A sub-network or a “subnet” is by definition a ‘smaller’ network. The word describes a specific network of nodes, computers or other devices that have a common designated IP (Internet Protocol) address prefix for router direction. The act of grouping and breaking down the network into smaller, more manageable pieces or “subnets” is called subnetting. This technique is applied as a way of “dividing and conquering” whereby a unicast address prefix is allocated among the several subnets of an enterprise’s network. Subnetting can be done for a variety of reasons apart from making the network more controllable, chief among them is to cut down on excessive rates of packet collision in Ethernet networks and to localise management of the entire nodal plane.
Subnets in a large network, such as a company office’s building’s network are usually arranged hierarchically – in a tree like structure. The first three digits in the IP address are usually reserved for the entire company’s prefix, while the second three might be for a specific department or wing and the third and fourth may constitute the departmental office’s specific node’s address. Routers used in such a hierarchical network are very easy to configure and manage since the network itself has an identifiable schema of addresses.
Subnetting results in an efficient network as fewer IP addresses are wasted in allocation, and all the routers can be configured to locate the node without wasting too much time. Subnetting is also required when the same network uses two different types of LAN connectivity options, i.e. Ethernet and Token Ring, or when there is need to manage the bandwidth requirements of the hosts and to separate the high bandwidth needing clients and the low bandwidth needing clients to better consolidate and stream packet speed into the clients needing more bandwidth.
Subnetting is not without its list of problems, however. Ideally, subnetting is to be done before setting up the network complex as in vitro changes are hard to execute as it is not a one time, fluid process when there is work going on.
Secondly, subnetting increases router complexity by adding as many as one router per node, thus increasing cost. To avoid this, an articulate design of the network must be come up with, or an easier routine must be applied before going head on into subnetting. It is only useful with newer routers or those supporting netmasks – older routers do not work well with subnetting.