Token Ring Protocol (IEEE 802.5)

Introduction

In a token ring, one or more tokens are passed from host to host as part of a local area network (LAN) data link where all devices are connected in a ring or star topology. Whenever data is sent between network nodes, it is called a token. Tokens are released once the recipient of the data has been verified, and only a host that holds a token can send data. Token rings were well-liked because they functioned effectively with high volumes of traffic, but they weren't ideal for huge networks, especially if those networks were dispersed extensively or contained physically far-flung nodes. Multistation access units (MSAs), which are similar to hubs on Ethernet, were introduced to get around some of these restrictions. MSAs, also called concentrators, are centralized wire hubs.

Modes of Operation

Data is broadcast sequentially from one ring station to the next on a Token Ring LAN, where stations are logically arranged in a ring topology. Access is controlled by a control token that circulates the ring. The ARCNET, token bus, 100VG-AnyLAN (802.12), and FDDI all employ similar token-passing systems, and all have theoretical advantages over the CSMA/CD of early Ethernet. One server serves queues in a cyclical order in a Token Ring network, which can be described as a polling system.

The following list of different operating modes is available:

• The incoming bits are merely sent to the output line in the listen mode, and nothing further is done with them.
• When a station is linked to the ring interface and acquires a token, the ring interface is configured to the talk or transmit node. The single-bit buffer that served as the direct input-to-output connection has been cut off.
• When the node is down, this mode—by-pass—is activated. Just much all data is ignored. There is no one-bit lag in this setting.

Access Control

The steps involved in data transfer are as follows:

• On the ring, empty information frames are constantly being passed around.
• The token is grabbed by a computer when it needs to transmit a message. The frame can then be sent by the computer.
• The following workstations each have a turn looking at the frame. The workstation that declares itself as the message's intended recipient copies the message from the frame and resets the token to 0.
• The token has been changed to 0, and the message has been duplicated and received when the frame returns to the sender. The message is taken out of the frame.

Multistation Access Units and Controlled Access Units

MAU may manifest as a hub or a switch; as Token Ring had no collisions, hubs were the most common form of MAU produced. Despite using LLC to operate, Token Ring also has source routing to send packets outside of the local network. The majority of MAUs are set up by default in a "concentration" mode, while later MAUs allow the option to also operate as splitters and not just concentrators, such as the IBM 8226. Later, IBM would provide Controlled Access Units known as Lobe Attachment Modules that could support several MAU modules. The CAUs provided features including Dual-Ring Redundancy for backup routing in the event of a dead port, modular concentration with LAMs, and numerous interfaces like the majority of subsequent MAUs. Compared to an unmanaged MAU hub, this allowed for a more dependable setup and remote management.

Cabling and Interfaces

A heavy two-pair, 150 Ohm shielded twisted pair cable known as IBM "Type-1" is the standard for cabling. This was the fundamental wire for the "IBM Cabling System," a structured cabling system that IBM believed would be extensively used. The connectors utilized were special hermaphrodites, sometimes known as Boy George connectors or IBM Data Connectors in colloquial usage. Due to their size, fragility, and requirement for at least 3 3 cm of panel space, the connectors have a few drawbacks. The connectors' advantages include gender neutrality and better shielding than the unshielded 8P8C standard. The computer typically used DE-9 female connectors. Later Token Ring implementations supported Cat 4 cabling as well, allowing 8P8C ("RJ45") connectors to be utilized on the MAUs, CAUs, and NICs. Many network cards also supported both 8P8C and DE-9 for backward compatibility.

Frame Types

On a Token Ring network, token, abort, and frame are three different types of frame formats that are supported. The method by which access to the ring is transferred from one network-connected computer to another is through the token format. The starting and ending delimiters, which are used to mark the start and end of a token frame, are the first two bytes of the token format in this instance, which is made up of three bytes. Access control is contained in the middle byte of a token frame.

While one bit serves as the token bit and another bit position serves as the monitor bit, three bits each serve as a priority indicator, a reservation indicator, and a reservation indication with three bits each.

Components

The Token Ring Frame Format is made up of the following elements:

• Start Delimiter (SD): The first field of the data/command frame, SD, is one byte long and serves as both a frame arrival notification and a means of synchronizing the retrieval timing of the receiving station.
• Control of Access (AC) The AC field has four subfields and is one byte long. The priority field is the first field of three bits. The token bit is the fourth bit in the sequence.
  • Implementing priority is made easier by priority bits and reservation bits. Reservation bits plus priority bits equals 3. For instance, the client has a priority of 0, while the server has a priority of 7
  • Token frame presence is signaled by the token bit. Token frames are created if the token bit is set to 1; otherwise, they are not created.
  • A solution to the orphan packet problem is the monitor bit. As monitors are capable computers that can recalculate CRC while changing monitor bit, it is covered by CRC. If the monitor bit is 1, the monitor will stamp the document; if it is 0, the monitor won't stamp it yet.
• The FC field, which is one byte long and has two fields, is called "Frame Control." The first is a one-bit field that serves as a flag for the kind of data the Protocol Data Unit contains (PDU).
• Destination Address (DA) The two to six-byte DA field contains the physical address of the frame's subsequent destination. The DA is the address of the router to the following LAN on its path if its final destination is another network.
• The Source Address (SA) field, which is two to six bytes long, contains the physical address of the sending station. The SA is that of the originating station if the packet's final destination is a station connected to the same network.
• Data The PDU is located in the sixth field, data, which has a 4500-byte allotment. A PDU length or type field is absent from a token ring frame.
• Four bytes make up the checksum field. The data at the sending station is double-checked using the checksum field. The number of bytes in the frame is totaled in this field. After determining how many bytes are in the received frame, the number is verified at the receiving end.
• End Delimiter (ED): The ED denotes the conclusion of the sender's data and control information and is represented by a second flag field of one byte.
• The frame's final byte contains the FS field, which stands for frame status. Both the monitor and the receiver can configure it to indicate that the frame has already been read or that it has already circled the ring.

Active and standby monitors

The Active Monitor

The active monitor is a designated station on the ring and is in charge of carrying out this mission. Based on an election procedure known as the claim token process, the active monitor is chosen. After being chosen, the active monitor is in charge of handling specific error scenarios that could arise on the ring, such as misplaced tokens and frames or timing issues. The active monitor's job includes clearing the ring of any continuously floating frames. The frame might keep going around the ring indefinitely if a device that has already sent a token to the network fails. An active monitor (AM) or standby monitor (SM) station is the type of station that each node in a Token Ring network is. At any given time, a ring can only have one active monitor. A vote or monitor contention mechanism is used to determine which monitor is active.

When one of the subsequent events occurs, the monitor contention process starts:

• The ring has a signal loss.
• All other stations on the ring are unaware of an active monitor station.
• When a station doesn't see a token frame for seven seconds in a row, for example, a specific timer on the station expires.

If any of the aforementioned circumstances occur and a station determines that a new monitor is required, it will transmit a "claim token" frame indicating its desire to take over as the new monitor. It's okay for that token to change roles and become the monitor if it returns to the sender. If more than one station simultaneously attempts to take over as the monitor, the election will be won by the station with the highest MAC address. The remaining stations become standby monitors. Every station must be able to switch from being a passive to an active monitor station at any time. Only one Active Monitor is permitted on the ring at once; all other stations turn into Standby Monitors. The Active Monitor's performance is checked by the Standby Monitor. The Token Claiming process is started if a Standby Monitor finds that the Active Monitor is not functioning properly.

The Standby Monitor

• The Standby Monitor determines whether a Token is moving around the ring. It is aware that a Token must circulate within that time and has a timer called the "Good token" timer. The Standby Monitor is aware that the Active Monitor is not performing its duties if the Token does not pass within this allotted time. After that, the Standby Monitor starts the Token Claiming procedure to choose an Active Monitor once more.
• Every time an Active Monitor Present frame passes by, the Standby Monitor resets a Timer labeled "receive a notification." The SM assumes that the Active Monitor is absent or broken if the Timer expires before another Active Monitor Present frame appears. To re-elect an Active Monitor, the SM starts the token-claiming procedure.

Token Insertion Process

Before being permitted to take part in the ring network, Token Ring stations must go through a 5-phase ring insertion process. The Token Ring station won't insert into the ring if any of these steps fail, and the Token Ring driver might indicate an error.

• An initial lobe media check is conducted by a station during phase 0 (Lobe Check). To send 2000 test frames along its transmit pair, which will loop back to its receive pair, a station must be wrapped at the MSAU. The station verifies that it can successfully receive these frames by performing a test. Phase 1 relay opening is then accomplished by a station sending a 5-volt signal to the MSAU (Physical Insertion).
• The station then emits MAC frames with its own MAC address in the Token Ring frame's destination address field during Phase 2 (Address Verification). If the Address Recognized (AR) and Frame Copied (FC) bits in the frame status are set to 0, which signifies that no other station on the ring is currently using that address, then the station must participate in the periodic (every 7 seconds) ring poll procedure. It is at this point that stations on the network identify themselves as a part of the MAC management procedures.
• When a station learns the address of its Nearest Active Upstream Neighbour (NAUN) and transmits that information to its nearest downstream neighbor, Phase 3 (Participation in Ring Poll), which is when the ring map is created, takes place. The AR and FC bits are set to 0, and the station is ready to receive an AMP or SMP frame when it arrives. The station then queues an SMP frame for transmission and changes both bits (AR and FC) to 1 if enough resources are available. If no such frames are obtained within the next 18 seconds, the station reports a failure to open and de-insert from the ring. If it successfully takes part in a ring poll, the station advances to the last phase of insertion, request initiation.
• Phase 4 (Initialization of the Request): To collect configuration data, a station at this point makes a particular request to a parameter server. This frame is transmitted to a particular functional address, usually, a Token Ring bridge, which could contain timer and ring number data the new station has to be aware of.

In the Token Ring priority MAC, there are eight priority levels that range from 0 to 7. The priority bits are set to the station's requested priority when the station that wishes to transmit receives a token or data frame with a priority that is lower than or equal to the station's requested priority. The station does not immediately broadcast; instead, the token travels throughout the medium before returning to it. Following the transmission and reception of its data frame, the station resets the token priority to its starting position.

For devices that implement 802.1Q and 802.1p, the following eight access priorities and traffic kinds are listed:

Optional Priority Scheme

A complex priority mechanism is used in Token Ring networks to allow some user-designated, high-priority stations to access the network more often. The priority field and the reserve field are the two fields that govern priority in Token Ring frames. Only stations having a priority greater than or equal to the value of a token's priority can seize that token. Only stations with a priority value greater than the transmitting station can reserve the token for the following pass around the network once it has been captured and converted to an information frame. The greater priority of the reservation station is included in the following token when it is produced. After their transmission is finished, stations that increase a token's priority level are required to restore the previous priority. It must have a higher priority if a non-destination station wants to reserve the token for the following pass, and it utilizes that priority to reserve the token by encoding it in the reservation bits. The Frame Check Sequence does not involve the Frame Delimiter or Access Control, therefore updating the reservation bits does not necessitate a new CRC.