A New Network Topology Evolution Generator Based on Traffic Increase and Distribution Model

Following traffic increase near to network backbone capacity in a domain network, the evolution of backbone topology takes place. In this paper, three main types of evolution methods: link upgrading only method (upgrade link capacity only), node upgrading method (add nodes and links with the same link capacity), and the combinations of previous two are explored. To shunt the saturated traffic efficiently, we propose several node upgrading generators - Adapting Traffic Evolution Topology gEnerators (ATETEs). ATETEs consider traffic distribution and node burden conditions and select the congested link as the main shunting object. The difference among ATETEs lies that they have diverse strategies to choose a node in the congested link as the second shunting aim. Simulation shows that: the type of ATETEs which picks the burdened node in the saturated link as another shunting object has the most effective shunting result. Compare with most of current topology generators such as [BRITE] in [1] which consider little about traffic increase distribution, ATETEs are more efficient to satisfy traffic increase.