Supernet

Supernetting is synonymous with Classless Inter-Domain Routing (CIDR) although CIDR is rather just the concept that is implemented when subnetting or supernetting.

In Internet networking terminology, a supernet is a block of contiguous subnetworks addressed as a single subnet. Supernets always have masks that are smaller than the classful mask, otherwise it isn't a supernet.

Supernetting alleviates some of the issues with the original classful addressing scheme for IP addresses by allowing multiple networks address ranges to be combined, either to create a single larger network, or just for route aggregation to keep the "Internet Routing Table" (or any routing table) from growing too large.

For supernetting to work, you must be using static routing everywhere or be using a routing protocol which supports classless routing, such as RIPv2 or OSPF (or BGP for Exterior Routing) which can carry subnet mask information with the routing update. The older RIPv1 (or EGP for Exterior Routing) protocol only understands classful addressing, and therefore cannot transmit subnet mask information.

EIGRP is also a Classless Routing Protocol capable of support for CIDR or VLSM (Variable Length Subnet Masking). By default EIGRP will summarize the routes within the routing table and forward these summarized routes to its peers. This can be disastrous within heterogeneous routing environments if VLSM has been used with Discontiguous Subnets and therefore Auto-Summarization should be disabled unless VLSM has been carefully designed and implemented.

The family of Classfull Routing Protocols are RIPv1, and IGRP - these protocols can not support CIDR as they do not have the ability to include subnet info within the Routing Updates.

The Family of Classless Routing Protocols are RIPv2, OSPF, EIGRP and BGP. EIGRP can handle multiple Routed Protocols such as IPX and Appletalk. RTP (Reliable Transport Protocol) is used by EIGRP as it's layer 4 protocol as opposed to TCP. This keeps EIGRP Protocol Independent because RTP is not native to THE TCP/IP ip stack like TCP.

Another way to look at supernetting is:

Let's take a class B mask of 255.255.0.0 - If we borrow 2 network bits, the mask changes to 255.252.0.0, this is called Supernetting. If on the other hand we borrow two host bits, the mask changes to 255.255.192.0, this is called subnetting.

This same method can be used on class A and C addresses.

External links

* [http://www.firewall.cx/supernetting-chart.php The Supernetting/CIDR Chart]