Token Ring Switching
Token Ring switching offers many of the advantages found in switched Ethernet networks. Most notably, it offers more bandwidth by reducing the size of the Token Ring domain. In the extreme case, the smallest Token Ring domain is a dedicated workstation on a switch port. Catalyst Token Ring switching creates two virtual entities: the Token Ring Concentrator Relay Function (TrCRF) and the Token Ring Bridge Relay Function (TrBRF). Together, these functions form a switched Token Ring network with smaller token domains.
Source-route switching described in the later section “Source-Route Switching” provides a service within the TrCRF for transferring frames between ports within a ring, or to the TrBRF to transfer to another TrCRF.
Another Catalyst Token Ring feature protects you from misconfiguring the system with a duplicate ring on another Catalyst. This is an erroneous configuration, which, if actually performed, could prevent your Token Ring network from operating correctly. DRiP is described in the later section “Duplicate Ring Protocol (DRiP).”
Token Ring Concentrator Relay Function (TrCRF)
In Token Ring networks, multistation access units (MAUs) interconnect workstations to form a physical star topology. The MAU, a type of concentrator, maintains ring functionality and passes tokens and frames among all stations attached to the ring. All the stations attached to the MAU share the same token and bandwidth.
In the Catalyst, the network administrator assigns ports to a virtual concentrator called a Token Ring Concentrator Relay Function (TrCRF). All the ports assigned to a TrCRF belong to the same ring number. However, the workstations attached to each port experience full media bandwidth and have no perception that other stations share the ring.
The TrCRF defines the port to ring number association. The Catalyst in Figure 3-8 defines four TrCRFs to interconnect fourteen ports. (A port is a member of one and only one TrCRF.)
Figure 3-8. Defining TrCRFs to Interconnect Ports
In general, a TrCRF can reside in only one Catalyst and cannot span outside of a Catalyst. This is called an undistributed TrCRF. An exception to this, the default TrCRF, spans Catalysts and is referred to as a distributed TrCRF. (A backup TrCRF may also span Catalysts.) The default TrCRF enables all Token Ring ports to belong to a common TrCRF without any administrator intervention. Users can attach to any Token Ring port and communicate with any other station in the distributed TrCRF network. The default TrCRF behaves like a giant ring extending across all Catalysts and provides the “plug and play” capability of the Catalyst in Token Ring networks.
When configuring a TrCRF, you must define the ring and VLAN numbers, and you must associate it with an existing parent TrBRF (TrBRFs are discussed in the next section). The parent TrBRF is assigned a VLAN number and is the identifier used to associate the TrCRF to the TrBRF. In addition, you may define whether the TrCRF will operate in the SRB or SRT mode. If left unspecified, the TrCRF will operate in SRB mode.
Token Ring Bridge Relay Function (TrBRF)
The Token Ring Bridge Relay Function (TrBRF) acts like a multiport bridge and interconnects rings. The TrBRF defines the TrCRF-to-bridge association. A TrCRF belongs to only one parent TrBRF, but multiple TrCRFs may attach to a parent bridge. Figure 3-8, for example, shows more than one TrCRF attached to a TrBRF. The TrCRFs interconnect through the TrBRF the same as rings do through the source-route bridges and source-route transparent bridges previously described. Unlike the TrCRF, the TrBRF can span between Catalysts, as shown in Figure 3-9. This allows TrCRFs on various Catalysts to belong to the same bridge
Figure 3-9. TrBRF Extending Across Catalysts
Enabling a TrBRF requires a bridge and VLAN number. The TrBRF’s VLAN number is paired with the TrCRF VLAN number to create the parent-to-child relationship. Because a TrCRF must associate with a parent TrBRF, the default TrCRF belongs to a default TrBRF. When enabling Token Ring switching with a non-default TrCRF and TrBRF, you must first configure the TrBRF, then the TrCRF, and finally, group ports to TrCRFs. Referring to Figure 3-9, you would do the following:
- Configure the TrBRF at the top of the drawing. To do this, you define the bridge number and the VLAN number as in the following:1Console> (enable) set vlan 100 type trbrf bridge
- Work down the figure to the TrCRF(s). Create the TrCRF, specifying the VLAN number, the ring number, the bridge type, and parent TrBRF. Here is an example:1Console> (enable) set vlan 110 type trcrf parent 100 ring 10
- After creating the TrCRF, associate ports to the correct TrCRF.
Source-route switching describes the mechanism of bridging Token Ring traffic. The modes are determined based on the location of the source and destination devices relative to the bridging function. The source-route switch (SRS) decides whether to transparently bridge a frame within a TrCRF or to source-route bridge to another TrCRF. When a station on a switch port transmits to another station residing on a different port but belonging to the same TrCRF, the SRS forwards the frame based on the destination MAC address. The SRS learns source MAC addresses and makes forwarding decisions the same as a transparent bridge. However, if the source and destination are on different rings, the source creates a frame with a RIF. The SRS examines the RIF and passes the frame to the bridge relay function for forwarding.
Although this sounds like a source-route bridge, a significant difference distinguishes an SRS from an SRB. When a station transmits an ARE, an SRB modifies the RIF to indicate ring/bridge numbers. An SRS never modifies a RIF; it simply examines it. When a source sends an all-routes explorer, for example, it sets the RII bit to “one,” indicating the presence of a RIF. Examination of the initial explorer frame, however, reveals that the RIF is empty. The SRS notices that the RII bit value is “one” and forwards the explorer to the TrBRF unmodified. The SRS simply says, “This is a source-routed frame; I better send it to the TrBRF and let the bridge worry about it.” In contrast, an SRB or TrBRF modifies the explorer RIF by inserting ring and bridge numbers.
In Chapter 2, a broadcast domain was defined and equated to a VLAN; a broadcast domain describes the extent to which broadcast frames are forwarded or flooded throughout the network. The domain boundaries terminate at a router interface and include an interconnected set of virtual bridges. But Token Ring complexities present ambiguities when defining a VLAN. In a source-route bridged environment, Token Ring actually creates two kinds of broadcasts: the intra-ring and the inter-ring broadcast.
A device generates an intra-ring broadcast whenever it produces a broadcast frame without a RIF, and the explorer bit is set to “zero.” A station may do this whenever it wants to determine whether the destination is on the same ring as the source. The SRS function floods this frame type to all ports within the TrCRF. The frame does not cross the TrBRF.
In contrast, an inter-ring broadcast frame sets the explorer bit to “one,” enabling the frame to cross ring boundaries. The TrBRF floods the inter-ring frame to all attached rings; all rings receive a copy of the frame. Figure 3-10 illustrates Token Ring VLAN boundaries.
Figure 3-10. Token Ring VLANs Illustrated
A network may see both the intra- and inter-ring broadcasts. Which one actually describes a VLAN? A VLAN in the Token Ring network includes the TrCRF and the TrBRF. A VLAN exists whenever Token Ring networks must interconnect through a router.
Duplicate Ring Protocol (DRiP)
In a Token Ring environment, each ring has a unique ring number identifying it for source-route bridging. Similarly, in a switched Token Ring, except for the default and backup TrCRFs mentioned earlier, each TrCRF has a unique ring number. If an administrator accidentally misc onfigures another TrCRF with the same ring number, shown in Figure 3-11, the Token Ring switching process gets confused.
Figure 3-11. Do not attempt this. Duplicate ring numbers are not allowed on multiple switches.
To mitigate the effects of duplicate ring numbers in a switched network, Cisco developed a proprietary protocol to detect them and react accordingly. The Duplicate Ring Protocol (DRiP) sends advertisements with the multicast address 01-00-0C-CC-CC-CC to its neighbors, announcing VLAN information for the source device only. By default, DRiP announcements occur every 30 seconds or whenever a significant configuration change occurs in the network. DRiP announcements only traverse ISL links and are constrained to VLAN1, the default VLAN. The receiving Catalyst then compares the information with its local configuration. If a user attempts to create a Token Ring VLAN that already exists on another Catalyst, the local unit denies the configuration.