Introduction to Networking: and Data Communications
PrevNext

Chapter 40. Token Ring

STOP - You are now leaving Ethernet IEEE 802.3

"Please fasten your seat belts and place your trays in the fully upright position...."

Token Ring is a token-passing bus arbitration topology for the Physical and Data Link Layers. It is a logical ring and a physical star topology.

Token Ring uses a token-passing scheme for bus arbitration. A special packet, called a token, is passed around the ring . When a node requires access to the ring, the node claims the token, and then passes its information packet around the ring. All nodes read the destination address: if it is not addressed for them, the information packet is then passed to the next node. When the destination node reads the packet, it marks it as read, and passes it on to the next node. When the information packet completely circulates, the ring and arrives back at the source node. The source node then releases the token back on to the ring.

Token Rings are not usually drawn as the above drawing illustrates (i.e. a separate line between each node). They are usually represented as separate paths that exist between nodes, and are drawn as in the figure to the right.

IBM Token Ring

Token Ring was originally developed by IBM for their PC LAN networks. It started out in 1969 as the Newhall Network, named after the originator of the token ring concept. IBM's Token Ring is the basis for the IEEE 802.5 standard Token Ring. They are very similar, and have minor differences which we will cover.

IEEE 802.4 Token Bus

An industrial version of Token Ring is standardized under IEEE 802.4 Token Bus. It is used in manufacturing process equipment for plant operation, as well as in automobile plants for computerized assembly. Token Bus uses both a Logical Ring and a Physical Bus topology.

IEEE 802.5 Token Ring

IEEE 802.5 Token Ring standard is based on the IBM Token Ring network. Token Ring has been used mainly in large corporations, and was considered in the past to be the only way to handle data communications in large networks (1000+ nodes).

Token Ring equipment is more expensive than Ethernet, and is one of the reasons that Ethernet is more popular. The other reason is that Token Ring is a much more complex bus arbitration method than CSMA/CD, and few network personnel understand the full capabilities of Token Ring.

IEEE 802.5 Bus Arbitration

Token Ring is a token-passing bus arbitration. A token is circulated on the ring. If a node on the ring needs to access the ring (transfer information), it claims the token.

The token is a special packet that is circulated around the ring. It is read from one node, then passed to the next node, until it arrives at a node that needs to access the ring (transfer information/data). When a node receives the token, the node is allowed to send out its information packet.

Example: A token is circulating in the ring. Node B needs to send some data to Node G. Node B waits for the token to arrive (there is only one token allowed on the ring). When it receives the token, it can then send out its information packet. Node G is the destination address in this example.

Node C receives the packet, reads the destination address, and passes it on to the next node. Node D, E & F do likewise.

When the packet arrives at node G, node G reads the destination address and it reads the information. Node G marks the information packet as read, and then passes it on.

Note: the Source and Destination addresses remain unchanged after passing through Node G. Node B is still the Source address and Node G is still the Destination address.

The packet continues around the ring until it reaches the source address Node B. Node B checks to make sure that the packet has been read (this indicates that Node G is actually present). The information packet is erased and Node B releases the token onto the ring.

Information marked READ is passed through the ring back to the Source (Node B).

The information packet is called the "Token Frame." The token itself is just called the "Token" (sometimes referred to as the "Free Token"). This can be confusing. Remember, when we talk about a frame, we are talking about data/information. When talking about a token, we are talking about bus arbitration and permission to use the bus.

/ 16 Mbps Transfer Rate

The transfer rate for Token Ring is 4 Mbps (for older systems), or 16 Mbps for newer systems (1990 and newer). There are several products in development (and some are now available) that will increase Token Ring's transfer rate. These protocols accomplish faster rates by using Switching Hubs, and even faster transfer rates by going over existing cabling.

NOTE: 16 Mbps NIC cards will operate at both 16 and 4 Mbps speeds.
4 Mbps NIC cards will only operate at 4 Mbps.

To identify the speed of an unknown card, examine the integrated circuits on the card. There is only 1 chipset that implements IEEE 802.5's 4 Mbps standard for Token Ring. It was developed jointly by Texas Instruments and IBM. It is a 5 chip set and consists of the following:

4 Mbps Token Ring NICs are usually full-length expansion cards.

16 Mbps NICs have typically 1 large IC (with 132 pins), and several small ones. They are typically 1/2 length cards. The IC number is TMS380C16 for the Texas Instrument version, TROPIC for the IBM version, or DP8025 for the National version.

IEEE 802.5 Topology

Token Ring is a Logical Ring / Physical Star topology. So far, we've been only discussing the logical portion. Nodes on the network are physically connected--via their NICs--to a central concentrator or hub. The concentrator is called a MAU (or MSAU): both stand for Multi Station Access Unit. To avoid confusion with Ethernet MAUs, we will refer to a Token Ring hub as a MSAU (pronounced "M sow"), or as a concentrator.

MSAUs

A Token Ring MSAU has connections to connect to the nodes, and it also has special connections called Ring In and Ring Out (to connect to other MSAUs).

The Ring In connector is abbreviated RI and the Ring Out connector is abbreviated RO. The nodes (PCs) would be attached to connectors 1 to 8 for this 8 node MSAU.

The MSAU logical connection would be drawn as indicated below:

The connection from the Node to the MSAU is called the Lobe. The connection to the ring is via the Ring In and Ring Out connectors.

MSAUs are passive devices (there isn't any built-in "intelligence.") MSAUs come in these 2 flavors:

Token Ring connectors

The wiring between the NIC card and MSAU consists of the 2 pairs of wires shown below:

There are 4 types of connectors that are used with Token Ring.

Signal LeadHermaphroditicRJ45RJ11DB9Tx+Orange (O)329Rx+Red
(R)431Rx-Green (G)546Tx-Black (B)655

UTP wiring pin outs:

Note: The receive pair (Rx) is the center pair of wires; the transmit pair (Tx) is the outside pair.

MSAU Relay

When a Token Ring NIC is first turned on, it goes through a process called Ring Insertion. It checks the Lobe to see if the wiring is okay, and then applies a DC voltage on the Transmit pair of wires. The DC voltage is often called phantom power.

This voltage energizes a relay in the MSAU, and attaches the Lobe to the ring. If you disconnect a cable at the MSAU, the relay will de-energize, and automatically disconnect the lobe from the ring. You can actually hear the relays clicking in and out.

Ring In / Ring Out

There are 2 connectors on a MSAU that are called Ring In (RI) and Ring Out (RO). These are used for connecting MSAUs together. Two pairs of wires are run between MSAUs to connect them together: one pair is used for the Main Ring and one is used for the Backup Ring.

The following figure indicates the Main Ring and the Backup Ring. Notice that the Backup Ring runs in parallel with the Main Ring, and is not normally used. Also, notice that the direction of data flow on the Backup Ring is opposite to that of the Main Ring.

Wrapping

If the Main Ring fails--due to cable faults or MSAU problems-- the Main Ring can then be wrapped to the Backup Ring. Wrapping is a term that is used to indicate that the Backup Ring is used in addition to the Main Ring.

The Backup Ring is connected to the Main Ring. The Main Ring (or a portion of the Main Ring) is still being used. Wrapping is only associated with the Ring In and Ring Out connectors (on the MSAUs).

Main Ring wrapped to Backup ring

This can be done in either of these 3 ways:

Physical Star / Logical Ring

With an understanding of how an MSAU works, it is easier to see how we get a Logical Ring for Token Ring. The Physical Star results from the Lobe cabling fanning out to the Nodes.

IEEE 802.5 and the OSI Model

Token Ring Cabling

There are 2 basic types of Token Ring cabling.

Shielded Twisted Pair

STP--or Shielded Twisted Pair--is balanced shielded twisted pair cable (150 +/-15 ohms impedance). It is typically used with the Hermaphroditic connectors. It is referred to as IBM Type 1, 1A, 2 or 6 cabling. STP is the most expensive cabling to use; the cable and the connectors are both expensive.

Max Lobe Distance# Stations per ringConcentrator4 Mbps1000 ft/ 305 m
250Passive16 Mbps550 ft/ 168 m250Passive4/16 Mbps1000 ft/ 305 m250Active

Unshielded Twisted Pair - Type 3

UTP--or Unshielded Twisted Pair--is used with phone style connectors (RJ11 or RJ45). It is 100 +/-15 ohms impedance, typically 22 to 24 AWG wire. It is categorized into the following categories:

Max Lobe Distance# Stations per ringConcentrator4 Mbps328 ft/ 100 m
72/54Passive (old/new)16 Mbps328 ft/ 100 m250Passive4 Mbps328 ft/ 100 m54Active16
Mbps328 ft/ 100 m250Active

IBM Cabling System

Type 1

Two shielded, solid wire, twisted pairs, 22 AWG. They are available for plenum or non plenum interior use (and underground or aerial exterior use). Use of Type 1 permits transmission at 16 Mbps, and a maximum of 260 stations on the network.

Note: Plenum is heating ducts and air returns. To be qualified for plenum installation, it must meet certain standards for both hazardous fume release and temperature ratings.

Type 2

Two shielded, solid wire twisted pairs, 22 AWG, plus four twisted pairs of solid 26 AWG wires added between the shield and the insulating cable sheath. Type 2 supports 16 Mbps transmission.

Type 3 (RJ11 and RJ45)

Unshielded, telephone grade (22 or 24 AWG) twisted pairs, typically found inside of a building. They are basically equivalent to Cat 5 cable. See the previous section on Unshielded Twisted Pair cabling for more details.

Type 5

100/50 micron fiber-optic cable, is used to connect distant MSAUs with fiber optic repeaters.

Type 6

Patch cables consisting of data-grade, stranded, shielded twisted pairs, 26 AWG. The distant limits are 66% of Type 1 cable.

Type 8

Under carpet cable, data-grade twisted pair cable, 26 AWG. The distance limits are 50 percent of Type 1.

Type 9

Shielded twisted pair, 26 AWG approved for plenum installations. The distance limits are 66% of Type 1 cable.

Ring Insertion

When any node or host wishes to attach to the ring, it initiates the Ring Insertion process. The Ring Insertion process has these 5 phases:

Phase 0 Lobe Media Check

The Lobe Media Check is performed by the NIC and it verifies the Lobe cable (by looping the station transmit signal to the station receiver at the MSAU). A Lobe Media test MAC frame is then issued. The relay in the MSAU is not energized at this time. A special packet is sent from the NIC to the de-energized MSAU lobe relay. This packet loops back from the MSAU, and returns to the NIC. The integrity of the wiring (that makes up the lobe) should be checked.

The NIC applies Phantom Power (DC voltage) on the Transmit pair to activate the relay at the MSAU port. The NIC is now physically connected to the Ring.

Phase 1: Monitor Check

The ring station waits for an Active Monitor Present frame, Standby Monitor Present frame, or Ring Purge MAC frame. If the ring station does not receive one of these frames before the T(attach) timer runs out, then the ring station initiates Token Claiming to re-elect an Active Monitor.

Phase 2: Address Verification

The station verifies that its MAC address is unique with the Ring. It sends a Duplicate Address Test MAC frame onto the ring. The Duplicate Address Test frame has the source and destination address set to its own MAC address. If the frame returns marked read, then the station knows that there is another node with an identical address.

Phase 3: neighbor Notification

The station learns its Nearest Active Upstream neighbor (NAUN) address, and informs its downstream neighbor of its own address (through the neighbor Notification process).

Phase 4: Request Initialization

The workstation sends the Request Initialization MAC frame to the Ring Parameter Server (RPS). The RPS then responds with an Initialize Ring Station MAC frame (containing the station's parameters), such as the local ring number, ring parameter timer values, and so on. If no RPS is available, the station will then insert with its default parameters.

When Phase 4 is complete, the station is physically and logically attached to the ring.

CAUs & LAMs

Smart concentrators (or Hubs) are called CAUs (pronounced "cows") in a Token Ring. CAU stands for Control Access Unit. It has a CPU built-in, and the smarts to control and determine when a Node is operating incorrectly. It can determine if the RI or RO main ring is operating properly. CAUs can make decisions on disconnecting nodes, or wrapping the Main Ring to the Backup Ring. They can also be controlled and programmed from a remote station - SNMP compliant (Simple Network Management Protocol). Nodes can be remotely disconnected from the Ring (CAUs controls LAMs).

A CAU can control up to 4 LAMs (pronounced lambs). LAM stands for Lobe Access Module and LAMs have the Lobe connections. The CAU is connected to the LAMs by a Power Connection and a Data Connection. A LAM has 20 lobe connections (A LAM is an active concentrator).

Active Concentrators

An active concentrator re-times and regenerates the data signal: it does the job of a repeater. Since it re-times and regenerates the data signal it is not used in Ring Length calculations.

Ring Calculations

Maximum Ring Length

The ring length of a Token Ring network is based on the length of the cable used in concentrator-to-concentrator connections, and in the longest concentrator-to-node connection. It is based on Type 1 cabling.

Ring SpeedMaximum Ring Length4
Mbps1200 ft/ 360 m16 Mbps550 ft/ 168 m

Ring Length Calculations

The following ring has Type 1 cabling, 4 passive concentrators, and the lobe with the Maximum Lobe Length (MLL) indicated. The cable length for this ring is calculated by adding all the cable lengths-- between the passive concentrator's Ring In and Ring Out connectors-- together with the Maximum Lobe Length (MLL).

Total Cable Length = MML + RI/RO cable lengths

Total Cable Length = 40 ft + 6 ft + 75 ft + 230 ft + 80 ft = 431 ft

The passive concentrators (MSAUs) will also have an effect on the ring length. Each passive concentrators will appear as 25 ft of Type 1 cable. The Ring Length has to be adjusted for the presence of each of the MSAUs (see below):

Ring Length = Cable Length + (number of MSAUs x 25 ft)

Ring Length = 431 ft + 4 x 25 ft = 531 ft

If you check the maximum ring length parameters that were mentioned earlier, you will see that this ring would function within the specifications for both a 4 Mbps and a 16 Mbps Token Ring.

Mixing Cable Types and Ring Length

The Ring Length is always calculated based on Type 1 cable. All other types of cable used in the network are converted to Type1 cable first, before determining the Ring Length. The conversion factors for other cable types is indicated in the following table:

Cable TypeConversion FactorType 1, 1A or 2 STP1.0(this is the reference)
Type 6 STP1.3Cat 5 UTP1.7Cat 3 UTP3.0

For example, in the following Token Ring, there is a mixture of cable types. The first step is to convert the cable lengths to their equivalent Type 1 cable length.

CableTypeLengthConversion FactorType 1
equivalent lengthACat 5 UTP80 ft1.7136.0 ftBType 6 STP12 ft1.315.6 ftCCat 5
UTP65 ft1.7110.5 ftDCat 3 UTP127 ft3.0381.0 ftEType 2
STP185 ft1.0185.0 ftTotal Cable Length =828.1 ft

Ring Length = Total Cable Length + (number of MSAUs x 25 ft)

Ring Length = 828.1 ft + (4 x 25 ft) = 928.1 ft

If you check the maximum ring length parameters, you will see that this ring would function within the specifications for a 4 Mbps (but not for a 16 Mbps Token Ring).

Active Concentrators and Ring Length

Active concentrators--and the cables connecting the Ring In/Ring Out connectors-- are not counted in determining the Ring Length. Since it re-times and regenerates the data signal, it is not used in Ring Length calculations.

In the following 16 Mbps Token Ring, all cable types are Type 1. If the active concentrator is ignored, a quick addition of cable lengths will indicate that the ring length is over the maximum allowable ring length for 16 Mbps Token Ring. In fact, the active concentrator allows ring lengths over the maximum allowed ring length by discounting the cables attached to it.

Total Cable Length = MML + RI/RO cable lengths

Total Cable Length = 75 ft + 12 ft + 210 ft = 297 ft

The 165 ft and 95 ft cables are not counted because they are attached to the active concentrator's Ring In/Ring Out connectors.

Ring Length = Cable Length + (number of MSAUs x 25 ft)

Ring Length = 297 ft + 3 x 25 ft = 372 ft

The active concentrator is not counted in the calculation of Ring Length, so only 3 MSAUs are used. The ring length appears to be 372 ft (which is within the limits of 16 Mbps Token Ring: 550 ft). In fact, the use of an active concentrator allows this 16 Mbps to operate with a total cable length greater than the allowed ring length for 16 Mbps Type 1 cabling.

Token Ring Monitors and Servers

Active Monitor (AM)

The Active Monitor (AM) is the active node with the highest address: it wins active monitor status by a token claiming process that takes place between all active nodes. Any node can become the AM and all other nodes become Standby Monitors (in case the Active Monitor fails or turns off).

The duties of the Active Monitor are as follows:

Maintaining the Master Clock

The Active Monitor maintains the ring's master clock. The master clock controls timing, and ensures that all other clocks on the ring are synchronized. The AM "beats the drum" for the other nodes to follow.

Ensuring Proper Ring Delay

It inserts a latency buffer (delay) to guarantee a minimum ring length. The delay is 24 bits long for 4 Mbps Token Ring, and 32 bits long for 16 Mbps Token Ring.

Initiating neighbor Notification

The Active Monitor periodically broadcasts the Active Monitor Present MAC frame to all ring stations on its ring, allowing each to acquire the address of its Nearest Active Upstream neighbor (NAUN). The NAUN address is used during error isolation to determine if there is a failing component in a given ring station's fault domain (next node over).

The nodes on the ring are aware of the MAC addresses of their Nearest Active Upstream Neighbor. Notice the word "active": the neighbor must be connected to the ring.

Monitoring Neighbor Notification

At any time during the neighbor notification cycle, certain events could happen that could affect the neighbor Notification process. The Active Monitor checks for these conditions and takes appropriate action as follows:

1. The Active Monitor's neighbor Notification Timer runs out. The ring is taking too long to complete the neighbor Notification process. The Active Monitor restarts the neighbor Notification process, and reports a neighbor Notification Incomplete MAC frame to the Ring Error Monitor (REM - just another node on the ring that has the job of monitoring ring errors).

2. The Active Monitor Present Frame takes too long to circle the Ring. The AM initiates token claiming so it can retransmit an Active Monitor Present frame.

3. If a Standby Monitor Present Frame is received after neighbor Notification is complete, the neighbor Notification is ignored and restarted. Another station has just connected to the ring and by inserting into the ring (has changed the NAUN order).

4. If another Active Monitor Present Frame is received with a source address different from its own. This means that there is another Active Monitor on the Ring. The receiving Active Monitor shuts down and becomes a Standby Monitor.

5. A hard error (cable fault, student playing with RI and RO ports) causes the ring to go down. After the hard error is fixed, the neighbor Notification process is restarted.

Monitoring Token and Frame Transmission

The AM monitors the ring to make sure that Tokens and Frames only circle the ring once. There is a Monitor bit in the MAC frame, and whenever a MAC frame is repeated by the AM, the AM sets the Monitor bit to "1". All frames that are received with the Monitor bit set to 1 are not repeated - this means that the frame has already circulated the ring once.

Detecting Lost Tokens and Frames

The AM has a timer to check that there are Tokens and Token Frames circulating the ring. The timer is set for the absolute longest time that it would take for a Token or Frame to circulate the ring. If the timer times out before a new Token or frame is received, the ring is purged, and a new Token is then released.

Purging the Ring

The AM broadcasts the Ring Purge MAC frame to all ring stations on its ring before originating a new Token. Receipt of the returned frame indicates to the AM that a frame can circulate the ring without incident. The Ring Purge Frame resets the ring stations to Normal Repeat mode.

Standby Monitor (SM)

There is only 1 Active Monitor allowed on the ring at a time: all other stations become Standby Monitors. Standby Monitors determine if the Active Monitor is functioning properly. If a Standby Monitor determines that the Active Monitor is not operating properly, the Standby Monitor initiates the Token Claiming process.

Duties of the Standby Monitor

a. The Standby Monitor checks to see if a Token is circulating the ring. It has a Timer called the "Good_token" timer, and knows that a Token has to circulate within that time. If the Token does not go by within this designated period, the Standby Monitor knows that Active Monitor is not doing its job. The Standby Monitor then initiates the Token Claiming process to re-elect an Active Monitor.

b. The Standby Monitor restarts another Timer called "receive notification" whenever an Active Monitor Present frame comes by. If the Timer runs out before another Active Monitor Present frame arrives, the SM assumes that the Active Monitor is not present, or has malfunctioned. The SM initiates the Token Claiming process to re-elect an Active Monitor.

Ring Parameter Server (RPS)

The Ring Parameter Server provides these 3 main services to the ring:

1. It assigns operational parameters to the station at the time of insertion onto the ring. These are parameters such as: Ring Number, Physical Location and Soft Error Report Timer Value. If there is no RPS present, the ring station uses its default values.

2. It ensures that all stations on the ring have the same operational values.

3. It forwards registration information to the LAN Managers from stations attaching to the ring.

Configuration Report Server (CRS)

A Configuration Report Server accepts commands from the network management software to get station information, set station parameters, and remove stations from the ring. It also collects and forwards configuration reports generated by stations on its ring to the LAN manager.

The network management software is a program that monitors the Network, and is used by the System Administrator. It can monitor many Rings and it may include Ethernet segments and connections to WANs.

Ring Error Monitor

The Ring Error Monitor observes, collects, and analyses hard-error and soft-error reports (sent by ring stations on a single ring, and assists in fault isolation and correction).

Hard Error Processing Function

Hard errors are detected, isolated, and bypassed through the use of a Beacon MAC frame. Hard errors are broken cables, failed equipment, improper signal timing, and incorrect voltage levels.

Any ring station that detects a hard error can generate a Beacon MAC frame. The frame is addressed to all other stations on the ring. A Beacon Frame contains the address of the station that discovered the Hard Error, its NAUN and a physical location (the RPS gives this information to the NIC during initialization).

1. Station G hasn't received any frame for a while. Station G starts a Beacon Frame with Station F as its NAUN

2. When Station F receives the Beacon Frame and reads that it is the NAUN of the Fault Domain. It disconnects from the ring and re-attaches to the ring using the Ring Insertion process.

Note: The cable between Station F and Station G is called the Fault Domain.3. If the fault still remains on the ring, the Beacon Frame originator Station G disconnects from the ring, and it re-attaches using the Ring Insertion process.

A fault that can be cured in this manner is called a Temporary Fault. If the fault cannot be cured, then it is called a Permanent Fault.

The Ring Error Monitor monitors the Beacon Frame. It reports (to the network management program) the location, NAUN, Beacon Originator address, and if it is a Temporary or Permanent Fault.

Soft Error Processing

Soft Errors are errors in the Bit pattern or encoding. There are 5 Soft Error types:

1. Line Error:

A code violation between the starting and ending delimiters in the MAC frame, or a Frame Check Sequence error. The FCS doesn't add up!

2. Internal Error:

The ring station recognizes a recoverable internal error. This can be used for detecting a ring station in marginal operating condition.

3. Burst Error:

The absence of transitions for 5 half-bit times. Manchester encoding is used for Token Ring and a low to high transition is a 1 (a high to low transition is a 0).

4. A/C Error:

The Token Ring Frame has 2 bits, called Address-Recognized (A) and Frame-Copied (C). During neighbor Notification, there should not be 2 Standby Monitor Present frames with AC=00 in a row. This would indicate a copy or framing error.

5. Abort Delimiter Error

A station sends out a special frame (called an Abort Sequence or Abort Delimiter) when it discovers an error (soft or hard) while transmitting a frame.

The Ring Error Monitor keeps track of the number of Soft Errors who reported them, and also the NAUN. The REM has thresholds of acceptable Soft Error levels (can be adjusted), and reports excessive Soft Errors to the LAN Manager.

There are two categories of Soft Errors: Non-Isolating and Isolating.

The REM keeps track of all stations and the Soft Errors associated with them.

Where are these Monitors?

Active Monitor

Any station can be the Active Monitor - the first station on with the highest MAC address.

Standby Monitor

All other stations besides Active Monitor

Ring Error Monitor

Usually in a Bridge / Router - something that is always on.

Configuration Report Server

Usually in the same Bridge/Router as the Ring Error Monitor.

Ring Parameter Server

Usually in the same Bridge/Router as the Ring Error Monitor

Token Ring Hierarchy

NOTE: The LAN Manager discussed here is a Network Management program running on OS/2. It is not Microsoft or IBM's Lan Manager network operating system!

IEEE 802.5 Frames

IEEE 802.5 Token Ring Standard has these 3 frames specified:

Abort Sequence

The Abort Sequence is used to indicate to stations on the ring that a temporary internal error has occurred. The Source NIC has discovered a soft error during transmission, and is aborting the transmission. It is a broadcast frame that goes to every station on the local ring.

The Abort Sequence consists of the following:

Token

The Token, sometimes called the free token, is used for Bus Arbitration in a Token Ring. Whichever station claims the Token has the right to transmit its information/data (Token Frame) on the ring. The Token consists of 24 bits:

It is a broadcast frame that goes to every station on the local ring.

Token Frame

The Token Frame is used for transmitting information/data on the ring. It is the only frame with Source and Destination addresses. The Token Frame consists of the following fields:

IEEE 802.5 Frame Fields

Start Delimiter (SD)

The Start Delimiter (SD) is used to:

The SD is 8 bits long and consists of the following:

J K 0 J K 0 0 0

J & K are Manchester coding violations, and have no change of state during the half bit. The J code violation is a steady high state and the K code violation is a steady low state.

Access Control (AC)

The Access Control (AC) field is used to do the following:

The AC field consists of 8 bits:

P P P T M r r r

Where:

The Priority Bits are used to indicate the priority of the Token. Each workstation is assigned a priority for their transmissions: 000 is the lowest, and 111 is the highest (7 levels of priorities). The Lan administrator sets the priority levels. For a workstation to claim a Token, it must have a priority equal to or greater than the priority of the Token.

It is the responsibility of the node (when finished transmitting data) to release the Token and to return the priority bits to the Reservation Bits.

The Reservation Bits are used to negotiate the priority of the next token as a transmission passes by. When a Token or Token Frame goes by, a node is allowed to reserve the priority of the next Token to be released (by placing its priority in the Reservation Bits). In order to change the Reservation Bits, the node's priority must be greater than the existing Reservation bits.

The Token Bit is used to indicate whether the frame is a Token or Token Frame (information frame). T = 0 indicates a Token, T=1 indicates a Token Frame.

The Monitor Bit is used by the Active Monitor (AM) to stop frames from continuously circulating the ring. The Monitor bit is set to M=0 by the Source. When a frame passes by the AM, the monitor bit is set to M=1. If the AM receives a frame with the monitor bit set to M=1, the frame is removed from the ring, purges the ring, and issues a new token.

Frame Control (FC)

The Frame Control (FC) field is used for the following:

The FC field consists of these 8 bits:

F F Z Z Z Z Z Z

Where:

The Frame Type indicates if it is either a MAC layer addressed communication or an LLC layer addressed communication.

The Control bits indicate the type of MAC-level communication.

Destination Address (DA)

The Destination Address (DA) is use to indicate the destination address of the Token Frame (information frame): it consists of 48 address bits.

The first bit is the Individual/Group (I/G) bit - this is used to indicate an individual or broadcast (to everyone). The 2nd bit is used to indicate whether a Universal or Local naming convention. A Universal naming convention would use the MAC address burnt into the NIC's ROM. The Local naming convention would be set by the Lan Administrator. The remaining 46 bits are the stations unique address.

Source Address (SA)

The Source Address is identical to the Destination Address field, except that the Individual / Group bit is always set to Individual (0) for IEEE 802.5. For IBM Token Ring, the first bit is the Routing Bit, which instructs bridges to pass the Token Frame (1) or to ignore it (0).

IEEE 802.5

IBM Token Ring

This difference is very important. Only IBM Token Ring allows Source Route Bridging.

Functional Addresses

Token Ring provides MAC addresses that are reserved for special functions.

Null Address:0000 0000 0000Broadcast Address:FFFF FFFF FFFF
or C000 FFFF FFFFActive Monitor Address:C000 0000 0001Ring Parameter Server:
C000 0000 0002Ring Error MonitorC000 0000 0008Configuration Report ServerC000 0000 0010
Netbios Address:C000 0000 0080Bridge Address:C000 0000 0100Lan Manager Address:
C000 0000 2000User Defined addresses:C000 0008 0000 to C000 4000 0000

Information Field (INFO)

The Information Field is used for the following:

The Frame Control bits FF determine if the information is for either the MAC (FF=00) or LLC (FF=01) layer. The LLC information is processed identical to Ethernet (as covered earlier under IEEE 802.2), and will not be covered here.

The Routing Bit determines if the frame uses either normal MAC layer communication, or if Source Routing is used. Only the IBM Token Ring uses Source Routing: IEEE 802.5 does not.

If the Routing Bit is set, then the MAC Frame INFO field contains routing information that's used during Source Routing. The Routing Bit instructs the bridge to pass the Token Frame (or ignore).

If the Routing Bit is not set, then normal MAC layer local ring communication is active. This includes specific subcommands called MVID (Major Vector IDs) that work with the Frame Control control bits to: Beacon, Claim Token , Ring Purge, etc.. (See Token Ring MVID Table)

Frame Check Sequence (FCS)

The Frame Check Sequence (FCS) is used for error checking: it uses 32 bit CRC (cyclic redundancy checking). It checks the FC, DA, SA and INFO fields (it is 4 bytes long).

End Delimiter (ED)

The End Delimiter is used to indicate the end of the frame. It consists of 8 bits in a "J K 1" sequence, where J & K are Manchester coding violations.

J K 1 J K 1 I E

Where:

Frame Status (FS)

The Frame Status (FS) field is used to indicate if the address was recognized by the destination, and if the frame was copied. This acts as an indication to the source that the destination station is present (and accepting data). This FS field consists of these 8 bits:

A C r r A C r r

Where:

Note: The A/C bits are provided twice for redundancy.

If this section was helpful, why not donate to further development?

Donate $1.25.

Donate $2.50.

Donate $5.00.

PrevHomeNext
Gateways Linux on Token Ring