Research On SDN-based Transmission Optimization Over Mobile Networks

In order to cope with the explosive growth of mobile network traffic and meet the needs of various new business and application models,the researchers have proposed a series of new mobile network architecture based on software defined network.They hope to leverage the advantages of network control and forwarding separation,logical centralized control,to simplify the management and control of mobile network,shorten the deployment period of new business and reduce network operation cost.However,under the new architecture,there are still many challenges in how to effectively manage the network traffic and improve the users' experience.To address these challenges,this thesis focuses on the core objectives such as network-side traffic optimization,transmission optimization of computation-intensive mobile application and traffic-intensive mobile application.With these goals,this thesis focuses on 3 specific research issues,including SDN-based cooperative caches,SDN-based cooperative offloading and SDN-based streaming media transmission optimization.1.For the network-side traffic optimization,this thesis propose a novel SDNbased Cooperative Cache Network(SCCN),which through the deployment of SCCN Controller to capture the popularity changes of contents based on the proposed incremental record mechanism and place contents among cache nodes optimally.We formulate the optimal content placement as a multi-commodity facility location problem and prove its NP-hardness,which aims to jointly optimize the transmission delay and the external traffic.We design Relaxation Algorithm(RA)to solve the problem with an approximation ratio of 1/2 in the worst case.To solve large scale problems efficiently,we further design Heuristic Algorithm(HA),which can find a near-optimal solution with three orders of magnitude speedup compared to RA.We implement a prototype based on Open vSwitch and conduct extensive simulations based on real traces to show the feasibility and effectiveness of SCCN.2.For the computation-intensive mobile applications,this thesis designs a software-defined Energy-Efficient Cooperative Offloading Model(SDCOM)to provide a centralized control in distributed mobile cloudlet system.We formulate the task scheduling as an 0-1 knapsack optimization problem,which aims to minimize the energy consumption of mobile devices under the external traffic constraint from the perspective of the entire network.To compute the optimal solution as a benchmark,we design the Conditioned Optimal Algorithm(COA).Due to its exponential time complexity,COA is practical only for a moderate number of task types.We further propose the Energy-saving Greedy Algorithm(EGA),which has a polynomial-time complexity and achieves a good performance to the optimal solution.Based on the proposed traffic queuing mechanism,we design an Online Task Scheduling Algorithm(OTS)to address the problem in the online setting.Moreover,we extend OTS to design a Proportional Fair Online Task Scheduling Algorithm(PF-OTS)based on the concept of Relative Energy Consumption.It can not only exhibits similar performance to OTS,but also ensures the fairness of energy consumption among mobile devices,but also ensures the fairness of energy consumption among mobile devices.We conduct extensive trace-based simulations to show the effectiveness of SDCOM under various network conditions.3.For the traffic-intensive mobile applications,this thesis design an SDN-based SVC scheduling strategy for streaming transmission optimization.The strategy calculates the skip probability of the users' viewing video based on the user-requested behaviors recorded by the SDN Controller,and manages the users' cache based on the skip probability,which aims to maximize users' QoE and minimize the users' wasted traffic.We formulate the chuck scheduling as a Precedence-Constrained Knapsack Problem and prove its NP-hardness.Then RHC algorithm is designed based on Receding Horizon Control theory.We conduct extensive trace-based simulations to show that RHC can save traffic 30% for mobile users.