## CS Network Routing and Forwarding Notes ### Network Layer Functions The network layer (IP layer) handles host-to-host packet delivery. This contrasts with the transport layer (process-to-process) and the data link layer (node-to-node). Two key functions define the network layer: * **Forwarding:** Moving packets from a router's input link to the appropriate output link. This is a *local* function within each router. Think of forwarding as the "post office" aspect: getting the letter to the right outgoing mailbox. * **Routing:** Determining the path a packet takes from source to destination. This is a *network-wide* function. Routing is the "postal service" aspect: figuring out the entire route from sender to receiver. These two functions are implemented in two planes: * **Data Plane:** Handles the local forwarding decisions within each router. A packet arrives on interface GigabitEthernet0/0. The router's data plane uses the routing table to determine that the packet should be forwarded out GigabitEthernet0/1. * **Control Plane:** Handles the network-wide routing decisions, determining the overall path. The control plane uses routing protocols to exchange information about network topology and select the best path between two hosts. ### Routing Table Maintenance Routers maintain routing tables to make forwarding decisions. These tables can be populated in two ways: * **Static Routing:** Manually configured by a network administrator. This is simple for small, stable networks but becomes cumbersome and error-prone in larger, dynamic networks. * **Pros:** No extra processing overhead, no bandwidth consumption for routing updates, enhanced security. * **Cons:** Requires deep network knowledge, manual adjustments for topology changes, error-prone in large networks. * **When to use:** Predictable traffic, simple network design, stub networks (networks connected to only one router). * **Types:** * **Standard:** Specifies a specific destination network and the next hop. * **Default (Gateway of Last Resort):** A route matching all packets, sending them to a default gateway (e.g., 0.0.0.0/0). Used when no other route matches the destination IP address. This is like having a "return to sender" address for unknown destinations. Misconfiguring static routes can lead to network connectivity issues. Careful planning and testing are crucial. * **Dynamic Routing:** Routers automatically exchange network status updates using routing protocols. This adapts to network changes automatically. * **Purpose:** Discovering remote networks, maintaining up-to-date routing information, choosing the best path, adapting to path failures. * **Roles:** Dynamically sharing information about remote networks, determining the best path, adapting to topology changes. ### Routing Protocols Routing protocols are categorized into two main types: * **Distance-Vector Routing Protocols:** Routers share information only with their immediate neighbors. Each router only knows the distance (metric) to its neighbors and the next hop to reach other networks. They don't have a complete view of the network topology. * **Characteristics:** Share updates with neighbors, only aware of directly connected networks and neighbors, unaware of overall network topology. * **Examples:** RIPv1, RIPv2, IGRP, EIGRP. * **RIP (Routing Information Protocol):** RIPv1 broadcasts updates every 30 seconds; RIPv2 uses multicast addressing and supports authentication. Distance-vector protocols are simpler to implement but can suffer from slow convergence (adapting to changes) and the "counting to infinity" problem. * **Link-State Routing Protocols:** Routers build a complete map of the network topology. They share link-state information with all other routers in the network. They use algorithms like Dijkstra's shortest path first (SPF) to determine the best path. * **Characteristics:** Build a complete network map, use SPF algorithm, more complex but faster convergence. * **Update Process:** Each router learns about its links, exchanges information with neighbors, builds Link-State Packets (LSPs), floods LSPs to all neighbors, builds a complete topology map, and computes best paths. * **Metric:** A value assigned to routes to evaluate their "cost" (e.g., hop count, bandwidth, delay). The protocol selects the path with the lowest metric. ### Summary of Key Differences | Feature | Distance-Vector | Link-State | |-----------------|-------------------------------------------------|-----------------------------------------------| | Topology View | Partial (only neighbors) | Complete | | Information Sharing | With neighbors only | With all routers | | Convergence Speed | Slower, prone to instability | Faster, more stable | | Complexity | Simpler | More complex | | Examples | RIP, IGRP, EIGRP | OSPF, IS-IS | This concludes the study notes. Remember to review the concepts and examples to solidify your understanding. Practice applying the concepts to different network scenarios.