Baselining Network-Wide Traffic by Time-Frequency Constrained Stable Principal Component Pursuit

In this letter, we develop a novel network-wide traffic analysis methodology, named Stable Principal Component Pursuit with Time-Frequency Constraints (SPCP-TFC), for extracting the baseline of a traffic matrix. Following a refined traffic matrix decomposition model, we present new timefrequency constraints to extend Stable Principal Component Pursuit (SPCP), and design an efficient numerical algorithm for SPCP-TFC. At last, we evaluate the SPCP-TFC method by abundant simulations, and show it has superior performance than other traffic baseline methods such as RBL and PCA.