Achieving Real-Time Quality of Service with Software Defined Networking

In traditional networks, Quality of Service (QoS) has been a challenge due to the lack of global knowledge and a large amount of data to process at each device. This issue becomes much harder when we want to manage realtime QoS. As an innovative network paradigm, Software Defined Network (SDN) enables a centralized network control and a global view of network states. Meanwhile, the online algorithm can accommodate with unknown inputs in any coming sequence and provides a worst-case guarantee for the performances even compared with the off-line optimal result. In this paper, we utilize SDN and online algorithms to achieve real-time QoS in networks. By applying Primal-Dual techniques and the potential function method with phase partitions, we design two online algorithms that achieve real-time QoS and network load balancing at the same time with competitive ratios of $log(D)$ and $log(n)$ respectively ($D$ is the network diameter and $n$ is the node number). The simulations also support our theoretical analyses.