The bit in Token Ring data frames that is set to zero when the frame is initially transmitted and set to one by the destination station when it has seen the frame. The original transmitting station uses the setting of this bit to decide whether or not to make additional transmissions to the destination station. For example, when the A bit on a returning data frame is zero this may indicate the destination station is not inserted into the ring or that it is not functional.

The one monitor on the ring that is currently responsible for maintaining the ring's health. The duties of the active monitor include detecting and correcting error conditions, such as a lost token or a persistently circulating frame. The assignment of the active monitor is dynamically determined with the token claiming process each time the ring recovers from a fault.

A type of MAC frames (management frame) used to indicate the continuing presence of an active monitor on the ring. Only the active monitor issues these frames and it issues one about every 7 seconds. For the active monitor, this frame communicates the station's address to the next station (that is, the frame communites the upstream neighbor address, UNA).

A ring condition in which data frames have been or are being transmitted onto the ring. This state is opposed to the ring's idle state.

Also called a network adapter board or controller. A hardware device that is capable of communicating over a communications protocol medium; the adapter “speaks” the specific protocol (for example, 802.5 Token Ring, Ethernet, FDDI). Each station on a Token Ring must have an adapter through which it sends and receives data. The IRIS Token Ring board is an adapter. Two of the important functions performed by an adapter board are repeater and medium access control sublayer.

The cable that connects an adapter to the network. The connection to the network is usually through a wall faceplate or directly to a trunk coupling unit (for example, a MAU or LAM). Two types of adapter cables are available: shielded twisted pair cable ending in a DB-9 connector for the adapter and unshielded twisted pair cable ending in an RJ-45 connector for the adapter.

See active monitor present frame.

A generic term referring to any intelligent hardware and software system that can be inserted into a Token Ring (through an adapter and a trunk coupling unit). Some examples include printers, networked fax machines, main frames, and stations (such as PCs and IRISes).

A type of MAC frame (management frame) used to indicate a break in the ring. Only standby monitors issue these frames. The presence of this type of frame on a ring indicates two things: a station thinks it has detected a break in the ring, and the beaconing procedure is in progress.

A ring recovery process that starts when a station fails to see a token after a set period of time while token claiming. (The length of this time period is defined by the 802.5 Token Ring Standard.) During the beaconing process, one or more stations transmit beacon frames until the ring is again intact (that is, the break in the ring is fixed). All stations refrain from transmitting data during beaconing. When a beaconing station sees another station's beacon frame, it stops beaconing. When a beaconing station sees its own beacon frame, it starts token claiming.

A ring error condition where the signal cannot complete the loop from a repeater's output port around the ring to the same repeater's input port. Whenever a station fails to see the token or an active monitor present frame for a defined period of time, a break in the ring is assumed and beaconing starts.

A state in which a station is not inserted onto the ring. The station is not listening, repeating, or transmitting when in the bypass state. If a station is attached to a trunk coupling unit (TCU), the TCU may bypass the station if it determines that the station is dysfunctional. This action should cause the station to transition into its bypass state. Other events that may cause a station to transition into bypass state are the discovery that its MAC address is already in use on the ring or reception of a beacon or claim token frame while trying to add itself to the ring.

The act of receiving the token but not repeating it onto the ring. The purpose of capturing the token is to transmit data. On a Token Ring, only the station that has captured the token has permission to transmit data.

The bit in Token Ring data frames that is set to zero when the frame is initially transmitted and set to one by the destination station when it has copied the frame. The original transmitting station uses the setting of this bit to decide whether or not to retransmit the data within that frame. For example, whenever the C bit in a returning frame is still zero, the destination station has not been able to copy the frame, so the data has not been received (even though the frame may have been seen).

As a verb, see capture. As a noun, see claim token frame.

A type of MAC frame (management frame) used to manage the token claiming process. Any standby monitor may use this frame.

See controlled access unit.

A hardware device that acts as a trunk coupling unit and that allows multiple stations to attach to it so that they can participate in the ring without being directly attached to the trunk ring. Multistation access units and lobe attachment modules are examples of concentrators.

Also called network manager. An application-level software module, residing within one station on the ring, for managing ring parameters and statistics. The presence of a configuration report server on a ring is optional. (Silicon Graphics does not currently provide this software module.)

The configuration report server maintains ring configuration parameters and updates other stations on the ring in the event any of the parameters change. As stations insert themselves onto the ring, they request the current parameters from the configuration report server. This server communicates (via MAC frames) with each station's medium access control sublayer. For example, in IRIS Token Ring products, this functionality is provided by the Texas Instrument TMS38016C chip on the adapter board.

An intelligent hardware device that provides diagnostic support for the ring. A CAU is always associated with a lobe attachment module. CAUs reside in a wiring closet.

Token Ring frames that contain user data in the “Information” segment, as illustrated in Figure Gl-1.

Network layer 2 for both the Open Systems Interconnect and the Systems Network Architecture environments. The data link layer is responsible for data transfer across a single physical connection. For Token Ring, the services of this layer are divided into two sublayers: medium access control and logical link control.

In an SNA environment, the interface (language for talking) to the SNA data link layer.

See data link provider interface.

A method of representing a 4-byte (32-bit) Internet address (IP address) in ASCII characters (digits 0-9). Each byte of the address is represented as a decimal number ranging in value from 0 to 255. Bytes are separated by a dot (.), for example, 126.4.71.254. See Internet address.

A term indicating the relative positioning of stations on the ring. Downstream stations are those that see the signal after the station in question.

An error condition in which two stations on a ring use the same physical address (MAC address).

A type of MAC frame (management frame) used by a station during initialization to verify that its physical address (MAC address) is unique on the ring. All stations use this type of frame when they insert themselves onto the ring.

An optional set of rules for governing Token Ring data transmission, frame stripping, and token release. This enhancement is available only for 16 megabit per second Token Rings. The rules allow a transmitting station to regenerate (release) the token before stripping all of its data frames and allow the next station to start transmitting. Early token release makes it possible for more than one data frame to be on the ring at a time, thus increasing the ring's efficiency (or use of bandwidth).

A condition where a high percentage of the theoretical data-carrying capacity of a transmission medium (cable) is used. For example, a 16 megabit per second Token Ring theoretically can carry 16 million bits of information every second. If every second (on the average), the stations on a ring were able to fill the cable with data so that 95% of the 16 million available bits were actually bits of data (as opposed to fill or garbage), this would be very efficient use of bandwidth. If, however, only 20% of the bits carried data, this would be inefficient use. Items that affect how efficient the use of bandwidth is on a network include the following:

A signal pattern placed on the transmission medium (cable) when a station is waiting (for example, when a station that has captured the token is not transmitting data and is waiting for its data frames to return).

The method for organizing information for communication among Token Ring stations and medium access control sublayers. There are three types of Token Ring frames: MAC frames, data frames, and the token.

The amount of time it takes for a signal to make a complete loop around the ring from the output port of one station around to the same station's input port. A ring with a small guaranteed latency is perceived as fast to its human users. The guaranteed latency of a ring increases as the trunk cable lengthens and as the number of stations increases. For example, the guaranteed latency for a 10 meter ring is less than for a 1000 meter ring, and the guaranteed latency of a ring with 5 stations is significantly less than a ring with 200 stations. The ring's data speed also affects its guaranteed latency; the higher the data rate, the smaller the guaranteed latency.

See fill.

A state of the ring where no data frames are being transmitted or repeated. Only the token and MAC frames are circling. This state is opposed to the ring's active state.

The action taken by an attaching device by which it synchronizes its receive clock with the signal on the transmission medium (cable), verifies the uniqueness of its MAC address, and begins repeating the signal.

Also called IP address. A 4-byte (32-bit) number used by the Internet Protocol (IP or TCP/IP) software to identify computers (or more accurately, computers' network connections). As the commonly quoted dictum says: “In the IP world, hosts do not have addresses, network interfaces do.”
One computer (host) can have one or more IP addresses; each physical network connection for a host must have at least one unique IP address.

Internet addresses come in a number of classes; the major classes are A, B, and C. All IP addresses have three parts: class identifier, network identifier, and host identifier. The number of bits used to represent each part depends on the address' class, as described below:.

The class plus network identification parts are commonly referred to as the “network address,” while the class, network, and host identification parts are referred to as the “host address.” For example, the network address for a device with an IP (or host) addresss of 206.2.71.198 is “net 206.2.71.” IP addresses are usually represented in ASCII digits 0 to 9 in dotted decimal notation (for example, 126.13.69.237). Table Gl-2 summarizes the ranges of valid addresses within these three classes.  

In order to ensure global uniqueness, network addresses (or blocks of them) are assigned by the Network Information Center to requesting organizations. The network administrator for each organization allocates the individual addresses (host addresses within the assigned blocks) to specific devices. Local network administrators are responsible for ensuring that two devices at the same site do not use the same address.

See Internet address.

See lobe attachment module.

A measurement of time during which a transmission medium (cable) is not available for use. For example, there is latency associated with a token traveling from one station to the next; the longer the cable between the stations, the higher the latency; the slower the data rate used on that cable, the higher the latency. See also guaranteed latency.

See logical link control sublayer.

The section (consisting of one or more segments) of cabling that connects an adapter to a trunk coupling unit (TCU), as illustrated in Figure Gl-2. For example, a lobe can consist of a simple adapter cable connected directly to a TCU, or a lobe can be a drop (a segment of trunk cable) from a TCU to a wall faceplate connection and an adapter cable running from the wall connection to the adapter.

A hardware device that acts as a concentrator (and trunk coupling unit), allowing one or more adapters to access the main (trunk) ring through it. LAMs may reside in a wiring closet and have lobe cabling that extends out to wall faceplate connections. See also lobe and concentrator.

Local area network protocol functions (services) handled by the upper portion of the OSI data link layer (layer 2), as illustrated in Figure Gl-3. IEEE 802.2 (also known as ISO 8802-2 and upon which Token Ring is based) is an example of a protocol standard for this sublayer. LLC 802.2/8802-2 provides three types of service: type 1 is unacknowledged connectionless, type 2 is connection oriented, and type 3 is acknowledged connectionless.

A ring error condition. The condition exists when no station is transmitting and yet there is no token on the ring. The active monitor is responsible for correcting this error condition by regenerating the token.

See medium access control sublayer.

Also called the physical address. An address that uniquely identifies the medium access control module of a station. In 802.5 environments, this is a 2-byte (16-bit) or 6-byte (48-bit) address. In FDDI environments, this is a 6-byte address. There is one MAC address for each network adapter on the ring. All the adapters on a ring must use the same address format: either the 2-byte format or the 6-byte one. MAC addresses can be universally assigned (UAA), in which case they are guaranteed to be globally unique, or they can be locally assigned (LAA), in which case they are unique only to their particular ring (or site). LAA MAC addresses (both 2-byte and 6-byte formats) are usually assigned and administered by a site's network manager. Blocks of UAA MAC addresses (6-byte format only) are assigned by the Institute of Electrical and Electronics Engineers (IEEE). (The IRIS Token Ring product supports only the 6-byte format for MAC addresses.)

Token Ring frames that contain management information in the “Information” segment, as shown in Figure Gl-4. MAC frames are used to monitor and manage the ring. The frames are generated and interpreted by the medium access control sublayer (MAC) on each network interface board. The information exchanged with these frames is used by each station's monitor module. Table 1-2 describes the types of MAC frames.

See MAC frames.

See multistation access unit.

Local area network protocol functions (services) handled by the lower portion of the OSI data link layer (layer 2), as illustrated in Figure Gl-3. The MAC interfaces to the logical link control sublayer. Each instance of a MAC must have an address that is unique on that ring. A MAC controls and mediates a station's access to the ring as well as assembles frames (on transmission), disassembles them (on reception), and handles physical addressing procedures. ISO 8802-5 Token Ring and IEEE 802.5 Token Ring are examples of a protocol standard for this sublayer. See also MAC address.

A software module within each Token Ring station that maintains and manages the ring and recovers from various error situations. Only one monitor on a ring is active; all other monitors are standby.

A hardware device that acts as a concentrator (and trunk coupling unit), allowing one or more adapters to access the main (trunk) ring through it. MAUs may reside in a wiring closet and have lobe cabling that extends out to wall faceplate connections. See also lobe and concentrator.

The central authority that assigns blocks of Internet Protocol (IP) network addresses to worldwide public and private organizations. The current address for this organization is Government Systems, Inc., Attn: Network Information Center, 14200 Park Meadow Drive, Suite 200, Chantilly, VA 22021 (at telephone 1-800-365-3642).

A generic term referring to any device attached to the ring. All nodes must have a physical layer address (MAC address). However, not all nodes have network layer addresses; user data cannot be sent to nodes that do not have network addresses. A bridge is an example of a node without a network address; a station is a node with a network layer address.

A condition whereby a frame circles the ring endlessly. For example, the station that put the frame onto the ring may fail before it can strip the frame.

The point at which the ring's signal passes into or out of a node. Each adapter has one input port and one output port.

A method for indicating the relative urgency or importance of data. Token Ring provides eight levels of priority and a set of rules for ensuring that higher-priority data be transmitted before lower-level data, as explained in the section “802.5 Mechanism for Prioritizing Data” in Chapter 1.

A sequence of bits (a field) within the access control portion of each token that is used to indicate the ring's current priority level. Stations can transmit data that has the same or a higher priority level than the level indicated in the token's priority field. Data with a lower priority must wait for the ring's priority level to decrease.

To clear the ring of all signals. This is one of the first actions performed by the active monitor upon taking on its role.

Also called release the token. The act of placing the token frame on the transmission medium (cable) whenever this action is not repeating. Only a station in either of the following states is allowed to regenerate the token: a station that has captured the token or a station that is the active monitor. In the former case, the station releases the token when it has finished transmitting data. In the latter case, a token is regenerated when the ring does not currently have a token (for example, a lost token condition or when the active monitor first assumes its duties).

See regenerate the token

A hardware device that repeats back onto the ring each signal it receives (reads) from it. Each attaching device on a Token Ring must be connected to its own repeater; the repeater is usually located on the network adapter board.

(In some texts, but not this one, this term is used to refer to a special node used only for boosting the ring's signal, which enables lengthening the geographical distance covered by a network.)

The act of reading a signal from the transmission medium (cable) and retransmitting it. Every token ring station is required to constantly do this, except when it is in the bypass state or when it is transmitting data. When repeating a frame, a station may alter the settings on some of the bits (for example, the reservation field, the A bit, the C bit). See also repeater.

A sequence of bits (a field) within the access control field of each frame (data, MAC, or token) that is used to indicate a desired priority level. When a station has higher-priority data to transmit, it indicates the desired priority level in the reservation field of any frame that it is repeating. Sometime later (governed by the priority mechanism described in “802.5 Mechanism for Prioritizing Data” in Chapter 1), a token is generated with the requested priority level.

A set of repeaters connected by transmission media (cabling) in a manner such that the signal travels from repeater to repeater in a closed loop. A number of cable layouts can be used to construct a ring. Two examples are illustrated in Figure Gl-5.

A scheme for controlling action in which all the devices in a circle take turns sequentially. In Token Ring, the controlled action is permission to transmit data onto the ring. At any point in time, only one station has the permission (the turn). When the turn-taker finishes transmitting, it passes the turn (token) to the next downstream entity. This system guarantees that only one station at a time is transmitting and that each station is given a turn.

Also referred to as a link. A group of addresses that define a path from one station (the source address) to another station (the destination address). Just as the stepping stones of a garden path provide the sequential steps that move you along the path, each address within a route provides the next step along that route. And, using this same analogy, as stepping stones provide places to stop and transfer your body from one foot to the other, each address for an intermediate station is a location where the packet stops (is picked up) and is transferred to another local area network. A complete route from source to destination may consist of one or numerous addresses. In Token Ring, routes are limited to a maximum of 7 addresses.

Also referred to as intermediate system. A computer, with connections to two or more networks, that performs routing services for those networks.

The process of discovering and assigning a route from a sending computer (the source address) to an intended receiving computer (the destination address). The routing method used for any particular packet/datagram can be either source routing or transparent routing. The exact manner in which routing is done differs from protocol to protocol.

In Token Ring environments, source routing at the medium access control sublayer is a major routing method. Each station maintains the routing information for all the stations with which it wishes to communicate. Each route (or link) consists of one to seven addresses that, when followed sequentially, lead to the destination. The number of routes that a particular station (or adapter) can handle depends on the amount of memory available.

In Ethernet environments, routing is handled by the upper layers of the protocol stack (for example, the Exterior Gateway Protocol, the Routing Information Protocol, the source routing option in the Internet Protocol). For BSD UNIX environments (including IRIX) using the TCP/IP protocol, both source and transparent routing are available. Software modules acting as routers and gateways maintain the routing information and handle transparent routing. There is virtually no limit to the number of addresses that can be specified in a source route; however, there is a practical limit set by available memory.

See standby monitor present frame.

See Systems Network Architecture.

A method of routing whereby the originator (sender) dictates the complete route that must be followed by the message in order to reach its destination. Each router/gateway along the route honors the route specified by the sender instead of following any of its usual or known routes. Some types of source routing (for example, Token Ring) are handled by the medium access control sublayer. Other types (for example, IP) are managed by the network layer. For a different routing method, see transparent routing. See also routing.

A ring condition where no stations are being added or removed. When a ring is stable, the upstream neighbor addresses (UNAs) that stations receive in circulating frames match the UNAs that are stored in their memories.

An act whereby a station raises the ring's priority level by generating a token whose priority field contains a value that is higher than the value that was in the captured token.

A monitor that watches the ring to verify that one active monitor exists and that there is no break in the ring. If an active monitor does not exist, standby monitors take action to remedy the problem by token claiming. If a break in the ring occurs, standby monitors starting beaconing. All of the stations on the ring, except one, are standby monitors.

A MAC frame (management frame) generated by each station whose monitor is in the standby monitor state. This frame is used to communicate the upstream neighbor address (UNA) that is essential for monitoring and maintaining ring integrity.

A system (hardware and software) with a unique MAC address and a network layer address (for example, an IP address or an LU address). Each station is attached to the ring through an adapter.

Remove a data frame from the ring. Each station is responsible for stripping the data frames that it transmits.

A family of protocols developed by IBM that make it possible for SNA-speaking computer systems to exchange data and interoperate. SNA has seven layers, as illustrated in Figure 1-5; each layer handles a specific set of communication functions and provides these services to the adjacent layers. The services defined for each SNA layer do not correspond necessarily to those in the OSI layer at the same relative position.

A control signal that consists of a unique signaling sequence (pattern), shown in Figure Gl-6. At any point in time, a ring can have zero or one token circulating, but no more. The token is captured (not repeated onto the ring) by a station that wants to transmit data. It is regenerated (put onto the ring) when the transmitting station finishes transmitting.

A ring management process whose goal is to make one station into the active monitor for the ring. The absence of an active monitor causes token claiming to start. During the token claiming process, one or more stations transmit claim token frames. All stations refrain from transmitting data during token claiming. When a token claiming station sees its own claim token frame return, and when that frame's upstream neighbor address (UNA) matches the station's previously received (stored) UNA value, the station becomes the active monitor. The event that causes a station to start token claiming is failure to see an active monitor present or token frame for a period of time. (The length of this time period is defined by the 802.5 Token Ring Standard.)

A timer that controls the maximum amount of time a station can use (occupy) the transmission medium (cable). The transmission medium is “occupied” as long as the station holds the token. When this timer expires, the station must place the token back onto the ring. The length for this timer is defined by the 802.5 Token Ring Standard.

A method of routing whereby the originator (sender) of a packet/datagram specifies only the destination address, and a router/gateway on the sender's LAN does the routing to either the destination system or to the next stop along the way to the destination. The sender does not specify, and does not need to know, the complete route. Each router (intermediate system) that receives the packet looks at the destination address, decides which of its known routes is best for reaching the destination, and sends the packet to the next-step along that route, until the destination is reached.

An adjective denoting the main ring or something attached to the main ring (as opposed to an entity indirectly attached to the ring through a trunk coupling unit). Figure Gl-7 illustrates a trunk ring.

A hardware device that enables one or more stations to connect to the trunk cable, as illustrated in Figure Gl-7. The TCU can insert the station onto the ring as well as bypass the connection to a station that is missing or dysfunctional. Examples of TCUs include the following: multistation access units, lobe attachment modules, and other concentrators.

A term describing the relative positioning of stations on the ring. Upstream stations are those that see and repeat the signal before the station in question. A station can become aware of instability in the ring only as a change in its upstream neighbor or the failure to receive a signal from the upstream neighbor, hence the importance of the upstream neighbor address (UNA).

The physical address (MAC address) of the station that transmits/repeats immediately before the station in question. The upstream neighbor's address (UNA) is one of the most important items for monitoring the health of a ring. Each station on a Token Ring keeps track of its UNA. As long as a station continues to receive frames indicating the same UNA, the ring is stable. When the UNA in the current frame does not match the previously received (stored) UNA, the ring configuration is changing. For example, the old upstream neighbor may have failed or may have been removed.

See efficient use of bandwidth.

See concentrator.