After ordering your line, you waited your 60 days; you sacrificed a couple of old ISA network cards to the circuit gods; and lo and behold, your circuit is ready a few days early. Now that you have a line, what do you need to know to talk to another location? DLCIYou need a DLCI to talk to the service provider's switch. A DLCI is basically a pointer for the switch telling it where this particular packet needs to go. Each switch at the service provider's central office (CO) has a configuration the switch can access. This configuration is like the address on an envelope. The DLCI says, "This packet needs to go to port 8, switch 32, at CO 68 for service provider 3," or wherever your other end is. As the packet moves through the service provider's WAN cloud, the intervening switches just need to look at the service provider information and pass the packet along. Once the packet is at the target service provider, the switches forward the packet to the correct CO. This description is an oversimplification, but this book is not designed to discuss Stratacom switch, the Cisco telco line of products, or methods of operation. You get the DLCI from the service provider. How you use it depends on how you configure your router, which is discussed later. LMILMI is a signaling standard between your router and the CO switch. LMI is responsible for making sure both devices know the other is there. In addition to acting as keepalives, LMI also acts as a form of Cisco Discovery Protocol (CDP). LMI can provide the router with its DLCI number and IP information regarding the device on the other side of the cloud.
IOS version 11.2 enabled routers to autosense the LMI type the Frame Relay switch is using. With versions prior to 11.2, you have to configure the LMI type. There are three options, and the thing to remember is that you want to use the LMI type used by the Frame Relay switch:
EncapsulationYou need to specify Frame Relay encapsulation on the interface, but there are two types, Cisco and IETF. If you do not specify, the router defaults to Cisco encapsulation. This default causes problems if you are not connecting to a Cisco router on the other side of the service provider's cloud. If you are connecting to a non-Cisco router, make sure you specify IETF. MappingThe router must understand that if it needs to send packets to a given destination, it needs to use the Frame Relay connection. You convey this understanding by mapping a Layer 3 address to the DLCI. Think of where a static route says, "To get to network www.xxx.yyy.zzz, I need to send packets out interface serial 1." The mapping works the same way, except it says, "To get to device www.xxx.yyy.zzz, I need to use Frame Relay DLCI xxx." You can establish mappings manually or automatically. Both ways require you to tell the interface what DLCI it is connected to, but Inverse-ARP is enabled by default. It takes the LMI information and sends a query across the service provider cloud to find out the address of the device on the other end, and then it creates a mapping. Frame Relay ConfigurationYou can use several commands to establish Frame Relay. Dynamic address mapping is enabled by default. If, for some reason, you need to, you can manually configure the map with the following commands. Table 9.1 explains each part of the commands:
Suppose a company has two routers it wants to connect via Frame Relay, router A and router B. Router A's IP address on its Frame Relay interface is 10.1.2.1, and it connects to the local CO switch with DLCI 100. Router B's IP address is 10.1.2.2 using DLCI 200. The company wants to use static mappings and provide for routing updates and for compression. The following line is the router A example: frame-relay map ip 10.1.2.2 100 broadcast payload-compress packet-by-packet The next line is the router B example: frame-relay map ip 10.1.2.1 200 broadcast payload-compress packet-by-packet What you must remember is that the DLCI is a local identifier between the router and the telco switch. Because the DLCI is local, both routers can be using the same DLCI something that people often don't realize. The other thing to remember is that you need to map the remote interface address to the local DLCI address. |