Routing Table

Static Routing is a static way to configure the #Routing Table of a Router#. It is static as not reactive to external changes such as topology changes. Usually, this is done manually by the network administrator in order to have some sort of control over route configuration. One particular example is using static route in a stub network, where there is only one route for the network to be accessed. We could simplify the network routing by using route summarisation which represents all similar routes with a single network (using network class’s default route# but smaller subnet mask is possible) if the routes share a single exit interface.

Routing Table# Principles, written by Alex Zinin, stated that:

The determined routing path will be later stored in the #Routing Table of the #Router. If there is a topological change in the network, the protocol will update the routing table accordingly and advertise it to its neighbours.

RIP only propagate the network address of known routes. There is no need to include subnet mask as part of the routing information since it could determine the subnet mask by just looking at the Address Classes default subnet masks. The boundary Router (the entry router to the RIP network) will summarise the RIP subnets from the internal network to the external network. This reduces the size of the Routing Table and resulted in a faster lookup in the routing table (single route can represent multiple routes). However, it comes at a cost of not able to support discontiguous network. Furthermore, if the subnet mask doesn’t match the class or a subnet of the class, RIP will not advertise the route to the other RIP enabled routers.

Router is a #Network Device operates at Network Layer# which forwards packets from source to destination based on network layers protocols (reading the source and destination IP Address). It will communicate with other routers using Routing Protocol# (dynamic routing) or be manually edited by network administrators (static routing) to maintain or update their Routing Table# that contain optimal routing paths. Of course Router needs to be connected to a network to establish either dynamic routing or static routing.

The physical structure of Router mainly consists of CPU, RAM, ROM, Flash, NVRAM and interfaces or ports. RAM not only stores the Routing Table#, but also holds Address Resolution Protocol (ARP) cache and fast-switching cache, performs packet buffering, and provides temporary memory for the running configuration file. NVRAM provides permanent storage for startup configuration file. Various versions of #Cisco Internetwork Operating System (IOS) are stored in Flash. ROM is where the power-on self test (POST) diagnostic software and bootstrap program (boot up partial IOS) located.

The process of forwarding the packet from its received port to the outgoing port is called switching. Despite the naming, Router will immediately drop the packet when it can’t figure the destination in its routing table maintained by itself. Otherwise, it will simply modify the destination Media Access Control Address (MAC) (the next-hop, can be another Router or the final destination) in the packet without touching the IP Address which defines the final destination. Though the primary forwarding decision will be based on the Routing Table# or routing information, assuming there is a mapping of logical address to its interface or port. If there is no path defined in the routing table to where the packet should be forwarded to, Router will redirect it to a default route if set (usually 0.0.0.0 in IPv4 Address and ::/0 in IPv6 Address), or simply discards it and sends Internet Control Message Protocol (ICMP)# Unreachable message to the source. Additionally, we can divide a physical interface into at most 65,535 logical interfaces using the command interface {interface-name} {interface}.{subinterface}. This is particularly useful in Virtual LAN (VLAN) routing.

Recursive Route Lookup occurs when the #Router has to perform multiple lookups in the #Routing Table before forwarding the packet.

Using Equal Cost #Load Balancing, the Router will send packets over multiple exit interfaces listed in the Routing Table.

In Distance Vector Routing, the #Routers share their Routing Table with their direct neighbour(s) (often termed “routing by rumour”) periodically. The table stores routing information which describes the direction (router and the corresponding port) where the packets should be forwarded to and the distance using one of the metrics (See in #Routing Protocol). When a router receives the information via this method, it will undergo a process called distance accumulation, that is, increments the distance by one. By this way, routers within the network are ignorant to its topology. The process is usually done by the Bellman-Ford algorithm.

Default Route is a route that will match all packets when the packet’s destination IP Address# is not known. It reduces the size of the Routing Table#. Typically, Default Route is used to connect edge Router# to Internet Service Provider’s (ISP)# network.

Convergence is a state when all router’s routing table are at a state of consistency. The convergence time is the time it takes to routers to share information, calculate the optimal path, and update their Routing Table.

By using CIDR, we could reduce the size of the Routing Table as the supernet (network that is at a higher hierarchy) is able to summarise multiple network addresses with a subnet mask that is less than the classful subnet mask. Classless Routing Protocols# will be used for routing update.