A New Topology Aggregation Algorithm in Hierarchical Networks

Large scale networks are usually hierarchically structured for the reason of scalability and security. Topology aggregation is introduced for the state information reduction of domains, but it also brings distortion of the original state information. The tradeoff between information reduction and distortion is the key to topology aggregation. In this paper, a new topology aggregation algorithm called SA approach is presented for hierarchical networks with delay and bandwidth constraints. This approach uses an approximation staircase generated by a line segment to represent the QoS supported area for the purpose of representing the parameters of a logical link in the full-mesh; furthermore, the spanning tree and star topology are employed to compress the full-mesh. Experimental results show that the new topology aggregation method can reduce more information distortion and achieve a better tradeoff between information reduction and accuracy than existing approaches.