Is Camens a Good Clustering Method?

Some methods like camins plus plus can work very well if you have a clear cluster structure with clusters of the same size and density. But real world data is sometimes not even visually clustered at all. And one application of this would be actually wector quantization to transmit a large data set of high dimensional wectors over a thin communication line. So in this i see breezing came si'm not trying to find any clusters, but just a set of sentrees that no one has ever found before.