This section covers SRB explorer frames, SRT, SR/TLB, RSRB, and DLSw Plus (DLSw+).
IBM developed SRB in the mid-80s as a way to bridge between Token Ring LANs. The IEEE adopted most of IBM's proposal into the IEEE 802.5 standard. As shown in Figure 4-20, in SRB, the source determines the route to the destination node by sending an explorer frame to it. SRB bridges do not keep a MAC table of hosts and do not worry about bridge loops. This is different to Ethernet transparent bridging where the bridges have the smarts of MAC layer destinations and determine the path by creating a spanning-tree. SRB networks are limited to a 7-bridge hop count. (Some implementations can extend to 13 hops.) In SRB, the source node obtains the knowledge of routes to destinations on the network. This is accomplished using explorer frames.
Figure 4-20 SRB Network
When a source node wants to send information to a destination that is not on the local LAN, it sends out an explorer frame. Bridges pick up the explorer frames and forward them out the other interfaces. The bridges add route information to the frames as they travel through the network. The route information includes a Token Ring number and bridge number pair that the explorer frame travels on. As seen in Figure 4-21, when the explorer frame reaches Bridge 6, a route descriptor is added to the frame that includes Ring 7/Bridge 6. At Bridge 4, another route descriptor with Ring 3/Bridge 4 is added. Finally, Ring 2 is added at the destination node with the bridge value set to 0 to indicate that the destination has been reached.
Figure 4-21 SRB Ring/Bridge Routing Information
When the explorer frames arrive at the destination, the destination frame sets the direction bit (also known as the D-bit) to 1 and sends the frame back to the source node through the same route that was used to arrive to the destination. When multiple frames reach back to the source node, it usually uses the route of the first frame received. Other decision metrics include the minimum number of hops and the path with the largest MTU allowed.
There are two types of explorer frames. The first is the all-routes explorer (ARE) frame. This explorer takes all possible paths to the destination. The second type is the spanning-tree explorer (STE) frame; these explorer frames use a spanning tree to reach a destination to prevent loops.
The RIF is contained in the 802.5 frame and is composed of a 2-byte RIF header. The RIF can contain one or more route descriptors. Figure 4-22 shows the RIF header and route descriptor format.
Figure 4-22 RIF
The first two most significant bits indicate the type of frame:
00Non Broadcast (specific route). Indicates that this is a regular frame that should be routed to the destination following the information in the route descriptors.
10All-routes broadcast. Indicates this is an All-rings explorer packet that should take all possible routes to the destination.
11Single-route broadcast. Indicates that this is a limited broadcast that should take one path to the destination.
The next bit is unused.
The 5 least significant bits in the first byte indicate the total length of the RIF field including the 2-byte header. A RIF with only one bridge hop is 6-bytes long.
The most significant bit in the second byte is the D-bit (direction bit). A value of 0 indicates forward direction, which means the source to the destination. The route descriptors are read from left to right when the D-bit equals 0. A value of 1 indicates reverse direction, meaning the destination station has seen it and the frame travels from the destination to the source. The route descriptors are read right to left when the D-bit equals 1.
The next 3 bits indicate the maximum transmission frame size for the 802.5 frame. This is the maximum frame size that this station is willing to accept. A value of 011 indicates 4136 bytes, and a value of 100 indicates 8232 bytes. (Some references might say the values are 4472 and 8144, respectively.) Table 4-9 shows possible maximum frame size values.
Table 4-9 RIF Max Frame Size Values
RIF MTU Bit Value
Max Frame Size (bytes)
The lower 4 bits are not used.
The route descriptors are 2 bytes long. The first 12 bits indicate the ring number. The 4 least-significant bits indicate the bridge number.
Figure 4-23 shows Token Ring and bridge numbers between Hosts A and B. Some numbers are in decimal and others are in hexadecimal format.
Figure 4-23 RIF Example
The RIF header from Host A to Host B is constructed as follows:
00001000 00110000 = 0x0830
From the bits, you can determine the following:
This is a regular frame (type = 00).
It is 8 bytes in length (length = 01000).
The maximum frame size is 4472 (MTU = 011).
The direction bit is set to 0.
The route descriptors are constructed as follows. The last bridge number is set to 0x0:
Ring 20 (0x014) to Bridge 12 (0xC) to Ring 0x003 to Bridge 10 (0xA) to Ring (0x010) = 014c.003A.0100
The RIF field in hex is 0830.014c.003a.0100.
The CCIE candidate should know how to read and construct the RIF field for regular SRB and in other scenarios, such as RSRB and DLSw. RIF examples are included in each corresponding section.
SRB Configuration with Cisco Routers
SRB is limited in that it is a flat network topology in the data-link layer. Another limitation is the maximum hop limitation of 7 bridges. Routers can be introduced into the network to help limit the number of hops on the network. Remote Token Ring networks can be attached to routers to reduce bridge hops. This is accomplished with the concept of the virtual ring, as seen in Figure 4-24. Consider each router interface a minibridge that connects the external ring to the internal virtual ring. You use the ring number in the RIF field just like any other physical ring.
Figure 4-24 SRB Configuration
To configure the virtual ring group, use the global command source-bridge ring-group virtual-ring-group-number. Each Token Ring LAN is bridged to the virtual ring by using the interface command source-bridge ring-number bridge-number virtual-ring-group-number. The bridge number can be from 1 to 15. The ring number is from 1 to 4095.
The configuration for the router in Figure 4-24 is displayed in Example 4-7.
Example 4-7 Configuration Example of Source Route Bridging
source-bridge ring-group 5 ! interface tokenring 0 source-bridge 1 10 5 source-bridge spanning ! interface tokenring 1 source-bridge 2 11 5 source-bridge spanning ! interface tokenring 2 source-bridge 3 12 5 source-bridge spanning
The frame in this example from Token Ring 1 to Token Ring 2 is source routed as follows: Ring 1 to Bridge 0xA to Ring 5 to Bridge 0xB to Ring 2. The RIF is 0830.001a.005b.0020.
The CCIE candidate must master how the RIF is built. At the end of some of the following sections, there is a paragraph on how the RIF looks.
SRT is specified in the IEEE 802.1d Appendix C standard. SRT bridges can forward traffic from both transparent and source-route end nodes and form a common spanning tree with transparent bridges. SRT bridges combine the implementations of transparent bridges and source-route bridges and can distinguish between source-route and transparently bridged frames. SRT allows transparent or source-route stations to communicate with other stations of the same type.
SRT bridges use a Routing Information Indicator (RII) bit to distinguish between SRB and transparent bridge frames. The RII values are as follows:
0, means a RIF is present; use the SRB algorithm.
1, means a RIF is not present; the frame is transparently bridged.
SR/TLB bridges provide bridging between Ethernet and Token Ring networks. The difference between SR/TLB and SRT bridges is that SR/TLB bridges can forward frames between source-route and transparently bridged networks. Considering the differences between these two technologies, SR/TLB has many challenges to overcome so that transparent Ethernet networks can communicate with SRB token ring networks and vice versa.
As shown in Figure 4-25, source-route translational bridges overcome the following issues:
Bit ordering of MAC addressesAs explained in an earlier section, Ethernet NICs expect to receive the least-significant bit first for each octet; Token Ring considers the first bit to be the most significant bit of each octet. Translational bridges reorder source and destination addresses when translating between Token Ring and Ethernet.
MTU sizeEthernet's MTU is 1500 bytes; Token Ring has a MTU of 4472 bytes on 4 Mbps networks and 17,800 on 16 Mbps networks. Translational bridges usually set the MTU on the Token Ring networks to 1500 to facilitate the transfer of frames to Ethernet. They check the MTU of every frame.
Frame StatusEthernet does not use frame status bits that are used in Token Ring. Some bridges set the C (copied) bit to 1 when a frame is transferred from Token Ring to Ethernet.
Explorer FramesEthernet does not use explorer frames. These frames are dropped when converting from Token Ring to Ethernet segments.
RIFEthernet does not understand the concept of source-routed frames.
Spanning-tree algorithmToken Ring bridges do not understand Ethernet's STP.
Frame ConversionEthernet V2 frames are converted to Token Ring SNAP frames, and vice versa.
Figure 4-25 SR/TLB
Configuring SR/TLB involves the configuration of SRB and transparent bridging as described in earlier sections. An additional command ties in the SRB domain with the transparent bridged domain:
source-bridge transparent ring-group pseudo-ring bridge-number tb-group
The arguments are as follows:
ring-groupThe virtual ring group number created with the source-bridge ring-group command.
pseudo-ring:A virtual ring group number created for the transparent bridge group. The Token Ring side sends frames to this ring number to reach the host in the transparent bridge side.
bridge-number:A bridge number is assigned for the bridge between the virtual ring group and the pseudo ring.
tb-group:The transparent bridge group number configured with the bridge-group command.
To perform SR/TLB the router configuration of the network in Figure 4-25 is displayed in Example 4-8. The virtual ring group number is 10. From the source-bridge transparent 10 2 5 1 command, you can determine that transparent bridge group 1 uses pseudo ring 2, which is bridged to virtual ring 10. Bridge number 5 bridges between pseudo ring 2 and virtual ring 10.
Example 4-8 Example of SR/TLB
source-bridge ring-group 10 source-bridge transparent 10 2 5 1 ! interface tokenring 0 source-bridge 5 6 10 source-bridge spanning ! interface ethernet 0 bridge-group 1 ! bridge 1 protocol ieee
The RIF of a source route frame from Host A to reach Host B routes from Ring 5, Bridge 6, Ring 10, Bridge 5, to Ring 2. In hex, the RIF is 0830.0056.00a5.0020.
RSRB permits the bridging of Token Rings that are located on separate routers across non-Token Ring media. The routers are remotely connected through serial lines, Ethernet, or other methods. RSRB routers create the same virtual ring number that is a logical ring that encompasses both routers. There are four methods of transport:
- Frame Relay
- Fast Sequenced Transport (FST)
If you use direct encapsulation over serial lines, they need to have high-level data-link control (HDLC) encapsulation. Direct encapsulation can run over Ethernet, Frame Relay-ATM Interworking, or serial interfaces. If you use TCP encapsulation, RSRB traffic can be prioritized over WAN links.
The configuration of RSRB involves the regular SRB configuration with virtual rings. All routers in RSRB need to have the same virtual ring number configured. The additional configuration command is the configuration of source-route remote-peer statements. The configuration of every RSRB encapsulation is not presented in this subsection. Example configurations follow for direct and TCP RSRB encapsulation. More information on RSRB is at the site www.cisco.com/warp/public/701/2.html.
RSRB with Direct Encapsulation
Looking at Figure 4-26, Router A connects to Router B through a serial line. The RSRB configuration is shown in Example 4-9. The source-bridge remote-peer 10 interface serial 0 command defines the RSRB remote peer as being available through serial 0.
Figure 4-26 RSRB with Direct Encapsulation@
Example 4-9 RSRB Configuration with Direct Encapsulation
!Router A source-bridge ring-group 10 source-bridge remote-peer 10 interface serial 0 ! interface tokenring 0 source-bridge 1 5 10 source-bridge spanning ! interface serial 0 encapsulation hdlc !Router B source-bridge ring-group 10 source-bridge remote-peer 10 interface serial 0 ! interface tokenring 0 source-bridge 2 6 10 source-bridge spanning ! interface serial 0 encapsulation hdlc
The RIF of a source route frame from Host A to reach Host B routes from Ring 1, Bridge 5, Ring 10, Bridge 6, to Ring 2. In hex, the RIF is 0830.0015.00a6.0020.
RSRB with TCP Encapsulation
Looking at Figure 4-27, Router A reaches Router B through an IP network. With TCP, the local peer IP address is configured with one remote-peer statement, and the remote peer IP address is also configured with another remote-peer statement. The router can also be configured to acknowledge sessions destined for the remote peer by using the local-ack argument. The RSRB configuration is shown in Example 4-10. You use the shaded source-bridge remote-peer commands to define the local peer with the local IP address, and the remote peer with the IP address of the remote router.
Figure 4-27 RSRB with TCP Encapsulation@
Example 4-10 Configuration of RSRB with TCP Encapsulation
!Router A source-bridge ring-group 10 source-bridge remote-peer 10 tcp 192.168.1.1 source-bridge remote-peer 10 tcp 192.168.2.1 local-ack ! interface tokenring 0 ip address 192.168.1.1 255.255.255.0 source-bridge 1 5 10 source-bridge spanning !Router B source-bridge ring-group 10 source-bridge remote-peer 10 tcp 192.168.2.1 source-bridge remote-peer 10 tcp 192.168.1.1 local-ack ! interface tokenring 0 ip address 192.168.2.1 255.255.255.0 source-bridge 2 6 10 source-bridge spanning
The RIF of a source route frame from Host A to reach Host B routes from Ring 1, Bridge 5, Ring 10, Bridge 6, to Ring 2. In hex, the RIF is 0830.0015.00a6.0020.
DLSw version 2 is documented in RFC 2166 and DLSw version 1 is documented in RFC 1795. It was originally submitted by IBM to the IETF as RFC 1434 to overcome some of the limitations of SRB networks, especially in WAN networks. DLSw serves as a replacement of SRB and serves SNA data-link connections (DLC) and NetBIOS traffic.
Cisco's implementation of DLSw is called DLSw+. Some of the benefits of implementing DLSw+ are that link-layer acknowledgments and keepalive messages of SNA and NetBIOS traffic do not have to travel through the WAN. Also DLSw+ traffic can reroute around link failures and can be prioritized on WAN links, whereas SRB does not provide this capability. Cisco routers can be configured to communicate with devices running standards-based DLSw.
Figure 4-28 displays a sample DLSw network.
Figure 4-28 DLSw Network
End systems can attach to the DLSw network from the Token Ring, Ethernet, Fiber Distributed Data Interface (FDDI), Qualified Logical Link Control (QLLC) (X.25), and SDLC networks. Although there are other methods, the preferred method for establishing peer connections is by using TCP. After a connection is established, the peer routers can exchange capabilities. Circuits are established between end-systems (SNA and NetBIOS). NetBIOS names can be configured to prevent NetBIOS Name Queries to traverse the DLSw network; MAC addresses can be configured in the same manner to reduce SRB explorer frames.
DLSw uses the Layer-2 Switch-to-Switch Protocol (SSP) between DLSw peer routers. SSP switches frame at the SNA layer. DLSw also encapsulates frames into TCP/IP for transport over IP networks.
DLSw supports termination of RIFs at the local virtual ring, which is the default mode. DLSw local termination eliminates the requirement for link-layer acknowledgments and keepalive messages to flow across a WAN. With RIF termination, the virtual rings do not have to be the same between the peers, and all remote hosts appear to be connected to the local router's virtual ring. DLSw also supports RIF Passthru, which does not terminate the RIF at the local end. With RIF Passthru, the virtual ring numbers of the DLSw peer routers must match, and the rif-passthru keyword must be in the remote-peer statement.
SNA devices on a LAN find other SNA devices by sending an explorer frame with the MAC address of the target SNA device. When a DLSw router receives an explorer frame, the router sends a canureach frame to each of the DLSw partners. If one of its DLSw partners can reach the specified MAC address, the partner replies with an icanreach frame. The specific sequence includes a canureach ex (explore) to find the resource and a canureach cs (circuit setup) that triggers the peering routers to establish a circuit.
NetBIOS circuit establishment is similar, but instead of forwarding a canureach frame that specifies a MAC address, DLSw routers send a name query (NetBIOS NAME-QUERY) frame that specifies a NetBIOS name. Instead of an icanreach frame, there is a name recognized (NetBIOS NAME-RECOGNIZED) frame.
Most DLSw implementations cache information learned as part of the explorer processing (from the icanreach responses) so that subsequent searches for the same resource do not result in the sending of additional explorer frames. This is also true with DLSw+.
The basic configuration of DLSw is quite simple. Each router with attached networks is configured with a local -peer command. Remote peers are configured to remote routers. It is preferred to use the loopback IP addresses to configure the peers because the DLSw peers stay connected regardless of the state of local or remote LAN interfaces. For designs with multiple branches connecting to a hub site, the promiscuous keyword can be used on the hub router's local-peer command to permit remote peers to connect without having to explicitly configure each remote peer on the hub router. The configurations of the routers in Figure 4-29 are shown in this section.
Figure 4-29 DLSw Example Network
The configuration for Router A is displayed in Example 4-11.
Example 4-11 DLSw Configuration of Router A
source-bridge ring-group 10 ! dlsw local-peer peer-id 10.1.1.1 promiscuous dlsw remote-peer 0 tcp 10.2.1.1 ! interface loopback 0 ip address 10.1.1.1 255.255.255.255 ! interface tokenring 0 source-bridge 1 5 10
The configuration for Router B is displayed in Example 4-12.
Example 4-12 DLSw Configuration of Router B
source-bridge ring-group 20 ! dlsw local-peer peer-id 10.2.1.1 dlsw remote-peer 0 tcp 10.1.1.1 ! interface loopback 0 ip address 10.2.1.1 255.255.255.255 ! interface tokenring 0 source-bridge 2 6 20
The configuration for Router C is displayed in Example 4-13.
Example 4-13 DLSw Configuration of Router C
bridge-group 1 protocol ieee ! dlsw bridge-group 1 ! dlsw local-peer peer-id 10.3.1.1 dlsw remote-peer 0 tcp 10.1.1.1 ! interface loopback 0 ip address 10.3.1.1 255.255.255.255 ! interface Ethernet 0 bridge-group 1
Router A only defines a remote peer for Router B. It does not need to define Router C as a peer because the promiscuous keyword accepts connections from remote peers without having to define them. Router C defines Router A as a remote peer.
The RIF for a host on Router A to reach a host on Router B takes the path from Ring 1 to Bridge 5 to Ring 10. Because the virtual ring numbers are different, the RIF is terminated on virtual ring 10 on Router A. The RIF in hex is 0630.0015.00a0.
If the router configuration on Router B is changed to virtual ring 10 and DLSw is configured for RIF Passthru, the RIF reaches the far end ring. The RIF from the host on Router A to Router B takes the path from Ring 1 to Bridge 5 to Ring 10 to Bridge 6 to Ring 2. The RIF in hex is 0830.0015.00a6.0020.