A Network Coordinate System Constructing Algorithm Based on Optimal Neighbor Nodes

Network coordinate systems (NCS) are often used to predict network latency. For the problem that there are large errors in existing classical NCS based on selecting neighbor nodes randomly, a NCS constructing algorithm based on optimal neighbor nodes is proposed. The algorithm firstly builds the NCS by randomly selecting neighbor nodes and calculate the prediction delay corresponding to different neighbor nodes. Secondly, the relative error of the NCS is calculated according to prediction delay and measured delay. Thirdly, the neighbor nodes with the smallest relative error are selected as the optimal neighbor nodes. Lastly, the coordinates of unknown nodes are determined using the optimal neighbor nodes. Experimental results based on PlanetLab nodes show that the prediction delay of proposed algorithm is more accurate compared with that of the NCS constructed by selecting neighbor nodes randomly.