TOKEN RING: IEEE STANDARD 802.5
Token ring LAN is defined by the IEEE standard 802.5. Like Ethernet, the token ring is a MAC protocol sitting between the Logical Link Control (LLC) and the physical layer in the OSI model.
Stations on a token ring LAN are connected in a ring using a NIC, as shown in Figure. A station can send directly only to its neighbours and in most cases, only to one neighbour (clockwise in Figure). If a station wants to send to another station on the ring, the frame must go through all the intermediate interfaces. Ring contention is handled through a token (a special frame) that circulates past all the stations. The specifics of claiming tokens and sending frames are involved.
The network access mechanism used by Ethernet (CSMA/CD) is not infallible and may result in collisions. Stations may attempt to send data multiple times before a transmission makes it onto the link. This redundancy may create delays of indeterminable length if the traffic is heavy. There is no way to predict either the occurrence of collisions or the delays produced by multiple stations attempting to capture the link at the same time.
Token ring resolves this uncertainty by requiring that stations take turns sending data. Each station may transmit only during its turn and may send only one frame during each turn. The mechanism that coordinates this rotation is called token passing.
FIBRE DISTRIBUTED DATA INTERFACE (FDDI)
The fibre distributed data interface (FDDI) standard for a 100 Mbps fibre optic LAN was developed during the mid-1980s by a subcommittee of ANSI and was completed in 1990. As LANs based on the IEEE 802 standards reached capacity, optical fibre LANs based on the FDDI standard became an alternative growth path. In their first implementations FDDI LANs were used to provide high-speed backbone connections between distributed LANs. Now, however, FDDI is used for normal LANs that have a large number of users or need to operate at very high speed.
The primary transmission medium used for FDDI is optical fibre, and the standard defines two types: Single mode fibre (SMF) and multimode fibre (_'VIIVIF). Single mode fibre can deliver connectivity over longer distances, with higer performance than MMF. It is used, consequently, for connections between buildings or over greater geographical areas. Multimode fibre is usually used to connect devices within a building or a small geographically contained area.
It is interesting that FDDI has also been implemented over twisted pair copper wire. The copper distributed data interface (CDDI), as it is now called, uses only shielded twisted pair or unshielded twisted pair category 5 cabling but supports distances of 100 metres and data rates of 100 Mbps. Other than the media, the implementation of FDDI and CDDI is the same.
FDDI also uses redundancy to overcome failures. An FDDI network contains two complete rings-one that is used to send data when everything is working correctly, and another that is used only when the first ring fails. Physically, the two fibres connecting a pair of computers are not completely separate. Instead, each fibre is covered with a flexible plastic jacket, and the jackets of pair are joined in the same way as the plastic jacket on the wires in the power cord of a household appliances. Therefore, the fibres needed for two rings can be installed at the same time.
Interestingly, the rings in an FDDI network are called counter rotating because data flows around the second ring opposite of the direction data flown around the main ring. To understand the motivation for counter rotating rings, consider how catastrophic failures occur.
First, because the pair of fibres connecting two stations usually follows the same path, an accident that breaks one fibre often breaks both. Second, if data always passes in the same direction ocross both rings, disconnecting one station from the ring (e.g.,)while moving a computer) will prevent other stations from communicating. However, if data travels in the reverse direction across the second ring, the remaining stations can reconfigure the network to use the reverse path.
(a) An FDDI network with arrows showing the directions that data flows and
(b) the same network after the station has failed.
Although Figure 6.9(a) shows the directions data can travel in a counter rotating rings, only one of the two rings is normally used. For example, in the figure, a station always transmits and receives frames on. the outer rings, while the network hardware forwards bits on the inner ring without interpreting them. Figure 6.9(b) illustrates the data following a failure. Hardware in the stations adjacent to a failure detect the disconnection and reconfigure so they loop incoming bits back along the reverse path.Thus. the failed station is removed and the remaining stations are connected to a contiguous ring. The process of reconfiguring to avoid a failure is called self-healing, and FDDI is known as a self-healing network. We can summarize. An FDDI network is called self-healing because the hardware can detect a catastrophic failure and recover automatically. To do so, FDDL uses a pair of counter rotating rings. One ring is used to transmit data. When a failure occurs that breaks the ring, stations adjacent to the failure automatically reconfigure, using the second ring to bypass the failure.
DISTRIBUTED QUEUE DUAL BUS (DQDB): IEEE STANDARD 802.6
Local area networks are usually restricted to a single such as a building or a floor or room in a building. Metropolitan Area Network (MAN) expands network coverage to include several buildings or sites within a limited area. These sites are usually located within a city's limits or limited to a specific region that mav include a number of cities. IEEE standard 802.6 defines the Distributed Queue Dual Bus (DQDB) which resembles a LAN standard. It is designed to be used in LAN. As the name implies, DQDB uses a dual bus configuration. Each device in the system connects to two backbone links. Access to these links is granted not by contention (as in 802.3) or token passing (as in 802.4 and 802.5) but by a mechanism called distributed queues. For the physical layer of DQDB, this protocol specifies a dual-bus topology to carry data in forward and reverse directions. The forward direction bus carries data while the reverse direction handles queuing and control information. In this illustration, the two unidirectional buses are labelled Bus A and Bus B.
DQDB consists of two unidirectional buses (cables) to which all computers are connected as shown in the figure. Each bus has a head-end, a device that initiates transmission activity. Each bus connects to the stations directly through input and output ports; no drops lines are used. To send data on one bus, a station must use the other bus to make a reservation.
BACK |
