Efficient time-extended TV viewing through hybrid data redundancy in networked appliances

Abstract

Online TV streaming services hosted by video distribution platforms have become commonplace, overcoming the imitations of conventional stand-alone appliances in the duration and range of broadcast TV programs available for time-shifted viewing. However, these services face significant challenges due to their extensive resource requirements. To mitigate the burden on service providers' constantly investing in increasing resources while adapting to the evolving TV consumption trends where users want more control over what, when, and on which device they view, we propose a time-extended TV viewing system to provide flexible and convenient access to broadcast TV programs. The main contributions of our study are threefold. First, the cooperative approach of our system enables time-shifted TV viewing services over a much longer period than the combined duration provided by all stand-alone appliances, and at a much lower cost compared to those that rely only on streaming servers. Second, our proposed hybrid data redundancy scheme improves storage efficiency using erasure-coded fragments for all segments and enhances streaming performance using replicas for popular segments requiring more service capacity. Third, our proposed replacement policy reduces server costs by minimizing server load through maximizing the hit ratio of data stored in participating appliances. Finally, we demonstrate through extensive experiments that our proposed time-extended TV viewing system performs better compared to when replication or erasure coding is employed.