A path adaptation scheme for minimizing quality disruption in peer-to-peer live video streaming over mobile heterogeneous networks

Abstract

With increasing needs of a live video streaming service to mobile devices, peer-to-peer video delivery is required to provide seamless quality of video even if a user moves across heterogeneous networks. The key challenge to provide a seamless peer-to-peer live video delivery over mobile heterogeneous networks is to handle dynamic changes of networks due to vertical handoff. In order to adapt to network dynamics, lots of path adaptation schemes have been proposed to construct a new overlay path for routing efficiency and to handle a bandwidth heterogeneity problem of different access networks. However, there has been little attention on user``s video quality during path adaptation. A mobile user may suffer from video quality degradation during path adaptation due to (1) insufficient path capacity of a new overlay path; (2) congestive losses when roaming with the previous rate preserved; and (3) adaptation delay imposed by a series of procedures for path adaptation. In order to minimize the service disruption due to adaptation delay, a multi-path handoff support which allows receiving packets via both of old network and new one in their overlapped coverage is required. However, it causes shared bottleneck problem which hinders proper path adaptation. In this dissertation, we propose a path adaptation scheme for peer-to-peer live video delivery to minimize the video quality degradations during vertical handoff. The proposed scheme consists of two specific adaptation algorithms: (1) QoS-aware path selection with active probing; and (2) conservative rate adaptation with redundancy control. We exploit active probing flows to obtain path capacity information by estimating available bandwidths of potential overlay paths and detecting shared bottleneck links. Using this path capacity information about the potential paths, a mobile user can choose a new overlay path which can provide the best QoS among the potential paths. After constructing a new overlay path, the ...