软件
軟件
연건
computer engineering & Software
2015年
9期
5-11
,共7页
流媒体%能耗%设备对设备%码率选择%移动云计算
流媒體%能耗%設備對設備%碼率選擇%移動雲計算
류매체%능모%설비대설비%마솔선택%이동운계산
Video streaming%Energy efficient%Device-to-Device(D2D)%Bitrate selection%Mobile cloud computing
针对传统的 HTTP 动态自适应流媒体(DASH)的邻近用户网络带宽竞争问题,提出一种设备对设备(D2D)协作式DASH系统,该系统采用了D2D协作的方式充分利用邻近用户的网络带宽能力下载和分享视频,客户端以中间件架构实现,使得该系统与主流播放器兼容。针对过度追求高视频质量而忽视智能手机能耗的问题,提出了一种能耗优化码率选择算法应用于 D2D 协作式 DASH 系统。根据智能手机当前的能耗、CPU 使用率、内存占用率与系统停留时间选出领队设备,领队设备负责视频的下载与D2D自组网的分享。实验测试结果表明:该算法与非合作算法相比在保证相同视频质量的情况下可以最多节省25%的能耗。
針對傳統的 HTTP 動態自適應流媒體(DASH)的鄰近用戶網絡帶寬競爭問題,提齣一種設備對設備(D2D)協作式DASH繫統,該繫統採用瞭D2D協作的方式充分利用鄰近用戶的網絡帶寬能力下載和分享視頻,客戶耑以中間件架構實現,使得該繫統與主流播放器兼容。針對過度追求高視頻質量而忽視智能手機能耗的問題,提齣瞭一種能耗優化碼率選擇算法應用于 D2D 協作式 DASH 繫統。根據智能手機噹前的能耗、CPU 使用率、內存佔用率與繫統停留時間選齣領隊設備,領隊設備負責視頻的下載與D2D自組網的分享。實驗測試結果錶明:該算法與非閤作算法相比在保證相同視頻質量的情況下可以最多節省25%的能耗。
침대전통적 HTTP 동태자괄응류매체(DASH)적린근용호망락대관경쟁문제,제출일충설비대설비(D2D)협작식DASH계통,해계통채용료D2D협작적방식충분이용린근용호적망락대관능력하재화분향시빈,객호단이중간건가구실현,사득해계통여주류파방기겸용。침대과도추구고시빈질량이홀시지능수궤능모적문제,제출료일충능모우화마솔선택산법응용우 D2D 협작식 DASH 계통。근거지능수궤당전적능모、CPU 사용솔、내존점용솔여계통정류시간선출령대설비,령대설비부책시빈적하재여D2D자조망적분향。실험측시결과표명:해산법여비합작산법상비재보증상동시빈질량적정황하가이최다절성25%적능모。
To solve the problem of network bandwidth competition among the group of adjacent users caused by the traditional Dynamic Adaptive Streaming over HTTP (DASH), a Device-to-Device (D2D) cooperative DASH was pro-posed to take full advantage of adjacent users’network capabilities to download and share video streaming. A layered middleware infrastructure was designed to make the proposed system compatible with major players. Concerning the problem of achieving the competitive video quality at the cost of significant energy consumption, an adaptive energy efficient bitrate selection method in D2D cooperative DASH was proposed. Captain devices were selected based on the state of devices such as energy state, CPU utilization, memory usage and the time the devices joint the group. Captain devices were in charge of downloading the video and sharing video in the D2D network. The experimental result shows that the proposed bitrate selection algorithm can reduce energy consumption of up to 25% as compared with the non-cooperative method while maintaining competitive video quality.