This thesis proposes a novel routing approach for Software-Defined Networks (SDN) that takes into account the cumulative load of flows on the links. In the proposed scheme, the knowledge of network topology and traffic is leveraged to select forwarding path. The proposed method enhances the network utilization, through which the average end to end delay and congestion probability are improved too. Moreover, the algorithm also considers the differences between two direction of a link, i.e. capacity, load. Simulation using Mininet confirms that the efficiency of proposed method outperform with conventional shortest path routing in the terms of end to end delay, bandwidth. Further, the implementation in OpenIRIS controller using specific designed route caching solution made an impressive enhancement in controller¡¯s computational performance.