Centralized QoS Routing Using Network Calculus for SDN-Based Streaming Media Networks

Abstract

Streaming media transmission requires strict quality of service (QoS) parameters such as maximum delay and delay jitter. An effective streaming media routing algorithm is a key factor in ensuring QoS. The existing solution only considers a single parameter indicator in the performance parameters such as bandwidth, delay, and utilization of the link, and fails to comprehensively measure the data flow in the network. It is not possible to comprehensively measure the relationship between the business attributes and the QoS parameters. Firstly, the deterministic upper bounds of QoS parameters in streaming media networks are solved by using network calculus theory, and the QoS parameters are normalized, and a multi-constrained QoS resource allocation model is established; the separation of control and forwarding planes is defined by using software-defined networking (SDN) to deploy the multi-constrained QoS resource allocation model in the control plane; the QoS routing system of streaming media network based on the SDN is designed and implemented, including flow table scheduling model, routing function, measurement and forwarding modules. In the routing function module of the SDN controller, a multi-constrained QoS routing algorithm based on network calculus is implemented. Experimental results show that the proposed multi-constrained QoS resource allocation model based on network calculus and the multi-constrained centralized QoS routing algorithm based on the SDN have good performance. INDEX TERMSSDN, network calculus, streaming media, multi-constrained QoS, centralized QoS routing. I. INTRODUCTION A. MOTIVATION: QOS OF STREAMING MEDIA

Energy Efficiency in Software Defined Networking: A Survey

“…In centralized model, a single central controller takes the overall decision of the network [43]. Being the single point of decision maker, the central controller globally monitors the network despite of its heterogeneous underlying network architecture.…”

“…Similarly, in [50], the discussed model comprises of two-layer control architecture. On the other hand, Maple [40], DMCQR [43] architecture follows centralized model.…”

“…Maple. [40,42] Centralized model Multi core scheduler is used to record reusable decisions 6TiSCH [42,43] Centralized model 6TiSCH and DetNet is used for Low Power Wireless Sensor Networks (WSNs) DMCQR [43,44] Centralized model Multi-constrained QoS resource allocation model Kim et al [44,45] Centralized model Network QoS control framework for converged fabrics Maestro [45,46] Distributed Model Simple programming model with additional throughput optimisation technique Procera [46,47] Distributed model A control architecture for software-defined networking (SDN) using a declarative policy language Hyperflow [47,48] Distributed model An event-based control plane Onix [48,49] Distributed model A platform for distributed system Devolved [49,50] Distributed Model Use of multiple independent controllers Kandoo [50,51] Distributed model Two-layer controller architecture Fig. 1 Traditional network architecture vs. three-layered software-defined networking architecture SDN controller works with the four interfaces: Northbound [56], Southbound [57], Eastbound [58] and Westbound [59] interfaces.…”

Terminal handover in software-defined WLANs

“…The emergence of software-defined networking (SDN) [3][4][5][6][7][8][9] provides a new way to thoroughly solve this problem. Its main idea is to decouple the control function of the forwarding function of the network and hand over the control to the centralized controller.…”

Research on semi-supervised multi-graph classification algorithm based on MR-MGSSL for sensor network

“…According to the characteristic of subgraphs, the multi-graphs are expressed through vectors and then classify the models by the existing semi-supervised learning methods. And it is summarized as two multi-map feature subgraph measurement model with label and without label [29].…”