Figure 4-3. Token Ring frame formats
Token Ring is a ring topology LAN technology with a token-passing mechanism for eliminating contention. There are actually two standards for Token Ring. It was originally developed by engineers at IBM who created the initial specification. Subsequently, the IEEE took on the responsibility of making an industry standard Token Ring specification, under the designation 802.5. There are slight differences between the two standards, but they interoperate without any issues. One of these minor differences is that the IEEE reduced the maximum number of devices on a ring from 260 to 250 for Type 1 shielded cabling. The maximum number of devices for a ring built with Category 5 unshielded twisted pair cabling (UTP) is only 72, however.
Figure 4-3 shows the formats for both an empty token and a frame that carries user data. Several features are remarkably similar to Ethernet and some are quite different. For example, the maximum Ethernet frame is 1518 octets long, while the maximum Token Ring frame size can be as much as 4550 octets in a 4Mbps ring, or 18,200 for a 16Mbps ring. In general, the Token Ring frame size is governed by the ring's hold-time parameter, which governs how long any one device is permitted to have the token.
Both Ethernet and Token Ring use 6-byte addresses. This fact makes bridging between the two much easier. However, there is one significant twist. Token Ring orders its bytes so that the most significant bit comes first, while Ethernet does exactly the opposite. If you bridge Ethernet and Token Ring segments together, you have to translate the addresses from one format to the other by changing the bit ordering.
A quick side note on the subject of Ethernet to Token Ring bridging is that there are in fact two different options for running LLC2 on Ethernet. The most common method is to simply use 802.3 format frames. The second option that appears in some installations is the so-called 80d5 format. This strange name refers to the Ethernet type code for the Ethernet II style frames. I have seen some overly helpful LLC2 server software that uses both frame types. This software tends to cause serious confusion on the bridge and should be avoided.
In general, bridging works well between Ethernet and Token Ring networks, but it is important to be careful of Maximum Transmission Unit (MTU) and watch out for these address translation issues. This warning implies, in turn, that it is possible to create a bridge between two Token Ring segments through an intermediate Fast or Gigabit Ethernet segment. If this is the goal, it would usually be more practical to use a TCP/IP tunnel protocol such as DLSw. This protocol would then allow the Token Ring MTU to be preserved through the Ethernet leg of the connection.
There are two common standards for Token Ring; one operates at 4Mbps and the other at 16Mbps. A high-speed standard also operates at 100Mbps. But, unlike the 4 and 16Mbps standards, the 100Mbps version has not seen widespread use to date.
In most cases, running the same equipment at either 4 or 16Mbps is possible. This possibility is useful, since it allows the network to use the lowest common technology. But there is an unfortunate converse to this property: once a ring has started operating at either speed, introducing a new device at the other speed can cause serious problems for the entire ring.
The token in Token Ring refers to the way that the ring avoids packet collision problems. There is a simple token packet that contains no user information. This packet is passed from station to station around the ring from an upstream to downstream neighbor. If a device has something to send, it waits until it gets the token. Then it replaces the token packet with a data packet, which it sends to its downstream neighbor instead of the token. The first two fields of this data packet look similar to the original token packet, except for one bit (the token bit) that indicates that data will follow.
This device then continues to transmit data until it either runs out of data to send or until it sends a predefined maximum number of bytes. The data can be sent either as one packet or it can be broken up into several separate packets, which can be useful if the device is talking to many other devices. Once it is done talking, it places a new token on the ring so that other devices will have an opportunity to talk as well. This method ensures that every device gets to participate fairly.
The packets that are sent this way all pass from the sender to its downstream neighbor, which forwards them on to the next downstream neighbor, and so forth, until they reach all the way back around the ring to the original sender. The sender is then responsible for removing the frames from the ring (which is called "stripping"). Once these frames are removed from the ring, the sender replaces a token onto the ring so the next device can talk.
Another interesting feature that is available in the 16Mbps Token Ring (but not in the 4Mbps standard) is Early Token Release (ETR). In this case, the sender doesn't wait until it has seen its own frames finish circulating around the ring. Instead, it puts the token back onto the ring for the next device as soon as it has finished its transmission. This placement makes the ring much more efficient. What's particularly useful about ETR is that not all devices on the ring need to support it. In fact, it should be fairly obvious that the whole ring benefits from improved throughput if even a few devices can do this.
The most common problem that arises in Token Ring topologies is a broken ring. If, for whatever reason, the ring is not closed, then the packets never get back to their starting point. As I just described, the normal token-release mechanism depends on the source device receiving the frames it has sent before it passes the empty token along for another device to use. A broken ring is a serious problem.
The ring deals with a break by sending around a beacon packet. The device that is immediately downstream from the break alerts the rest of the ring by sending this special type of frame. The beacon frame contains the Layer 2 MAC address of the upstream neighbor that has lost its connection. In this way the failure domain, the area in which the break has occurred, is identified. This identification allows the ring to go into its reconfiguration procedure.
A particular device on the ring is designated as the ring monitor. This device is responsible for putting new tokens onto the ring when they are lost. In a broken ring situation, the monitor device is able to keep communication working. However, this situation is considerably less efficient than having a properly closed ring.
Token Ring has, in general, seen significantly less industry acceptance than Ethernet, despite having some useful advantages. There are important exceptions, particularly in mainframe environments, where Token Ring is more popular, but Ethernet is definitely more popular. With the advent of higher speed versions of 802.3, the industry preference for Ethernet appears to be growing.
The main reason for this gap in acceptance is simply the implementation cost. Token Ring hubs and switches cost more than Ethernet hubs and switches. However, the largest portion of the cost difference comes from the higher prices of Token Ring NICs and the cost of the Token Ring chipsets used by these cards.
The development of 100Mbps Fast Ethernet finally eliminated the bandwidth advantages of Token Ring. By the time the new 100Mbps Token Ring devices were available, Gigabit Ethernet was also coming out. Although Token Ring has not reached any theoretical limitations, it seems unlikely that a Gigabit Token Ring standard will be available before the 10 Gigabit Ethernet makes it obsolete.
The result is that Token Ring tends to be used primarily in organizations with large legacy mainframe environments that make extensive use of Token Ring and protocols such as LLC2. In organizations with modern mainframes, there is a trend toward adopting TCP/IP as the protocol of preference and toward the installation of Gigabit Ethernet modules into the mainframes or their front-end processors.
One of the most popular methods for connecting devices to a Token Ring is the Multistation Access Unit (MAU). This unit is traditionally an unmanaged device, with the classic example being the IBM 8228. The 8228 requires no external power. Rather, the individual end devices provide the power for the MAU to operate. When no device is connected to a port, a relay switch disconnects it from the ring. Then, when a device connects, it provides power to this relay and is inserted into the ring electrically.
At either end of the MAU are ports labeled "Ring-In" (RI) and "Ring-Out" (RO). These ports are provided to allow several MAU devices to be chained together into a larger ring. The RI port from one MAU is connected to the RO port of the next, and so on around the ring until the RO port of the first MAU is connected back to the RI port of the last to close the ring. These RI and RO ports are not equipped with the same sort of relay switching as the regular device ports, so they can only be used for interconnecting MAUs.
The IBM 8228 uses IBM UDC connectors (also called "hermaphroditic" connectors because they are both male and female), which are usually used with IBM Type 1 shielded cabling. However, there are also RJ45 MAU units from both IBM and other manufacturers. Generally, the RJ45 units require external power like Ethernet hubs. Other than this, no real functional differences exist between these two types of MAUs.
A managed Token Ring hub is also sometimes called a Controlled Access Unit (CAU). This hub generally operates in a similar manner, except that it becomes possible to monitor utilization and error statistics on individual ports, and activate or deactivate ports individually. Some of these hubs have the ability to assign individual ports to different physical rings internally, similar to an Ethernet VLAN.
Token Ring switching is really nothing more than Token Ring bridging, just like in Ethernet. Also as in Ethernet networks, Token Ring switches are becoming a popular method of connecting important end devices to the network. In effect, the switch becomes a multiport bridge that interconnects a large number of different rings.
Typically, source-route bridging is used in Token Ring networks. Besides the simple bridging functions that allow a separate token on each port, source-route bridging allows devices to find the optimal path through the network. They find this path by means of Routing Information Field (RIF) data that is added to packets as they travel through the network.
When a device wants to find a particular MAC somewhere on the network, it sends out an "explorer" packet. This broadcast frame finds its way throughout the whole network, hopefully to the desired destination. Along the way, the explorer packet picks the RIF information that describes the path it took. The destination device then responds using the shortest path.
As with Ethernet switches, Token Ring switches are able to implement VLANs. This is done in exactly the same way as on Ethernet. All ports that are part of the same VLAN are bridged. Meanwhile, distinct VLANs do not communicate directly, but must instead communicate through a router.
Each port on a Token Ring switch is a distinct ring, regardless of VLAN membership. This situation is analogous to Ethernet, where each port is a distinct collision domain, regardless of VLAN membership. Two ports that are members of the same VLAN have a bridge connecting them, while ports from distinct VLANs do not. As in Ethernet, when individual end devices are connected to a switch directly, it is possible to use a full-duplex version of the protocol allowing simultaneous sending and receiving of packets. This full-duplex version of Token Ring requires that the end device have Direct Token Ring (DTR) capabilities. DTR is not part of the 802.5 standard, but is nonetheless a widely implemented feature.