Research On Key Technologies Of Software Defined Content Delivery Networks

Content delivery network(CDN)has been drawn much attention from both industries and academies.The key characteristic of CDN is that the content is closely to the users,so that the delay could be decreased when the users fetched the content.Nowadays,the users have two choice to distribute their content,including deploying own content delivery service or renting the service.The development of CDN has been innovated and reformed in terms of the architecture,mechanism and algorithms,so as to provide the ability to adapt the variety of data type and transmission requirement.With the progress of cloud computing and software defined network,content delivery networks has been improving through holding the new technologies.In this dissertation,the Software-defined Content Delivery Networks(SDCDN)combines the advantages of Software-defined networking(SDN)technology and CDN,and utilizes the characteristic of individual centralized controller,dynamic programming and so on.Through integrating the resource and making the decision with the information of the underlying network and the application layer,the decision with the comprehensive view is in favor of the lowlatency,fine-grained and user-oriented services in the large-scale cases.The architecture of this dissertation is based on the SDCDN.First,we describe the evolution of the SDCDN,and analyze the workflow of SDCDN.In the meantime,we propose some critical strategies based on the SDCDN,and then we take full advantage of the ability of information collection and management.From the aspect of improving the performance of content delivery,we mainly make the research of the placement problem,pricing problem and updating problem through the theoretical analysis,modeling and experimental proof.The main contributions of this dissertation include:(1)With regard to the large volume of traffic,the state of the network is unstable and the delay is gradually increasing,so as to make quality of service decline.In this dissertation,we propose Cloud Co-CDNs(C3),which a prototype that transfers the management part of edge networks to the cloud with the software-defined network,and the cooperation among providers of edge networks is considered in the prototype.Next,we design a stochastic model for describing the workflow in the C3 and develop a heuristic algorithm for replica placement to trade off among the latencies that are produced by the delivery process,the updating process and the replication process.(2)With regard to the mobility of users,we focus on the mobile edge network based on the software defined content delivery network further.We propose Provision for Mobile Edge Computing(PMEC),a prototype that takes advantage of storage devices with flexible placement.In PMEC,we accommodate various considerations and select different storage devices to cache,deploying the cache with the relationship of a two-tiered structure in the MEN.Thus,we construct a flexible overlay network with the objective of minimizing the cost in a two-tiered edge network.Based on the analysis of the problem,we solve the twotiered placement from bottom to top using the dynamic minimal spanning tree(MST)algorithm,and design two algorithms for each tier including the basic algorithm and the improved algorithms to decrease the delay.The simulation is conducted on realistic data to demonstrate the performance of our algorithms.(3)With regard to the mismatch between the quality of service and pricing,we focus on how to guarantee the quality of service through the automatic market-driven approach.We propose Dynamic Pricing with Traffic Engineering(DPTE),a prototype that generates traffic distribution using a market-driven model.We design the pricing model and traffic engineering model in DPTE.The target is to make the system not only rely on objective properties to reflect the price but also utilize pricing rules to influence the choice of users,so as to improve the performance.Finally,we take advantage of the network utility function as the objective,and design a heuristic algorithm to achieve the load balancing.Meanwhile,it also increases the revenue for the provider.(4)With regard to decreasing the cost of adjusting,we propose the updating strategy based on the lyapunov`s stability theory,so as to smooth the fluctuation of the bitrate.In the updating model,we considers the fluctuation of bitrate,change of routing and choice of servers to minimize the total cost as guarantee the bitrate of each users.Through the solution of LP(linear programming)model,we adjust the resource of application layer and underlying layer with decreasing the amount of variation,so as to guarantee the quality of service.