Gateway-to-Gateway (GGP)

This protocol was implemented in the Internet from its beginning [10]. The
original architecture consisted of small LANs connected to the backbone of
the Internet network, then called Arpanet.
The communication between this network and the LANs was carried
out through a few gateways controlled directly by the Arpanet manager. Each
gateway knew the routing information for all possible destinations and interchanged
its tables with other neighboring gateways to maintain information
coherency, using the GGP.
The GGP is based on the distance vector and the information
exchanged consisted of pairs (network address, distance from the gateway to this
network). The distance is measured in intermediate gateways (i.e., zero for a
directly connected network with an increase of one for each intermediate
gateway crossed in reaching the destination network). This distance vector–
based method has the drawback that it does not consider the real
capacity of each network. For example, a path through four Ethernet networks
interconnected through three intermediate gateways can be shorter
than a path through two networks connected by a low capacity serial line.
However, it is possible to take this fact into account by artificially increasing
the distance vector depending on its characteristics.