Token Ring is a popular local area network (LAN) technology developed by IBM that still has a large installed base in many shops but has been greatly outpaced in recent years by different forms of Ethernet. Token Ring was standardized in the IEEE 802.5 specifications, which describe the implementation of a token-passing ring network configured as a physical star topology.
In a Token Ring network, stations (computers) are wired in a star formation to a central wiring concentrating unit called a Multistation Access Unit (MAU). This unit concentrates wiring in a star topology but internally forms a logical ring topology over which network traffic can travel. Lobes connect the individual stations to the MAU. The maximum cable length for a lobe is 22.5 meters or 100 meters, depending on the cable type, but you can extend this distance up to 2.4 kilometers using repeaters designed for Token Ring networks. MAUs typically support 8 or 16 connections for attaching lobes.
You can extend a Token Ring network by connecting MAUs to ring-out and ring-in ports to form larger rings that can support larger numbers of stations. Stackable MAUs simplify this process. You can connect up to 33 MAUs to form a network. Many MAUs support being connected by fiber-optic cabling to create networks that span a building or campus. Most MAUs also support in-band management by using Simple Network Management Protocol (SNMP) plus out-of-band management by using a serial interface.
Graphic T-8. Token Ring.
Token Ring networks typically operate at speeds of 4 or 16 Mbps, although speeds of up to 100 Mbps are possible with equipment from some vendors. Token Ring networks come in two types, both of which can operate at 4 or 16 Mbps:
Type 1 is often considered more reliable than Type 3, but the larger installed base of UTP cabling makes Type 3 a viable option for new Token Ring installations. Type 1 configurations support up to 260 stations per ring, while Type 3 can support up to 72 stations per ring.
Token Ring stations pass a single data packet called a token from one computer to the next rather than let each node transmit independently, as in a contention-based network such as Ethernet. Only one token can be on the network at a time, so collisions do not occur in Token Ring networks as they do in Ethernet networks. This process is analogous to sending messages to a group of people by passing a hat.
In order to pass a token in a Token Ring network, each station must know who its neighbors are and must perform a check to make sure that the circuit is unbroken. Messages containing this information are continually sent around the ring. The token circulates so long as this message is received. To generate the required information, the first station online in the ring assumes the role of Active Monitor Station. It creates the token and is responsible for taking action if the token is lost or damaged. The Active Monitor Station sends out an Active Monitor Present frame every seven seconds to the next node down the line. Each node in turn informs its downstream neighbor that it is its Nearest Active Upstream Neighbor. An error-detection process called beaconing occurs if the ring breaks and the token fails to circulate. If the Active Monitor Station fails, another station assumes its role of monitoring the status of the network and generating a new token if the existing one is lost.
If a station wants to transmit data over the network, it waits until the token comes by; if the token has not been claimed by another station, it claims the token and inverts the monitor setting bit to mark it “busy” so that no other station can claim the token for a predefined but variable amount of time. The originating station then removes the last byte from the token (called the delimiter byte), appends data to the token, and appends the delimiter byte to the end to form a frame of variable length (up to 8000 bytes). The token with data circulates around the ring in one direction from station to station. (Each station acts as a repeater to regenerate and forward the token.) When it returns to the originating station, the token and the data are removed and a new token is generated and placed onto the network.
The term “Multistation Access Unit” is sometimes abbreviated as MSAU instead of MAU to distinguish it from “media attachment unit,” a term used in older Ethernet networking technologies.
Distances between MAUs and attached stations are specified as lobe lengths, which refer to round-trip signal paths. Thus, a station with a lobe length of 200 meters actually uses a cable 100 meters long.
STP cabling for Type 1 Token Ring comes in nine types, two of which are common now:
You can get both types of cable in an adapter cable version (terminated at one end with an IBM data connector and at the other end with a DB9 male connector) or a patch panel version (terminated at both ends with data connectors). Use patch panel cables to connect MAUs, and use adapter cables to connect stations to MAUs.
You can also get baluns, which can convert Type 1 IBM cabling to UTP cabling to connect different Token Ring types, and you can get special adapters that allow data connectors to be connected to RJ-45 ports so that you can use installed UTP cabling with Type 1 MAUs.
Most MAUs and NICs are dual-speed and can run at either 4 or 16 Mbps, but not both. However, you can use bridges or routers to connect 4-Mbps Token Ring networks to 16-Mbps Token Ring networks.
The following table provides suggestions for troubleshooting Token Ring network problems.
Mismatched ring speed
Be sure that all connected stations use 4 Mbps or that all use 16 Mbps. Do not mix stations of different speeds.
Stations cannot receive
Check cables and reset the MAU.
Conflicting MAC addresses
Use NIC configuration software to change the MAC address on one of the conflicting computers.
Traffic congestion on the network
Segment the network by using a bridge or a router.