Most of network operation/management tasks such as traffic accounting, traffic engineering, and network design require an accurate and timely network Traffic Matrix (TM). A TM presents the network traffic volume between origin and destination (OD) nodes. Many network management methods assumed that an accurate traffic matrix is given by other network entities, but this is not true. To apply those methods in the practice, the first priority challenge is to obtain an accurate and timely TM. Another issue to obtaining TM is to estimate the change of TM in the near future, not just reporting historical traffic measurement results. Meanwhile, Software Defined Networking (SDN) paradigm has attracted a significant interest from industry and academia as a future network architecture. Within SDN environment, many advantages exist compared to traditional IP based network. To acquire an accurate and timely TM for SDN, we propose FLAME, a TM estimation method based on Poisson Shot Noise process. FLAME is designed to take advantages of the SDN paradigm, and it reflects network traffic characterics revealed by current measurement studies. To evaluate FLAME, we compared the estimation results with a real data center network traffic trace. The evaluation results show that the estimated TM is 66.36% similar compared to the measured traffic.