In order to robustly characterize today's network traffic, a new multifractal theory based multiplicative cascade traffic model is proposed. The proposed multiplicative cascade traffic model is conservative in measure with its multiplier weights determined by Binomial Newton expansions. We also derive some major statistical characteristics of this new cascade process enabling generation of synthetic traffic used for traffic model validation and simulation purposes. Experimental investigation results have shown that the proposed model is promising in robust and accurate traffic characterization and modeling for both real wired and wireless traffic outperforming three well-known multifractal models in the literature.11511431155