An adaptive data replication technique with online advertising for P2P VoD services.

Ho, Chien-Peng
Chia-Ling Hsieh
Journal Title
Journal ISSN
Volume Title
International Association of Computer Science and Information Technology Press
This study aims to construct an agent-driven advertising framework for allowing agent-peers to distribute ad-videos to their potential customers and robust online marketing communication channels for smooth functioning of the Video-on-Demand (VoD) systems in unstructured Peer-to-Peer (P2P) overlay networks. In particular, the framework does not rely on a centralized collection server, but depends more on agent-peers that better meet customer needs and can transform a server load into scalable bandwidths with availability and high system reliability. However, in large-scale networks, it is important to reduce the load imposed on the server resources. This paper examines the problem of characterizing and assessing the local impact of the load imposed on the peer-level underlay networks by analyzing the kurtosis properties of packet inter-arrival time distribution on a P2P network. In addition, we analyzed the observed global impact of the load imposed for film-data replication to eliminate potential network bottlenecks using an extended density-based clustering (DENCLUE) algorithm of P2P overlay traffic on the agent-level underlay. The results indicate appropriate replication effective in system performance. It also points to the enormous potential of agent-level advertising in enhancing predetermined marketing objectives.