Site home page
Get alerts when Linktionary is updated
Book updates and addendums
Get info about the Encyclopedia of Networking and Telecommunicatons, 3rd edition (2001)
Download the electronic version of the Encyclopedia of Networking, 2nd edition (1996). It's free!
Contribute to this site
Electronic licensing info
Note: Many topics at this site are reduced versions of the text in "The Encyclopedia of Networking and Telecommunications." Search results will not be as extensive as a search of the book's CD-ROM.
Fast Ethernet is traditional CSMA/CD (carrier sense multiple access/collision detection) access control at 100 Mbits/sec over twisted-pair wire. The original Ethernet data rate was 10 Mbits/sec.
During the early development of Fast Ethernet, two different groups worked out standards proposals-and both were finally approved, but under different IEEE committees. One standard became IEEE 802.3u Fast Ethernet and the other became 100VG-AnyLAN, which is now governed by the IEEE 802.12 committee. The latter uses the "demand priority" medium access method instead of CSMA/CD.
100VG-AnyLAN has not caught on, while Fast Ethernet enjoyed great success in enterprise LAN environments. At about the same time, switching technologies came on the scene to further improve network performance. More recently, Gigabit Ethernet has come on the scene to provide a high-speed upgrade path for Fast Ethernet users. It provides high-speed backbones that interconnect Fast Ethernet LANs. See "Gigabit Ethernet."
Fast Ethernet takes advantage of the scalability of CSMA/CD hierarchical networking. It is designed after the 10Base-T standard and can be built into hierarchical networking topologies like that shown in Figure F-1. This type of configuration is compatible with structured wiring strategies, as discussed under "TIA/EIA Structured Cabling Standards."
The primary concern of Fast Ethernet developers was to preserve the CSMA/CD medium access method of 802.3 Ethernet while boosting the data rate. In addition, the developers kept the frame format. Because of this, Fast Ethernet fits in well with traditional Ethernet installations. One version of the Fast Ethernet standard will run on older Category 3 cable installations. Multiple Ethernet types (10 and 100 Mbits/sec) may coexist. A Fast Ethernet-compatible hub simply needs to perform speed matching when exchanging frames. An autonegotiate feature allows devices to detect the speed of incoming transmissions and adjust appropriately.
There are four Fast Ethernet schemes:
The higher frequency used in the Fast Ethernet standard is prone to attenuation, so cable distance is more limited than in the Ethernet 10Base-T specification. If the encoding scheme of traditional Ethernet were used with Fast Ethernet, the high-end frequency would be above 200 Mhz. That is double the maximum frequency rating of Category 5 cable. To get around this, new encoding schemes were implemented to allow higher-frequency transmissions. Refer to the paper called "Specifications for Data Encoding in Fast Ethernet" at the Optimized Engineering Web site. See the related entries and Web links page.
Note that the collision domain is 1/10th the size of the collision domain in 10-Mbit/sec Ethernet. This is because of the 10-times increase in speed of the network. It also means that Fast Ethernet networks can only be one-tenth the physical size of a 10Base-T network.
Fast Ethernet can support full-duplex switched networking modes to provide even better performance. A full-duplex nonshared link (i.e., workstations are attached to switches, not hubs) doesn't even need CSMA/CD, because no other stations are trying to use the link, and each end system has its own channel to transmit on. Collision detection and loopback functions can be disabled. This arrangement is especially useful for backbone connections. If both end systems are transmitting at the same time, the combined data rate is 200 Mbits/sec.
100Base-TX works with two pairs of UTP (unshielded twisted pair) or STP (shielded twisted pair). Transmission takes place on one pair of wires and collision detection takes place on the other pair of wires. Two types of cable can be used with this specification: Category 5 UTP and IBM's Type 1 STP. Category 5 cable has four pairs of wire. Since only two of the pairs are used, the other two are available for other uses or future expansion. However, it is not recommended that another high-speed network be used on the pairs.
Category 5 cable is designed to cancel out the effects of EMI and, therefore, has very stringent installation requirements. All the components in the cabling system must be Category 5 compliant, including connectors, patch panels, punchdown blocks, and hubs/switches. The twists in the cable must be maintained all the way up to within 1/2 inch of a connector. The twists in the cable help maintain proper signaling, especially at high data rates. Other 100Base-TX requirements are
ANCHOR HERE: Figure 2 (see book)
The 100Base-T4 standard has the same cabling specifications as 100Base-TX in terms of distances and hub configuration. The difference is that 100Base-T4 uses all four wires in a half-duplex signaling scheme. Three pairs are used to either transmit or receive data, and the other pair is used for collision detection. 100Base-T4 also uses a special three-level encoding scheme (as opposed to two levels in other media) to reduce the clock rate. Spreading the 100-Mbit/sec signal over three pairs reduces the signal frequency and allows it to run on older Category 3 cable. Basically, 33.33 Mbits/sec is transmitted over each of the three pairs.
Higher-grade cable such as Category 5 is recommended for future expansion. Like 100Base-TX, the 100Base-T4 specification has a maximum hub-to-station cable length of 100 meters and an end-to-end maximum of 250 meters. 100Base-T4 also uses the same repeater classing scheme as 100Base-TX. Class I hubs allow a mix of different Fast Ethernet network types, and class II hubs allow only one scheme (refer to the description in the previous section).
100Base-FX is the fiber-optic cable implementation of the Fast Ethernet standard. It is ideal for building backbone connections. Cable distance is limited to 412 meters; but if full-duplex mode is used, cable runs may be as long as 2 km. While fiber-optic cable can span longer distances, the limit is imposed to account for packet round-trip timing.
Fiber-optic cable is not prone to interference. It does not emanate a signal, so it is more secure (especially for wiring across public areas). In addition, fiber-optic cable can scale up to higher transmission rates for future expansion.
100Base-FX requires a cable with two strands of 62.5/125-micron fiber. One strand is used for signal transmission while the other is used to receive and detect collisions. 100Base-FX also uses the same repeater classing scheme as 100Base-TX. Class I hubs allow mixed Fast Ethernet networks, and class II hubs allow only one scheme (refer to the description in the previous section).
Copyright (c) 2001 Tom Sheldon and Big Sur Multimedia.