The massive adoption of Cloud services has led to the explosion of traffic transiting over the Cloud infrastructure. Such an impressive evolution of data demand will inevitably be the catalyst of Operator infrastructure transformation. In this context, Software Defined Networking (SDN) is the technology that is shaping the future of carriers' networks. SDN considerably reduces the complexity of managing the network infrastructure while providing tremendous computational power compared to legacy devices. In this paper, we address the resource allocation issue in Wide Area Networks (WAN) while considering the requested QoS. To do so, we design an SD-WAN architecture to enhance the network resources allocation and hence improve the QoS of distributed applications. We formulate first the path computation problem as an Integer Linear Program while taking into consideration both network application requirements and the network occupation status. The problem is then resolved in a polynomial time leveraging the Branch-and-Cut algorithm. Results obtained with our experimental platform, show that the proposed SD-WAN framework outperforms the most prominent related solutions in terms of applications' satisfaction level and consumption of network's resources.