This chapter introduces the basic concepts of the FDDI protocol. After reading this chapter, you will know how FDDI works and be familiar with the most common FDDI terms.
Fiber Distributed Data Interface (FDDI) is a local area network (LAN) communications protocol that is based on a basic token ring architecture. It is fast, reliable, and manageable. It is emerging as the standard alternative to slower protocols like Ethernet and 802.5 token ring. Table 1-1 compares FDDI with Ethernet (the built-in communications medium offered on Silicon Graphics workstations and servers) and token ring 802.5.
4 or 16 Mbps
Maximum packet size
4500 bytes for 4
18,000 for 16
Typical maximum length of LAN cable
< 2.5 kilometers
< 42 kilometers
(200 km wrapped)
Typical max. length between nodes
Maximum number of
FDDI is an international standard. It has been approved and accepted by the two major standards committees: American National Standards Institute (ANSI) and International Standards Organization (ISO).
The FDDI components of FDDIXPress and the accompanying FDDI board conform to the ANSI and ISO FDDI standards. The specific FDDI components (and the ANSI and ISO standards on which they are based) are listed below:
Figure 1-1 shows how the FDDI components correspond to ISO's seven-layer Open Systems Interconnection (OSI) reference model.
The OSI model defines a hierarchical structure for organizing the different functions (services) of telecommunications systems. In theory, each layer is completely independent, so changes to one layer have no effect on other layers. Standard interfaces are defined for communication between the adjacent layers. As Figure 1-1 shows, the FDDI standard occupies the two lowest layers—the entire physical layer and a portion of the data link layer—just as Ethernet and token ring do.
The physical layer defines the electrical, mechanical, and logical characteristics for transmitting bits across the physical medium. Examples of physical media include twisted pair, coaxial, and fiber optic cable. Dual ring FDDI specifies fiber optic cable as the physical medium.
The data link layer specifies the way a node (for example, the FDDIXPress board) accesses the underlying physical medium and how it formats data for transmission. FDDI specifies formatting data into frames, using a special set of symbols and following a special set of rules. The MAC sublayer within the data link layer specifies the physical address (MAC address) used for uniquely identifying FDDI nodes.
Functionally, FDDI is similar to the 802.5 token ring and Ethernet standards, as summarized below:
Like Ethernet and 802.5 token ring, FDDI uses the interface to the logical link control (LLC) sublayer of the data link layer, so switching from Ethernet to FDDI does not affect the higher layers. Layer 3 and 4 software (for example, TCP-UDP/IP) works over FDDI just as it does over Ethernet or token ring.
Like Ethernet and 802.5 token ring, FDDI uses frames to deliver data between stations.
Like 802.5 token ring (but unlike Ethernet), FDDI prevents collisions on its physical medium (cable) by passing a token; at any specific instant, only the station with the token may transmit onto the ring.
The subsections that follow describe each of the FDDI components. Figure 1-2 illustrates one possible configuration of these FDDI components.
The physical layer medium dependent protocol (PMD) defines the lowest FDDI protocol; it occupies the lower sublayer of the physical layer. PMD specifies the requirements for the cable (for example, fiber optic), the transmitter and receiver, the media interface connectors (MIC), and the optional optical bypass switch. PMD functionality is contained within a chip on the FDDI board.
The physical layer protocol (PHY) defines the upper sublayer of the physical layer. It establishes the connection between the PMD and MAC. In addition, the PHY provides encoding and decoding of data and control symbols. The PHY synchronizes incoming and outgoing code-bit clocks. This functionality is contained inside a chip on the FDDI board.
The media access control protocol (MAC) schedules and performs data transfer on the FDDI cable. The MAC is the FDDI component that contains the FDDI connection's identity, commonly referred to as a MAC address.
When a MAC begins to receive a block of information (a frame) from the FDDI cable, it checks the destination address field of the frame to see if the address is one of its own addresses. If the address matches one of its own addresses, the MAC simultaneously repeats the frame onto the physical medium and copies the frame into its local memory.
While repeating the frame, the MAC modifies the frame's status to indicate that the frame has been seen and received. The modified frame continues along the ring until it reaches the original transmitting station, which interprets the modified frame as an acknowledgment. This functionality is handled by a chip on the FDDI board.
The station management protocol (SMT) monitors and controls all FDDI activity on its station. SMT manages processes in the various FDDI layers (PMD, PHY, and MAC) at the station level and ensures the correct operation of the station on the ring. (See “FDDI Ring” for a description of the FDDI ring.) SMT's responsibilities include overseeing station insertion and removal from the ring, initializing the station to conform with the current ring status, and identifying, isolating, and recovering from faults on the ring.
An FDDIXPress station's SMT functionality is distributed. Some of it is contained within a software module that includes the SMT daemon (smtd) and a special database file called the management information base (MIB); some functionality is located within chips on the FDDI board.
The MIB resides in the local memory on each FDDI station. This database maintains statistical and operational information used to manage the ring.
Control within an FDDI ring is distributed among the SMT entities of all the stations on that ring; control is not handled by a master station. SMT entities communicate with each other to manage the administration of addressing, allocation of network bandwidth, and configuration and control of the ring. Some of these SMT parameters are site-configurable. For FDDIXPress, the SMT configuration file is /etc/fddi/smtd.conf.
(1M) man page.
An FDDI ring is a length of cable laid out in a closed loop. Current standards require that the ring cable be fiber optic cable. An optical signal (light) passes through the cable (around the ring) and returns to its point of origin. Whenever a station is connected to the ring, it is physically inserted into the ring so that the optical signal passes through the station (illustrated in Figure 1-3). Stations on the ring are referred to as upstream or downstream in relation to each other. The downstream neighbor station is the first station to see a transmitting station's transmission. In Figure 1-3, station A is station C's downstream neighbor and station B's upstream neighbor.
The FDDI dual ring (or trunk ring) has two separate loops (rings). One ring is called the primary ring and the other is the secondary ring, as illustrated in Figure 1-4. Most sites use the secondary ring as a backup ring. The light signal within each loop of a dual ring travels in the opposite direction from the signal in the other ring; in FDDI jargon this is referred to as counter-rotating. Because the signal travels in different directions, upstream and downstream neighbors are opposite on each ring. In Figure 1-5, where station 2 is station 1's downstream neighbor on the primary ring, station 2 is the upstream neighbor on the secondary ring.
The cabling for FDDI is available in a number of forms. Multimode (62.5 micron) fiber optic cable was the first transmission medium (cable) defined for FDDI. Recently, the use of single-mode (50-micron) fiber optic cable was approved. Copper cable has also been approved, for use only between concentrators and stations.
In addition to the FDDI components, the FDDI standard defines the types of devices that can be connected to the ring. These devices include (but are not limited to) the following:
When connected to the dual ring, each port connects to both the primary ring and the secondary ring (as shown in Figure 1-5). This dual connection is known in FDDI jargon as “connecting to the dual ring.” The station's SMT ensures that the station can continue to transmit and receive data even when the primary ring experiences a break. (A break in the ring occurs when the signal cannot make a complete trip around the ring; this can be caused by a station failing or by a faulty cable.)
When connected to a concentrator, the two ports can each be connected to one of the concentrator's M ports. A DAS station can behave as a single attach station (SAS) if configured to do so, in which case only one of its ports is connected to the concentrator and the other port is not used.
|Note: As illustrated in Figure 1-5, for DAS connections to the dual ring, port A must always be connected to port B of the downstream station, while port B connects to port A of the upstream station.|
A single attach station (SAS) has a single slave (S) port that attaches to the ring through a master (M) port on a concentrator. The concentrator routes the signal from the functioning ring through every SAS connected to that concentrator.
A concentrator allows many single-attachment FDDI devices to obtain their connection to the FDDI ring through one device—the concentrator. Concentrators have one or more master ports (M), each of which accepts a connection from one single-attachment device.
The FDDI standard defines two types of concentrators: dual-attachment and single-attachment. A dual attach concentrator (DAC) has two ports (A and B), each of which connects to both the primary and secondary rings, just like the DAS. A single attach concentrator (SAC) connects to an FDDI ring through another concentrator, in the same manner as an SAS. Figure 1-6 illustrates the use of concentrators on an FDDI ring.
The FDDI local area network consists of two or more stations or nodes connected serially by fiber optic cables to form a closed loop, the ring. Each FDDI local area network has two rings: a primary ring and a secondary ring. Figure 1-6 and Figure 1-10 show common FDDI ring configurations. The secondary ring is usually configured as a backup ring.
An optical signal (light), encoded to represent data, is beamed into the cable by a transmitting station. The signal travels through the cable and is read by each station on the ring, until it returns to the original sender. As long as the signal can make a complete trip around the loop, the ring is operational. When a break or fault occurs in the ring, the signal cannot complete the loop. Situations that break the ring include, among other things, a missing or damaged cable, a loose connection, and a dysfunctional station.
The optical signal travels in opposite directions in each ring. This design makes closure of a broken primary ring feasible. When the SMT module within a station notices that the primary ring is broken, it connects the secondary ring to the primary one to complete the loop. This action bypasses (cuts out) the faulty section, as illustrated in Figure 1-7. In FDDI jargon, fixing a broken primary ring in this manner is called “wrapping the ring.” The original two rings are joined to form a single loop (ring). Notice that the ring must wrap in two locations to complete the loop. In this condition, transmission proceeds without interruption for all the stations on the functioning portion of the ring.
When a ring wraps, two stations change their internal optical signal paths. Instead of the signal passing through both port A and port B (as illustrated in Figure 1-5), it is received and transmitted through a single port (either A or B). Figure 1-8 illustrates the altered optical signal paths. The two stations that make this change are located at the ends of the functional portion of the primary ring.
If more than one fault occurs on the FDDI ring, the ring may become fragmented, as shown in Figure 1-9. In this condition, communication continues among the stations within each fragment, but communication is not possible with stations located on a different fragment.
FDDI management tools such as smtstat and smtring (or the graphical product, FDDIVisualyzer) can be used to identify problems with the ring.
FDDI defines an optional device that allows a DAS to become dysfunctional without wrapping the ring. This device is called an optical bypass switch (OBS). The optical bypass switch is connected between a station's two ports and the dual ring.
Without an optical bypass switch, when a DAS becomes dysfunctional, the signal going around the ring cannot continue past the dysfunctional station; stations downstream from this station do not receive any signal. The ring is broken, which causes an automatic wrap.
When an optical bypass switch is present in this situation, it maintains an intact loop by simply routing the signal through the switch, bypassing the dysfunctional station as if it were not attached to the ring. The SMT modules of neighboring stations will notice that they have acquired different neighbors, but they will continue to communicate without the disruption caused by a wrapped ring.
A station on a ring gains access to transmit information onto that ring by capturing the ring's token. Only one token is allowed on each ring. Various controls are built into FDDI to limit or specify the length of time the token can be held. Once a station captures the token, it can transmit data onto the network. When the station finishes transmitting, or its time expires, it places the token back onto the ring, thus allowing the next station the opportunity to capture it. When a station does not have anything to transmit, it does not capture the token.
Once a frame is transmitted onto the ring, it moves around the ring in the following manner: Each station reads the frame and transmits it back onto the ring. If a station makes a local copy of the frame, it indicates this action by altering various bits in the copy that it retransmits onto the ring. As frames pass around the ring, the transmitting station recognizes the return of its own data and determines if reception has been successful and error free by checking the changed bits in the frame. Each station is responsible for removing (stripping) all the data that it placed on the ring.
You can use FDDI as a standalone network, or you can incorporate it into an existing internetwork. When incorporating FDDI with an existing network, it is standard practice to use FDDI as the backbone and the slower networks (Ethernet or token ring) as subnetworks. This involves using a router (for example, an FDDI-to-Ethernet router) that is connected to both the non-FDDI network and the FDDI ring. The router allows information (packets) to flow between the two networks even though they use different protocols. Figure 1-10 shows FDDI with an Ethernet network; the ring illustrated has five dual-attachment nodes, one of which is a concentrator. A Silicon Graphics workstation or server that has two network interfaces automatically and by default performs as a router.