IP routing specifies the manner in which an IP packet can be delivered from the host, where the data packet is created, to the destination host.

IP Routing
IP routing process consist of following steps -

1. When sending a packet, compare the destination IP address of the packet to the sending host’s perception of the range of addresses in the connected subnet, based on the host’s IP address and subnet mask.

a. If the destination is in the same subnet as the host, send the packet directly to the destination host.  ARP is needed to find the destination host’s MAC address.

b. If the destination host is not in the same subnet as the host, send the packet directly to the host’s default gateway.  ARP is needed to find the default gateway’s MAC address.

Routers use the following steps: the packet must first be received, whereas the sending host begins with the IP packet in memory.

1. For each received frame, use the data-link trailer frame check sequence (FCS) field to ensure that the frame had no errors; if errors occurred, discard the frame and don’t continue to the next step.

2. Check the frame’s destination data link layer address, and process only if addressed to this router or to a broadcast/multicast address.

3. Discard the incoming frame’s old data-link header and trailer, leaving the IP packet.

4. Compare the packet’s destination IP address to the routing table, and find the route that matches the destination address.  This route identifies the outgoing interface of the router, and possibly the next-hop router.

5. Determine the destination data-link address used for forwarding packets to the next router or destination host as directed in the routing table.

6. Encapsulate the IP packet inside a new data-link header and trailer, appropriate for the outgoing interface, and forward the frame out that interface.

Brief summary of the routing process: Router receives a packet, matches the packet’s destination address with the routing table, and forwards the packet based on the matched route.

On point-to-point links, an ARP table is not needed.  A point-to-point link can have at most one other router connected to it, so you can ignore the data-link addressing.  With Frame Relay, the routing process does consider the data-link addresses, called data-link connection identifiers (DLCI).

IP Addressing

- Unicast IP addresses are IP addresses that can be assigned to an individual interface for sending and receiving packets.

- Each unicast IP address resides in a particular Class A, B, or C network, called a classful IP network.

- If subnetting is used, which is almost always true in real life, each unicast IP address also resides in a specific subset of the classful network called a subnet.

- The subnet mask, written in either dotted decimal form or prefix notation form (/24), identifies the structure of unicast IP addresses and allows devices and people to derive the subnet number, range of addresses, and broadcast address for a subnet.

- Devices in the same subnet should all use the same subnet mask; otherwise, they have different opinions about the range of addresses in the subnet, which can break the IP routing process.

- Devices in a single VLAN should be in the same single IP subnet.

- Devices in different VLANs should be in different IP subnets.

- To forward packets between subnets, a device that performs routing must be used.  In this book, only routers are shown, but multilayer switches – switches that also perform routing functions – can also be used.

- Point-to-point serial links use a different subnet than the LAN subnets, but these subnets only require two IP addresses, one for each router interface on either end of the link.

- Hosts separated by a router must be in separate subnets.