1.6 Token Ring

---

Token ring is a type of LAN that was introduced by IBM in 1985. It had a top speed of 4 Mbps and was developed in response to the commercial availability of Ethernet, which was developed jointly by Digital Equipment Corp. (DEC), Intel, and Xerox. When Ethernet was introduced, IBM did not endorse it, mainly because its equipment would not work in that environment. Later, in 1989, the speed of token ring was boosted to 16 Mbps.

The ring is essentially a closed loop, although various wiring configurations that employ a multistation access unit (MAU) and patch panel may cause it to resemble a star topology (see Figure 1.3). In addition, today’s intelligent wiring hubs and token ring switches can be used to create dedicated pipes between rings and provide switched connectivity between users on different rings.

Figure 1.3: Token ring configured in a star topology through the use of MAUs.

The cable distance of a 4-Mbps token ring is limited to 1,600 feet between stations, while the cable distance of a 16-Mbps token ring is 800 feet between stations. Because each node acts as a repeater in that data packets and the token are regenerated at their original signal strength, token-ring networks are not as limited by distance as are bus-type networks. Like its nearest rival, Ethernet, token-ring networks normally use twisted-pair wiring, shielded or unshielded.

1.6.1 Advantages of Token Ring

The ring topology offers several advantages:

• Since access to the network is not determined by a contention scheme, as is Ethernet, a higher throughput rate is possible in heavily loaded situations, limited only by the slowest element—sender, receiver, or link speed.

• With all messages following the same path, there are no routing problems to contend with. Logical addressing may be accommodated to permit message broadcasting to selected nodes.

• Adding stations is easily accomplished—one connector is unplugged, the new node is inserted, and both nodes are plugged into the network. Other nodes are updated with the new address automatically.

• Control is simple, requiring little in the way of additional hardware or software to implement.

• The cost of network expansion is proportional to the number of nodes.

Another advantage of token ring is that the network can be configured to give high-priority traffic precedence over lower priority traffic. Only if a station has traffic equal to or higher in priority than the priority indicator embedded in the token can it transmit data onto the network.

The token ring in its pure configuration is not without liabilities, however. Failed nodes and links can break the ring, preventing all the other stations from using the network. At extra cost, a dual-ring configuration with redundant hardware and bypass circuitry is effective in isolating faulty nodes from the rest of the network, thereby increasing reliability. Through the use of bypass circuitry, physically adding or deleting stations to the token-ring network is accomplished without breaking the ring. Specific procedures must be used to ensure that the new station is recognized by the others and is granted a proportionate share of network time. The process for obtaining this identity is referred to as neighbor notification. This situation is handled quite efficiently, since each station becomes acquainted with the address of its predecessor and successor on the network upon initialization (power-up) or at periodic intervals thereafter.

1.6.2 Frame Format

The frame size used on 4-Mbps token rings is 4,048 bytes, while the frame size used on 16-Mbps token rings is 16,192 bytes. The IEEE 802.5 standard defines two data formats—tokens and frames (see Figure 1.4). The token, three octets in length, is the means by which the right to access the medium is passed from one station to another. The frame format of token ring differs only slightly from that of Ethernet. The following fields are specified for IEEE 802.5 token-ring frames:

Figure 1.4: Format of IEEE 802.5 token and frame.

• Start delimiter (SD): Indicates the start of the frame.

• Access control (AC): Contains information about the priority of the frame and a need to reserve future tokens, which other stations will grant if they have a lower priority.

• Frame control (FC): Defines the type of frame, either MAC information or information for an end station. If the frame is a MAC frame, all stations on the ring read the address, but only the destination station can read the user data.

• Destination address (DA): Contains the address of the station that is to receive the frame. The frame can be addressed to all stations on the ring.

• Source address (SA): Contains the address of the station that sent the frame.

• Data: Contains the data “payload.” If the frame is a MAC frame, this field may contain additional control information.

• Frame check sequence (FCS): Contains error-checking information to ensure the integrity of the frame to the recipient.

• End delimiter (ED): Indicates the end of the frame.

• Frame status (FS): Provides indications of whether one or more stations on the ring recognized the frame, whether the frame was copied, or whether the destination station is not available.

1.6.3 Operation

A token is circulated around the ring, giving each station in sequence a chance to put information on the network. The station seizes the token, replacing it with an information frame. Only the addressee can claim the message. At the completion of the information transfer, the station reinserts the token on the ring. A token-holding timer regulates the maximum amount of time a station can occupy the network before passing the token to the next station.

A variation of this token-passing scheme allows stations to send data only during-specified time intervals. The ability to determine the time interval between messages is a major advantage over the contention-based access method used by Ethernet. This time-slot approach can support voice transmission and video conferencing, since latency is controllable.

To protect the token ring from potential disaster, one station is typically designated as the control station. This station supervises network operations and does important housecleaning chores, such as reinserting lost tokens, taking extra tokens off the network, and disposing of “lost” packets. To guard against the failure of the control station, every station is equipped with control circuitry so that the first station detecting the failure of the control station assumes responsibility for network supervision.

1.6.4 Dedicated Token Ring

Dedicated token ring (DTR), also known as full-duplex token ring, lets devices directly connected to a token-ring switch send and receive data simultaneously at 16 Mbps, effectively providing each station with 32 Mbps of throughput.

Under the IEEE 802.5r standard for DTR, which defines the requirements for end stations and concentrators that operate in full-duplex mode, all new devices will coexist with existing token-ring equipment and will adhere to the token-passing access protocol. The DTR concentrator consists of C-Ports and a data transfer unit (DTU). The C-Ports provide basic connectivity from the device to token-ring stations, traditional concentrators, or other DTR concentrators. The DTU is the switching fabric that connects the C-Ports within a DTR concentrator. In addition, DTR concentrators can be linked to each other over a LAN or WAN via data transfer services such as asynchronous transfer mode (ATM).

1.6.5 High-Speed Token Ring

With 16-Mbps token-ring connections between switches easily becoming congested at busy times and high-performance servers becoming less able to deliver their full bandwidth potential over a 16-Mbps token-ring connection, the need for a high-speed solution for token ring has become readily apparent in recent years. Other high-speed technologies are already available for inter-switch links and server connections— Fiber Distributed Data Interface (FDDI), Fast Ethernet and ATM—but they are inadequate for the token-ring environment.

In 1997, several token-ring vendors teamed up to address this situation by forming the High Speed Token Ring Alliance (HSTRA). A year later, the alliance issued a specification for high-speed token ring (HSTR), which offers 100 Mbps and preserves the native token-ring architecture. However, to keep costs to a minimum and to shorten its development time, HSTR is based on the IEEE 802.5r standard for dedicated token ring, adapted to run over the same 100-Mbps physical transmission scheme used by dedicated Fast Ethernet. HSTR links can be run in either half-duplex or full-duplex mode, just like dedicated token ring.

HSTR can be mixed with existing 4/16-Mbps switches, hubs, bridges, routers, NICs, and cabling. HSTR was implemented for greater throughput where the enterprise needed it most—at the server and backbone. Upgrading these connections with HSTR only required that an HSTR uplink be plugged into a token-ring switch and that the existing 16-Mbps server network NIC be replaced with a 100-Mbps HSTR NIC. To complete the upgrade, the two devices are connected with appropriate cabling. The 100-Mbps HSTR operates over both Category 5 UTP and IBM Type 1 STP cable, as well as multimode fiber-optic cabling.

It was also possible to connect desktop systems to token-ring switches on dedicated 100-Mbps HSTR connections. Token-ring vendors offer 4/16/100-Mbps adapter cards that enable companies to standardize on a single network adapter and prepare their infrastructure for the eventual move to HSTR. While the HSTR standard does not define an auto-negotiation algorithm, individual vendors have a number of ways to implement the feature while adhering to the standard. With this feature, HSTR products operate at the maximum connection speed, automatically determining whether to transmit at 4 Mbps, 16 Mbps, or 100 Mbps. Corporations could equip stations with auto-negotiating 4/16/100-Mbps NICs, even if there is no immediate need for 100 Mbps throughput to the desktop. When the hub or switch at the other end of the connection is later upgraded to 100 Mbps HSTR, the token ring station can automatically adjust transmission to 100 Mbps.

Since Ethernet packets can be carried over token-ring links, HSTR makes a good backbone medium for the mixed-technology LAN. With support for the maximum token-ring frame size, an HSTR backbone segment is able to handle Ethernet and token-ring frames on the same VLAN connection, which Fast Ethernet would not be able to do without a lot of processing to break down the larger token-ring frames.

Token ring is a stable technology with a proven capacity for handling even today’s high-bit-rate applications. At the same time, network managers can protect their current investments in token ring by understanding application performance and the capacity of the network, and tuning it accordingly. The DTR standard prolongs the useful life of token-ring networks, while meeting the increased bandwidth requirements of emerging applications such as document imaging, desktop videoconferencing, and multimedia. Nevertheless, token ring has been overtaken by Ethernet, in terms of both technology and market share. Not only is Ethernet cheap to implement, it offers a migration path to higher speeds that token-ring standards lack. While Ethernet has reached gigabit-per-second speeds, token ring has not been standardized beyond 100 Mbps. ^{[1 ]}

^{[1 ]}Although some vendors like Cisco Systems offered a form of Gigabit Token Ring, critics claimed they were proprietary products and not authentic token ring.

---