was not used.
See Also Project 802
The Institute of Electrical and Electronics Engineers (IEEE) working group on cable modems. This working group is no longer active.
See Also Project 802
A set of standards being developed by the Institute of Electrical and Electronics Engineers (IEEE) for Wireless Personal Area Networks (Wireless PANs or WPANs).
Overview
The 802.15 working group is currently divided into a number of Task Groups (TGs) that are actively working on different aspects of wireless PANs. These TGs include
TG1: This group is working to develop standards for WPANs based on Bluetooth 1.x wireless technologies.
TG2: This group is trying to get 802.15 WPANs to coexist with 802.11 WLANs by overcoming issues related to electromagnetic interference between the two technologies.
TG3: This group is trying to develop High Rate (HR) WPANs with speeds in excess of 20 megabits per second (Mbps).
TG4: This group deals with developing Low Rate (LR) WPANs with speeds in the range of 10-200 kilobits per second (Kbps). These LR WPANs will be used for things such as home automation, remote control, interactive toys, sensors, and similar systems.
See Also Personal Area Network (PAN) ,Project 802
A set of emerging Institute of Electrical and Electronics Engineers (IEEE) standards being developed for fixed broadband wireless networking.
Overview
Broadband wireless, also called Wireless MAN (WMAN), involves wireless high-speed data transmission between fixed points at various frequency bands. In other words, 802.16 is to the metropolitan area network (MAN) what 802.11 is to the local area network (LAN): similar technology but operating at higher speeds and over longer distances. WMAN is the wireless equivalent of broadband wired technologies such as Digital Subscriber Line (DSL).
The IEEE working group on 802.16 is subdivided into a number of Task Groups (TGs) to work at these different frequencies, specifically:
TG1: This group is working on air interfaces in the 10- 66gigahertz (GHz) region of the electromagnetic spectrum.
TG2: This group is working on ensuring coexistence between different wireless broadband access systems.
TG3: This group is working on air interfaces in the 2-11 GHz region of the spectrum.
TG4: This group is working on the Wireless High-Speed Unlicensed Metropolitan Area Network (Wireless HUMAN) standard, which involves transmission in license-exempt portions of the spectrum in the 5- 6GHz range.
See Also Project 802
The Institute of Electrical and Electronics Engineers (IEEE) working group on Resilient Packet Ring (RPR) technologies.
See Also Project 802
An Institute of Electrical and Electronics Engineers (IEEE) standard for implementing 1000 megabits per second (Mbps) Ethernet (that is, Gigabit Ethernet, or GbE) over shielded twisted-pair (STP) cabling.
Overview
The CX in 1000BaseCX stands for short-haul copper and indicates that this version of GbE is intended for short cable runs over copper cabling. 1000BaseCX technologies are in the beginning stages of being widely implemented in enterprise-level networks and are primarily used for collapsed backbones and high-speed interconnects within wiring closets and equipment rooms. 1000BaseCX and other GbE standards are defined in the 802.3z series of standards of Project 802 developed by the IEEE.
Implementation
1000BaseCX is to some degree simply an extension of standard Ethernet technologies to Gigabit-level network speeds. 1000BaseCX is implemented using STP cabling. This STP cabling must be standard 150-ohm balanced cabling and should have a quality slightly better than IBM Type I cabling. Cable segments can have a maximum length of only 82 feet (25 meters).
1000BaseCX. A typical 1000BaseCX network.
Like other forms of GbE, 1000BaseCX employs 8B/10B coding with a transmission frequency transmission of 1.25 GHz. 1000BaseCX is intended mainly for connecting high-speed hubs, Ethernet switches, and routers together in wiring closets. Common implementations for 1000BaseCX are in switch-switch and switch-server connections, with switch-server connections being the most frequently implemented use for 1000BaseCX.
See Also 1000BaseLX ,1000BaseSX ,1000BaseT Gigabit Ethernet (GbE)
An Institute of Electrical and Electronics Engineers (IEEE) standard for implementing 1000 megabits per second (Mbps) Ethernet (that is, Gigabit Ethernet, or GbE) over fiber-optic cabling.
Overview
The LX in 1000BaseLX stands for long and indicates that this version of GbE is intended for use with long-wavelength (1300 nm) laser transmissions over long cable runs of fiber-optic cabling. 1000BaseLX technologies are in the beginning stages of being widely implemented in enterprise-level networks and are primarily used for long cable runs between pieces of equipment on a campus or within a building. 1000BaseLX and other GbE standards are defined in the 802.3z standards of Project 802 developed by the IEEE.
Implementation
1000BaseLX can be implemented using either single-mode fiber-optic cabling or multimode fiber-optic cabling. Transmission requires two optical fibers due to the full-duplex operation of 1000BaseLX. Cable segment lengths depend on the cable grade used, as shown in the table on the following page.
1000BaseLX. A typical 1000BaseLX network.
Cable Grade | Maximum Segment Length |
10-micron single- mode fiber | 10 kilometers (6.2 miles) |
50-micron multimode fiber | 550 meters (1800 feet) |
62.5-micron multimode fiber | 440 meters (1440 feet) |
1000BaseLX is intended mainly for connecting high-speed hubs, Ethernet switches, and routers together in different wiring closets or buildings using long cabling runs. 1000BaseLX is most commonly implemented in a switch-switch configuration.
Notes
When you use multimode fiber-optic cabling in 1000BaseLX implementations, a condition called differential mode delay (DMD) can sometimes occur. This condition occurs only in cabling of uneven quality, and it leads to signal jitter that can disrupt network communications. To resolve this problem, newer 1000BaseLX transceivers condition the signal to distribute its power equally among all transmission modes of the cable.
See Also 1000BaseCX ,1000BaseSX ,1000BaseT Gigabit Ethernet (GbE)
An Institute of Electrical and Electronics Engineers (IEEE) standard for implementing 1000 megabits per second (Mbps) Ethernet (that is, Gigabit Ethernet, or GbE) over fiber-optic cabling.
Overview
The SX in 1000BaseSX stands for short and indicates that this version of GbE is intended for use with short-wavelength (850-nanometer) laser transmissions over short cable runs of fiber-optic cabling. 1000BaseSX technologies are in the beginning stages of being widely implemented in enterprise-level networks and are primarily used for shorter cable runs between pieces of equipment within a building.
1000BaseSX is an extension of standard Ethernet technologies to gigabit-level network speeds. 1000BaseSX and other GbE standards are defined in the 802.3z standards of Project 802 developed by the IEEE.
Implementation
1000BaseSX is implemented using only multimode fiber-optic cabling. Cable segment lengths depend on the cable grade used, as shown in the table.
Cable Grade | Maximum Segment Length |
50-micron multimode fiber | 500 meters (1640 feet) |
62.5-micron multimode fiber | 220 meters (720 feet) |
1000BaseSX is intended mainly for connecting high-speed hubs, Ethernet switches, and routers together in different wiring closets or buildings using long cabling runs. 1000BaseSX is most commonly implemented in a switch-switch configuration.
Notes
When multimode fiber-optic cabling is used in 1000BaseSX implementations, a condition called differential mode delay (DMD) can sometimes occur. This condition occurs only in cabling of uneven quality, and it leads to signal jitter that can disrupt network communications. To resolve this problem, newer 1000BaseSX transceivers condition the signal to distribute its power equally among all the cable's transmission modes.
See Also 1000BaseCX ,1000BaseLX ,1000BaseT Gigabit Ethernet (GbE)
An Institute of Electrical and Electronics Engineers (IEEE) standard for a form of Gigabit Ethernet (GbE) involving transmission over copper cabling.
Overview
The T in 1000BaseT identifies it as operating over twisted-pair copper cabling and makes it an extension of the traditional 10BaseT and 100BaseT Ethernet technologies for transmission over unshielded twisted-pair (UTP) cabling. 1000BaseT is based on the IEEE 802.3ab standard and is intended to provide the simplest upgrade path for legacy 10BaseT and 100BaseT Ethernet networks.
1000BaseT is implemented using the commonly installed Category 5 (Cat5) cabling or enhanced Category 5 cabling version of UTP cabling. Cable segments for 1000BaseT must have a maximum length of 328 feet (100 meters). Category 6 (Cat6) cabling is expected to provide enhanced transmission characteristics for 1000BaseT networks when it becomes available.
1000BaseT uses all four pairs of wires in standard UTP cabling, as opposed to the two pairs of wires used in 10BaseT and 100BaseT networks. Using all four pairs of wires brings certain problems because of attenuation, crosstalk, and echoes arising from full-duplex transmission over single wires. Full-duplex transmission is used to enable each pair of wires to simultaneously transmit and receive data at a rate of 250 megabits per second (Mbps). The resulting four pairs of wires means that the total data rate is 4 x 250 = 1000 Mbps or 1 Gbps. Although each pair of wires carries data at 250 Mbps, the signaling is only at 125 Mbps. This is accomplished by using five-level pulse amplitude modulation (PAM) as the line coding mechanism, which encodes two bits of information for each signal pulse. The reason for this choice is that Cat5 cabling works well up to 125 megahertz (MHz) in standard installations.
1000 BaseT has been standardized by the IEEE under the 802.3ab specification since June 1999. This specification provides for the implementation of automatic link negotiation so that problems such as crossed cables can be detected and corrected for, that 100 Mbps interface can be detected and accommodated when connected to 1000BaseT ports, and that half-duplex network interface cards (NICs) and hub/switch ports can be recognized and used when detected. The standard also specifies special filters for hybrid circuits used in full-duplex transmission over single wires, a special five-level PAM encoding mechanism instead of binary signals, forward error correction techniques, and pulse shaping technologies to make 1000BaseT a functional and reliable networking technology.
Implementation
One popular use for 1000BaseT (also referred to as GbE over copper) is in wiring closets of enterprise networks. 1000BaseT GbE switches can be used to aggregate Fast Ethernet workgroup switches by first connecting the workgroup switches to 100/1000 Mbps ports on the GbE switches using Cat5 cabling and then uplinking the GbE switches to the network's GbE backbone switches using fiber.
1000BaseT. A typical 1000BaseT network.
Another growing use for 1000BaseT GbE switches is for connecting high-performance multimedia workstations and Web server clusters to the network backbone. The main limitation for 1000BaseT workstations is that they must be located within 100 meters of the switch. For interbuilding GbE links you need to use fiber. Note, however, that 1000BaseT only supports cable runs up to 328 feet (100 meters) in length, which means that you may need to modify existing cabling infrastructures in some large buildings when migrating from Fast Ethernet to 1000BaseT. Despite the potential application for supporting high-performance workstations, the most popular uses for 1000BaseT remain for backbone uses such as installing vertical risers between floors and horizontal cable runs between wiring closets, for interconnecting workgroup switches, and for connecting server farms and clusters to backbone switches.
Marketplace
Adapting the Ethernet standard to transmission at rates of 1 Gbps over existing Cat5 wiring means lower costs for companies planning to upgrade their switched backbones from 100 Mbps Fast Ethernet to GbE. A typical 1000BaseSX fiber port costs about 50 percent more than a comparable 1000BaseT copper port, so considerable cost savings can be achieved by implementing GbE over copper instead of fiber. However, copper is more susceptible to noise than fiber for gigabit transmission, and Cat5 cabling must be of the highest standard and deployed according to standards to ensure maximum reliability.
1000BaseT solutions are beginning to be implemented for short, high-speed interconnects within wiring closets in enterprise networks, for consolidating server farms in data centers, and for interconnects in carrier colocation sites. Some also envision using it for connecting high-performance workstations as well. Vendors of 1000BaseT equipment include Cisco Systems, 3Com Corporation, Hewlett-Packard, Asante Technologies, Nortel Communications, Extreme Networks, Intel Corporation, and Foundry Networks. Cisco's Catalyst 6500 for its Catalyst 6506 switch chassis module provides 16 1000BaseT ports and is popular with Internet service providers (ISPs) deployed at carrier colocation sites because of its fault-tolerant load balancing and realtime code upgrade capabilities. Foundry's eight-port 1000BaseT modules for its BigIron 8000 switch chassis is another popular choice with excellent performance.
A number of vendors are also providing 1000BaseT NICs for use in high-speed servers and high-performance networks connected to GbE backbones. These vendors include 3Com, Intel, Sun Microsystems, and SysKonnect. A popular choice in enterprise server farms is the SysKonnect 9822 64-bit 66 MHz PCI card, which comes in single-port and dual-port configurations. Alteon WebSystems also has a 10/100/1000BaseT NIC with support for auto-negotiation and jumbo frames.
Currently 1000BaseT is the fastest growing type of GbE being deployed in enterprise networks, because it leverages the investment of existing Cat5 cabling installations. In the second quarter of 2000, GbE over copper accounted for 25 percent of GbE port sales.
Notes
1000BaseT is also sometimes referred to as 1000BaseTX.
Before installing 1000BaseT switches in a building wired with older Cat5 cabling (as opposed to newer enhanced Category 5 [Cat5e] cabling), be sure to test the cabling using a cable tester to make certain the installation fully complies with IEEE TSB95 field specifications. Common reasons for existing wiring failing to support GbE are crosstalk caused by improper cable termination in punchdown block and wall plates. Hewlett-Packard and Fluke Corporation are two vendors that sell suitable cable test equipment. Also, be sure also to use high-quality Cat5e patch cables for connecting patch panels to switches in wiring closets and for connecting servers or workstations to wall plates in work areas.
See Also 1000BaseCX ,1000BaseLX ,1000BaseSX Gigabit Ethernet (GbE)
Another name for 1000BaseT Gigabit Ethernet (GbE) over copper.
See Also 1000BaseT
A generic designation used to describe all forms of Gigabit Ethernet (GbE), including 1000BaseCX, 1000BaseLX, 1000BaseSX, and 1000BaseT. 1000BaseX actually is the technical designation for the family of physical layer (PHY) technologies used in GbE and evolved directly from the Fiber Channel ANSI94 standards.
See Also 1000BaseCX ,1000BaseLX ,1000BaseSX Gigabit Ethernet (GbE)
An Institute of Electrical and Electronics Engineers (IEEE) standard for a high-speed serial transmission technology also known as FireWire (a popular term coined by Apple Computer) and iLink (a Sony Corporation trademark).
See Also FireWire
The original host-to-IMP (interface message processor) interface for the Advanced Research Projects Agency Network (ARPANET), the precursor of the Internet.
See Also ARPANET
An information display protocol for IBM mainframe computers.
Overview
The 3270 protocol is a family of protocols that includes separate specifications for terminal display and printer output. The 3270 protocol enables text-based connection-oriented conversations to take place between centralized mainframe hosts and supported peripherals, including terminals, printers, and controllers. These conversations originally took place over dedicated serial transmission links between the terminal and the mainframe controller. For long-distance computing, transmission over private wide area networks (WANs) such as the IBM Global Network enabled mainframe COBOL programmers to work from home instead of having to commute to the office.
Private WAN services such as IBM Global Networks have now been dismantled and replaced with the Internet, and Web-based access to mainframe hosts has led to the development of hardware-based and software-based emulators. A 3270 emulator is a device or program that enables a workstation to communicate with a mainframe as if it were a dedicated 3270 terminal. A typical 3270 software emulator might use ActiveX controls or Java applets to display the 3270 terminal interface running within a Web browser window. These controls or applets are downloaded from a Web server acting as an intermediary between the Transmission Control Protocol/Internet Protocol (TCP/IP) local area network (LAN) on which the workstations reside and the mainframe host's Systems Network Architecture (SNA) network. Web- based 3270 emulators are popular in many corporate intranets where they can help leverage the usefulness of legacy mainframes and enable offsite users to access mainframe databases through the Internet.
Notes
You can use Microsoft SNA Server to enable a Windows 2000 or Windows .NET Server network to access a mainframe's hierarchical SNA network. SNA Server accomplishes this by defining and assigning dependent 3270 Logical Units (LUs). SNA Server client software must also be installed on each Windows 2000, Windows XP, or Windows .NET Server client that uses these SNA Server LU services. The SNA Server client software enables communication between 3270 applications on the client and the SNA server that provides a gateway to the mainframe.
The original 3270 protocol transmits streams of clear text information. If you are running terminal emulation software on a remote workstation and accessing the mainframe over the Internet, you need to make sure you are using Secure Sockets Layer (SSL) or some other method of securing your transmission.
See Also Web-to-host
An information display protocol for IBM AS/400 systems.
Overview
The 5250 protocol enables text-based connection- oriented conversations to take place between AS/400 hosts and supporting peripherals, including terminals, printers, and controllers. The 5250 protocol is a family of protocols that includes separate protocols for terminal display and printer output. AS/400 display sessions use 5250 data streams over Advanced Peer-to-Peer Networking (APPN).
Recent years have seen the proliferation of Web-based software for enabling local area network (LAN) workstations to access information on AS/400 hosts. These solutions are called terminal emulators and may be either hardware- or software-based. A typical 5250 emulator is an application that uses ActiveX controls or Java applets to display the 5250 terminal interface running within a Web browser window. Emulation controls or applets are downloaded from a Web server acting as an intermediary between the Transmission Control Protocol/Internet Protocol (TCP/IP) LAN on which the workstations reside and the AS/400 host.
Notes
Printer emulation is the ability to print locally from a workstation running a terminal emulator. In order for 5250 terminal emulators to support printer emulation, they must conform to the TN5250 Telnet protocol. Otherwise, workstations must print to a centralized printer attached to the AS/400.
Microsoft SNA Server supports both 3270 and 5250 sessions between Microsoft Windows clients and IBM mainframes and AS/400 systems.
The 5250 protocol transmits streams of clear text information, so if you are running terminal emulation software on a remote workstation to access your company's AS/400 over the Internet, you need to make sure you are using Secure Sockets Layer (SSL) or some other method for securing your transmission.
See Also Web-to-host
