Towards SDN-based Deterministic Networking: Deterministic E2E Delay Case

The explosion of the number of things connected to the Internet gave birth to a new set of services. Customers now are expecting next-generation networks to satisfy vertical applications with different requirements. For instance, 5G networks can carry these various demands by logically dividing the physical network into multiple slices, each slice features specific characteristics based on the type of service. Software-Defined Network (SDN) and Network Function Virtualization (NFV) introduced the term 'Network Softwarization' which is the main enabler of network slicing and 5G networks. Specifically, SDN separates the control plane from the data plane, this concept brings many benefits such as dynamism, flexibility, and innovation. However, when it comes to the assurance of Quality of Service (QoS), SDN is still behind. Not only SDN was not optimized for real-time communications, but also SDN networks cannot offer a deterministic End-to-End (E2E) delay. In this paper, we study OpenFlow-based communications with a focus on the delay. The modeling of queuing delays shows a stable linear development of the mean waiting time under a probability of 0.2 that 10 switches generate packet_in message. After that, the increase becomes exponential and thus hard to predict.