Toward green media delivery: location-aware opportunities and approaches

Abstract

Mobile media has undoubtedly become the predominant source of traffic in wireless networks. The result is not only congestion and poor Quality-of-Experience, but also an unprecedented energy drain at both the network and user devices. In order to sustain this continued growth, novel disruptive paradigms of media delivery are urgently needed.We envision that two key contemporary advancements can be leveraged to develop greener media delivery platforms: 1) the proliferation of navigation hardware and software in mobile devices has created an era of location-awareness, where both the current and future user locations can be predicted; and 2) the rise of context-aware network architectures and self-organizing functionalities is enabling context signaling and in-network adaptation. With these developments in mind, this article investigates the opportunities of exploiting location-awareness to enable green end-to-end media delivery. In particular, we discuss and propose approaches for location-based adaptive video quality planning, innetwork caching, content prefetching, and long-term radio resource management. To provide insights on the energy savings, we then present a cross-layer framework that jointly optimizes resource allocation and multi-user video quality using location predictions. Finally, we highlight some of the future research directions for location-aware media delivery in the conclusion.

Energy-Saving Pushing Based on Personal Interest and Context Information

“…. , α T f during the T f frames [12]. Consider g t = α t Ts h t 2 /(N 0 W t ) and denoteg t α t Ts h t 2 /(N 0 W max ), then the equivalent channel gain can be rewritten as g t = Wmax W tg t .…”

Role of large scale channel information on predictive resource allocation

“…In practical systems, the large-scale channel gains can be estimated from the radio map [2] with the help of the predicted user location, which inevitably have estimation errors. Fortunately, the imperfect large-scale channel information has little impact on the estimated distribution information f (g) in (8), as demonstrated via simulations later.…”

“…As the smart phone popularizes, improving the performance of mobile networks such as energy efficiency (EE) by exploiting context information has drawn attentions [1][2][3][4][5].…”

“…For the traffic not having an urgent deadline for transmission, such as pre-subscribed file downloading, content pushing, and on-demand video streaming delivery, resource planning with user level context information, or predictive resource allocation, can save energy and improve performance of a network significantly [2-5, 8, 9]. With perfect achievable rate prediction, i.e., assuming perfect future channel information, half energy can be saved [2].…”

“…However, in wireless channels the small-scale channel gains in the future are hard to know accurately, where the channel prediction errors depend on the channel coherence time. By contrast, the large-scale channel gains can be estimated from user locations and the radio map [2], which however has estimation errors due to the prediction error of locations and the measurement error of the signal strength as reported by [10]. Recently, the simulation results provided in [5] demonstrate that the power-saving gain obtained from knowing the future large-scale channel gains is almost the same as that provided by perfectly knowing all the future instantaneous channel gains.…”