Motif-based QoS-aware Dynamic Optimization of P2P Streaming Networks
Kazuki Ono
Graduate School of Science and Engineering, Saitama University, Japan.
Andrii Zhygmanovskyi
Graduate School of Science and Engineering, Saitama University, Japan.
Noriko Matsumoto
Graduate School of Science and Engineering, Saitama University, Japan.
Norihiko Yoshida *
Graduate School of Science and Engineering, Saitama University, Japan.
*Author to whom correspondence should be addressed.
Abstract
In recent years, Peer-to-Peer (P2P) streaming networks have been becoming popular because of not only fault tolerance, but also load balancing. However, in P2P streaming networks, there are some problems such as topology imbalance, interruption of streaming, latency in segment delivery, and degradation of received content caused by connecting low-performance nodes. Therefore, in this paper, we propose a new approach which reconfigures streaming networks dynamically using network motifs which can be used as an approach to characterize the complex networks, and evaluate our approach by simulation. The results indicate that our approach greatly decreases height and latency in generated trees, and provides an ability to reconfigure dynamically the network in case of predecessor's disconnection or defection. Moreover, our approach improves the quality of contents in a streaming network regardless of its network size.
Keywords: Peer-to-peer streaming networks, network motifs, resilience, dynamic reconfiguration.