Follow us
Technical Glossary
FDDI Overview
Fiber Distributed Data Interface (FDDI) operates at 100 Mbps and specifies a token-passing, dual-ring topology over fiber-optic transmission media.
Although it operates at much faster speeds, FDDI is similar in many ways to IEEE 802.5 Token RingToken_Ring. Both networks use a ring topology, token-passing media access technique, redundant rings and other reliability features. One of the most important characteristics of FDDI is its use of optical fiber as a transmission medium. Fiber offers several advantages over traditional copper wiring, including security (no electrical signals to tap), a higher throughput potential than copper, and it is easier to install.
This section applies to 100 Mbps over fiber-optics only, and has no applicability to the STP/UTP 100 Mbps Copper Distributed Data Interface (CDDI). The information covered in this section follows the ANSI specifications X3T9.5 for Fiber Distributed Data Interface.
Dual-Ring ConnectionsFDDI specifies the use of dual rings, which consists of one primary ring and one secondary ring. Traffic on these rings travels in opposite directions. Each ring typically consists of two or more point-to-point connections between adjacent stations.
Single attachment stations (SAS) attach to one ring (usually via a concentrator to guard against SAS failures/power-downs interrupting the ring), while dual attachment stations (DAS) attach to both rings. The DAS has two ports to connect to the ring, one port to attach to the primary ring, and the other port to attach to the secondary ring.
Fault ToleranceThe dual rings in an FDDI network provide fault tolerance. If a station on a dual ring fails or is powered down, the ring automatically wraps or reverses to form a single contiguous ring (Figure 1). This removes the failed station from the ring, while still allowing the other stations to continue operation. If a cable between two devices fails, the devices wrap the ring within themselves in order to avoid the bad segment.
A second failure could cause the ring to wrap in both directions from the point of
failure, which would segment the ring into two separate rings that could not communicate
with each other. For example, if you have five sequential nodes in a ring, and node 1
fails after node 3 fails, nodes 2 and 4 will each be isolated because no communications
path exists between them. Subsequent failures will then cause additional segmentation. To
avoid this segmentation, optical bypass relaysbypass_relays, or switches, are used.
Optical bypass relays avoid segmentation by eliminating failed stations from the ring (Figure 2). During normal operation, an optical bypass relay allows the light signal to pass directly through itself uninterrupted. When a node with a bypass relay fails, the bypass relay reroutes the signal back onto the ring before it reaches the failed station, so the ring does not have to wrap back on itself and can keep its communications path and direction.
Optical bypass relays have a power penalty which may cause the maximum allowable loss between stations (see "Attenuation Budgets" below) to be exceeded. The following caveats should be noted when installing optical bypass units:
- Bypass relays introduce additional loss in the network, and they do not perform repeater functions of amplifying and restoring the bit stream. Thus, significant signal loss can occur when the next node on the ring is far away.
- By bypassing a station, the new distance between adjacent stations may exceed the maximum allowable value.
- Bypass relays, as with any mechanical device, may introduce less than reliable service
to the network.
Critical devices such as routers or mainframe hosts can use another fault-tolerant technique called dual homing to provide additional redundancy and sustain the ring. In dual-homing situations, the critical device is attached to two concentrators. One pair of concentrator links is declared the active link; the other pair is declared passive. The passive link stays in backup mode until the primary link (or the concentrator to which it is attached) fails. When a failure does occur, the passive link is automatically activated to sustain the ring.
MediaAs implied by its name the media for FDDI is fiber-optic cable. FDDI networks use two two types of fiber-optic cable: single-mode and multimode. Mode refers to the angle at which light rays are reflected and propagated through the optical fiber core, which acts as a waveguide for the light signals. Single-mode fiber has a narrow core (8.7 to 10/125 micron) that allows the light to enter only at a single angle. Multimode fiber has a relatively thick core (62.5/125 micron) that reflects light rays at many angles. However, the downside of multiple propagation paths is that it can cause the signals to spread out in time and limit the rate at which data can be accurately received. Thus, single-mode fiber is capable of higher bandwidth and greater cable run distances that multimode fiber.
The most commonly used media today is 62.5/125 or 50/125. ANSI X3T9.5 recommends the use of multimode 62.5/125 micron fiber; however 50/125 and 85/125 are also acceptable. Single-mode, or mono-mode (as dubbed in Europe), was nearing ratification at 8/125 and 10/125 micron sizes. Although many vendors claim to support single-mode fiber, they may not be in compliance with ANSI. The dual ring topology requires that at least four fiber pairs be installed between stations, that is, two-pairs for transmit and two-pairs for receive.
ConnectorThe Physical Media Dependent (PMD) sub-layer defines the method for physically connecting a multimode fiber-optic cable to an FDDI node. The Media Interface Connector (MIC) is used to make this connection. The MIC properly aligns the fiber with the transmit/receive optics in the node. The MIC consists of a keyed plug and keyed receptacle. Keying insures that the plug is installed correctly. If incorrectly installed, a ring failure is likely to occur.
ST type connectors are also acceptable for multimode fiber connections. FDDI nodes designed for ST-type rather than MIC connections are available from different vendors; however care must be taken to ensure that the transmit and receive fibers do not get crossed, since ST connectors are not keyed and are individually attached to each fiber strand (one pair per ring).
Port TypeX3T9.5 specifies connection rules to ensure against the construction of illegal topologies. In the FDDI topology there are four port types. These are described below:
| Connects to the incoming primary ring and the outgoing secondary ring of the FDDI dual ring. This port is part of the dual attachment station (DAS) or a dual attachment concentrator (DAC). | |
| Type B | Connects to the outgoing primary ring and the incoming secondary ring of the FDDI dual ring. This port is part of a DAS or a DAC and is also used to connect a DAS to a DAC |
| Connects a concentrator to a single attachment station (SAS), dual attachment station (DAS) or another concentrator. This port is only implemented in a concentrator (DAC or single attachment SAC). | |
| Type S | Connects a SAS or a SAC to a concentrator (DAC, SAC). |
Distances
A total fiber length of 100 km (60 miles) is specified, with a maximum length of 2 km
(1.24 miles) between stations.
Total Stations
FDDI specifies a total of 500 nodes maximum over the distances cited above.
Attenuation Budgets FDDI PMD standard specifies a power budget of 11.0dB between stations (including cable, splices, connectors and relays), and a maximum cable attenuation of 1.5dB/kilometer for 1300 nanometer fiber (1300 nm is the typical wavelength for 62.5/125 multimode fiber). The maximum cable length is defined as the maximum distance possible without violation of the 11.0 dB attenuation allowance which includes connector loss. This distance is around 2km.
