Service Provisionings Over Information-Centric Networking, Implementations And Optimizations

The proliferated Interert has become one of the most important infrastructure in our society. It is reported that the amount of internet users in China reaches 731 million by the end of Febuary 2016, which had increased sevenfold from 2005. The fast growing Internet leads to the emerging of more customerzied services and introduces more heavy network traffics, and thus causes unprecedured pressure on current TCP/IP networks, which was designed for end-to-end data transmission based on IP addresses. To tackle these challenges, Information-centric Networking(ICN) has been proposed rencently as a promising future Internet architecture, whose communication paradigm is based on data rather than IP address. ICN supports in-network caching which aims to reduce network latencies and mitigate the redundant traffic like video streaming services. The named data of ICN facliliates the content distribution over wireless networks, Internet of Things(IoT), and vehicular networks, etc.Current research focuses on data naming, in-network caching and other network layer techoniques such as routing and forwarding.However, few work has been done to bridge the gap between ICN and upplayer servers. We argue that the service provision over ICN is the key to the practical deployment of the new network architecture. Specifically,This thesis studies the web and video services and the contributions lie in three aspects as follows.1) To bridge the gap between ICN and service, we design and implement a web service platform over ICN and prototype an ICN web server termed CCNxTomcat, which can run over TCP/IP networks and ICN.The platform enables the user access the web service over ICN through a browser and CCNxTomcat facilitates the service deployment over ICN. We deploy the web platform and CCNxTomcat in our campus future internet testbed and verify its feasibilities. We also validate the performance of CCNxTomcat and it shows that CCNxTomcat outperforms the state-of-the-arts in terms of the concurrent requests, throughput as well as the latencies.2) To support the dynamic request in ICN webserver and further improve the performance, we design and implement ccncache, a high performance dynamic ICN web caching which supports named-data query, insertion and deletion, ccncache harnesses data structures including counter bloom filter, hash table, trie tree and skip list to well-manage the named-based dynamic content ICN packets. The experimental results show that ccncache achieves 2.5x look up time reduction against the-state-of-the-arts, with only 30% additional memery cost.3) It was reported that video streaming accounts for 73% of the totoal internet traffic and the user quality of experience (QoE) is critical to the business of the stakeholders. Hence, our final goal is to improve the QoE video streamings over ICN, and more specifically, the QoE of dynamic adaptive streaming(DAS) over Content-centric Wireless Networking(CCWN). To this end, we design and implement BC-DAS,a system frame work which leverages CCWN in-network caching and application layer bitrate adaptation to achieve the optimized QoE results.Markov Decision Process is employed to achive the optimal adapation.We conduct extensive results over a real topology using real datasets to valiate superior of BC-DAS. It shows that BC-DAS outperforms the existing approaches both in overall QoE and the metrics such as startup delay, average video bitrate and total bufferint time, etc.