The Layer 2 and 3 parts concern you the most. When a host wants to talk to another host and they are not on the same logical IP network, the host establishes a Layer 2 session between its MAC address and the MAC address of the router. The host passes its Layer 3 packet to the router. The router looks at it, determines where it needs to send it next, and passes it on to the next device, establishing a Layer 2 session with it. This process continues until the packet is received by Host B, who sends back an acknowledgment. The process starts over, but in the reverse direction. These sessions stay up until the host ends the conversation or there is some kind of unrecoverable error.
To put it another way, at Layer 2, Host A only talks to Router A; but at Layer 3, Host A is talking to Host B. To understand what a bridge does, replace the hub (in the figure) with a bridge. The Layer 2 session terminates at the bridge, and there will be a lot more Layer 1 sessions.
The most interesting thing about the preceding scenario is that in establishing the various sessions, the routing protocol never came into play because the routing protocol has nothing to do with setting up a session. The routing protocol keeps tables of where to send the routed packet next, and it always hands the packet off to the next device to decide what to do with the packet next. OSPF might confuse this a little because it does interact with other layers, but it still hands the packet off to the next device to decide what to do next.
Understanding Router Subinterfaces
One of the most difficult concepts to understand is the difference between point-to-point and multipoint interfaces on a router. This section will briefly discuss the different scenarios regarding the use of each.
A routers two different types of subinterfaces provide a flexible solution for routing various protocols over partially meshed networks. A single, physical interface can be logically divided into multiple, virtual subinterfaces. The subinterface might be defined as either a point-to-point connection or a multipoint connection.
The concept of subinterfaces was originally created to better handle issues caused by split horizon over nonbroadcast multiaccess (NBMA) networks (such as Frame Relay and X.25) and distance-vector based routing protocols (such as IPX, RIP/SAP, and AppleTalk).
Split horizon dictates that a routing update received on an interface cannot be retransmitted out onto the same interface. This rule holds even if the routing update was received on one Frame Relay PVC and destined to retransmit out onto another Frame Relay PVC. Assuming a Frame Relay setup of sites A, B, and C, this would mean that sites B and C can exchange routing information with site A, but would not be able to exchange routing information with each other. Split horizon does not allow Site A to send routing updates received from Site B on to Site C, and vice versa.
By dividing the partially meshed Frame Relay network into a number of virtual, point-to-point networks using subinterfaces, the split horizon problem can be overcome. Each new point-to-point subnetwork is assigned its own network number. To the routed protocol, each subnetwork now appears to be located on separate interfaces. Routing updates received from Site B on one logical point-to-point subinterface can be forwarded to site C on a separate logical interface without violating split horizon.
Cisco serial interfaces are multipoint interfaces by default unless specified as a point-to-point subinterface. Though less common than point-to-point subinterfaces, it is possible to divide the interface into separate virtual multipoint subinterfaces.
Multipoint interfaces/subinterfaces are still subject to the split horizon limitations, as discussed previously. All nodes attached to a multipoint subinterface belong to the same network number. Typically, multipoint subinterfaces are used in conjunction with point-to-point interfaces in cases in which an existing multipoint Frame Relay cloud is migrating to a subinterfaced point-to-point network design. A multipoint subinterface is used to keep remote sites on a single network number while slowly migrating remote sites to their own point-to-point subinterface network. Eventually, all remote sites can be moved to their own point-to-point subinterface networks and the multipoint subinterface will not be necessary.
The OSI model as discussed in Chapter 1, Foundations of Networking, was shown to be a framework for communication but was not to be itself a method of communication in the sense that the model could pass data. The actual communication between two devices, whether they are hosts or routers, is implemented by a protocol.
A protocol is a formal description (documented in RFCs) of a set of rules and conventions that govern how devices on a network exchange information. There are two basic types of protocols that will be discussed, and although there are others, they are beyond the scope of this book. If you are interested in reading on these other protocols, please refer to the bibliography, as many of the books referenced there cover multiple protocols in more depth. The two basic types of protocols that should concern you for the purposes of this chapter are as follows:
You now know why a protocol is needed and what a protocol actually is. You have also determined the difference between the two main types of internetworking protocols, routed and routing.