Link loss inference algorithm with minimal cover set and compressive sensing for unicast network measurements

To reduce the probe cost and improve the accuracy of the link loss inference, a novel algorithm under network tomography framework is proposed. The number of end-to-end paths is reduced by using minimal cover set measurements. Meanwhile, the accuracy of the link loss inference is improved by the implement of solving linear equations and compressive sensing techniques. Taking into account the constraints in compressive sensing theory, an approach for constructing a novel network tomography model that obeys the constraints of compressive sensing is developed. Simulation results show that this algorithm can obtain a higher accuracy with less end-to-end measurement paths.