Research On Video Data Distribution Algorithms For Opportunistic Networks

Opportunistic network is kind of Mobile Ad-Hoc Network characterized by delay and tolerance. In terms of structure of network, it is hard for network to maintain a stable and integrated link due to changes in structure of network. In the network, once nodes run within Communication diameter, nodes could transmit and exchange data during the limited staying time. Specifically, forwarding nodes will store the data until encounter another proper nodes.The process of exchanging data is that once nodes encounter with each other, nodes can transmit data with various conditions. At the same time, nodes can bring the relevant data until they meet the forwarding nodes or destination nodes to forward data. Given those networks, nodes that need to share data are not always in the same sub-networks. Thus, it is essential to study the movement model, routing algorithm, and data forwarding scheme as well as management of butter.With the spread of smart devices, users could take what kind of scenario they like. Therefore, there must be an effective algorithm to meet share between users. However, current algorithm focus generally on how to forward key information such as traffic information and weather in-vehicle network, less researchers concentrate on dissemination of video data with the needs of the user.Based on mentioned problems above, this thesis study how to receive a large amount of video block for users to get a better understanding of the whole video by looking through several video segments in a short time. First of all, the media, especially large multimedia contents in DTN (Delay Tolerant Networks), must be divided into pieces played alone by users in order to transport rapidly in DTN, and there must be not redundant frames included in these video blocks in order to prevent size of video blocks from being changed during the period of cutting.Besides, if users need to get a video consisted of sequential pieces in limited time, there must be an algorithm of conducting dissemination of pieces. Firstly, key features of forwarding pieces are theoretically investigated such as Length Priority, Intensity Priority, Position Priority and Area Priority. Secondly, the design and evaluation of RSCS (Random-Sequential Content Spreading) is presented, a completely distributed algorithm that selects pieces to transfer based on their priority of absence. Through this method, destination node can receive a sequential video queue during a short period of time, and users can look through the received pieces well. Besides, the performance of RSCS is evaluated by using synthetic traces from intermittently-connected networks in ONE simulator. When compared with sequential and randomized solutions, results show that RSCS significantly outperforms these approaches in terms of random-sequential index. Moreover, RSCS also achieves better performance levels in ratio of delivery and latency than sequential solution.