Token Ring

Token Ring is considered a half-duplex network implementation because only one host can transmit at any given time. Token Ring's full-duplex network implementation is known as Dedicated Token Ring (DTR) . In DTR implementations, LAN hosts connect to a DTR concentrator or switch and have all available link bandwidth to use for data transmission and reception .

Token Ring and IEEE 802.5 are two of the three chief examples of token passing networks; the third being FDDI. Token passing networks move a small frame, the token, around the network. Possession of the token by a host gives that host the right to transmit data. If a host receiving the token has nothing to send, it passes the token to the next host downstream in the ring. Each host can hold the token for a maximum period of time, called the Token Holding Time (THT), and the default is 10 ms.

If a host possessing the token has something to send, it seizes the token, alters one bit of the token to turn it into a start-of-frame (SOF), adds the data, and sends the complete frame (SOF and data). This frame is sent to the next host on the ring, known as the downstream neighbor , as illustrated in Figure 5-7.

While the data frame is circling the ring, no token is on the network, unless the ring supports early token release. If the ring does not support early token release, other hosts wanting to transmit must wait. If early token release is supported, a new token can be released when frame transmission is completed. If early token release is not in use, collisions cannot occur in Token Ring network implementations.

The data frame circles the ring until it reaches the intended destination host, which copies the information for further processing. The data frame continues to circle the ring and is removed when it reaches the originating host. The originating host then checks the returning frame, determining whether the frame was seen and copied by the destination.

Unlike Ethernet, Token Ring networks are deterministic , meaning that it is possible to calculate the maximum time that will pass before any host will be capable of transmitting. This deterministic nature makes Token Ring networks ideal for applications in which delay must be predictable, such as SNA-based (mainframe) applications.

FDDI

FDDI LANs were introduced in the mid-1980s as ANSI standard X3T9.5 and operate in a similar fashion to Token Ring LANs. FDDI LANs are a popular choice for LAN backbones because they have the following characteristics:

• Operate at high speeds - Data is transmitted around a FDDI ring at 100 Mbps.

• Reliable - Servers, workstations, or other network devices can be connected to dual rings. After a ring failure, a usable path on the other ring is automatically available.

• Support a large network diameter - Dual fiber- optic rings can have a network diameter of up to 100 kilometers each.

FDDI LANs are similar to Token Ring LANs because both are made up of a series of point-to-point links that connect a host to a host, a host to a concentrator, or a concentrator to a concentrator. FDDI specifies the use of fiber optic cabling for its infrastructure, but copper was later introduced and is supported by the Copper Distributed Data Interface (CDDI) specifications (similar to FDDI).

Figure 5-8 illustrates a FDDI LAN with single-attached stations (SAS), dual-attached stations (DAS), and FDDI concentrators (used to connect multiple stations to the FDDI LAN). To provide external access to a WAN, a router with dual-attached FDDI ports (ports A and B) is attached to the FDDI ring. Separate fibers are used in both transmit and receive directions, providing both an input and output point on each FDDI station.

Dual Homing

In an FDDI LAN environment dual homing is a fault-tolerant technique used for critical network devices, such as mainframes, server farms, or other mission-critical devices. In a dual- homed implementation, the host is attached to two FDDI concentrators, as illustrated in Figure 5-9.

Primary and Secondary Rings

FDDI trunks consist of two rings, known as a dual ring . During normal operation, traffic flows on the primary ring, where the secondary ring is the backup in the event the primary ring fails. In the instance of a primary ring failure, the systems adjacent to the break automatically reconfigure the ring path and create a new path that is a combination of both the primary and secondary rings, as illustrated in Figure 5-10.

The path length around the ring is limited to 200 kilometers. It is this limitation that restricts the LAN circumference of the dual-ring to 100 kilometers. If one of the rings fails, the dual-ring becomes a single large ring, combining the circumference distances of both the primary and secondary rings.