低暫存空間與使用者頻寬之有效率熱門影片廣播法;Efficient Schemes for Broadcasting Popular Videos at Low Buffer Demand and Client Bandwidth

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/54389

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/54389

Title:	低暫存空間與使用者頻寬之有效率熱門影片廣播法;Efficient Schemes for Broadcasting Popular Videos at Low Buffer Demand and Client Bandwidth
Authors:	陳應南;Chen,Ying-Nan
Contributors:	資訊工程研究所
Keywords:	網際網路協定電視;GDB;HB;FDPB;緩衝;隨選視訊;定期廣播法;periodic broadcasting;video-on-demand (VoD);Internet Protocol Television (IPTV);fixed-delay pagoda broadcasting;greedy disk-conserving broadcasting;buffer;harmonic broadcasting
Date:	2012-07-12
Issue Date:	2012-09-11 18:49:48 (UTC+8)
Publisher:	國立中央大學
Abstract:	隨著視訊壓縮與寬頻網壹取技術的進步，網際網路協定電視(IPTV)儼然成為直接提供用戶端多媒體服務傳輸的熱門技術，定期廣播法(Periodic Broadcasting)則提供IPTV隨選視訊(VoD)服務一個有效的傳輸方法。這個方法將熱門影片切割成許多片段(segments)，同時在不同頻道上重覆播放。先前研究主要著眼於降低用戶端等待時間，例如：Fixed-Delay Pagoda Broadcasting (FDPB)與 Harmonic Broadcasting (HB)廣播法。然而，這些方法必須事先暫存大量影片資料，因此無法有效支援能力有限的用戶端設備，如：Set-Top Box (STB)。為了因應這樣的議題，Staircase Broadcasting (SB)、Reverse Fast Broadcasting (RFB)與Hybrid Broadcasting (HyB)等廣播法分別提出解決方案。此蟹首先提出一個結合FDPB與RFB的廣播法，這個方法能夠同時減少暫存空間與降低用戶端等待時間。另外，我們提出結合HB與SB的新廣播法，也達到同時減少緩衝空間與用戶端延遲時間的目的。與SB、RFB及HyB等方法比較，本文所提出的方法能夠在相同的暫存空間條件下，達到更低的等待時間。由於用戶端必須同時間接收與伺服端相同頻寬的影片資料，這個限制導致上述廣播法並不適用於使用行動網路的用戶，因為這將增加用戶端所需頻寬，而讓用戶產生額外費用。為解決此問題，Skyscraper Broadcasting (SkB)、Client Centric Approach (CCA) 與Greedy Disk-conserving Broadcasting (GDB)等廣播法提出支援用戶端小頻寬播放影片的方法，其中GDB更提供伺服器廣播頻道、用戶端頻寬及緩衝空間的較佳折衷方案。藉由改進GDB廣播法，此蟹提出reverse GDB (RGDB)的廣播法以達到用戶端低暫存空間與小頻寬需求的目標，並證明在低使用頻寬下用戶端仍然保證能順利持續播放影片，另外我們亦導出RGDB所需用戶端最大暫存空間量。最後，分析的結果顯示RGDB除支援用戶端小頻寬播放影片且得到與GDB相同的用戶端等待時間外，亦降低其用戶端緩衝空間需求達33%-50%。With the advances in video compression and broadband access technologies, the Internet Protocol Television (IPTV) becomes a popular technology for the delivery of multimedia services directly to the end users. Periodic broadcasting is an efficient approach to deliver IPTV services. The approach divides a popular video into segments, which are then simultaneously broadcast on different data channels. Previous studies mainly focus on decreasing client waiting time, such as the fixed-delay Pagoda broadcasting (FDPB) and the harmonic broadcasting (HB) schemes. However, these schemes must buffer a large part of the video data in advance such that cannot support limited-capability client devices, such as set-top boxes (STBs). To address this issue, some literatures, including the staircase broadcasting (SB), the reverse fast broadcasting (RFB), and the hybrid broadcasting (HyB) schemes, were proposed. This Dissertation also concentrates on reducing buffering space. Motivated by HyB, this work proposes a new scheme, which combines FDPB and RFB to yield small buffer requirements, as well as short waiting time. In addition, another new scheme integrating HB and SB is also proposed to save client buffering space and waiting time. In comparison with SB, RFB, and HyB, the proposed schemes can yield the smallest waiting time under the same buffer requirements. For above schemes, client receiving bandwidth equals server broadcasting bandwidth. This limitation causes these schemes to be infeasible in mobile networks because increasing receiving bandwidth at all client sites is expensive, as well as difficult. To alleviate this problem, the greedy disk-conserving broadcasting (GDB) scheme supports a client with a small bandwidth. In comparison with other similar scheme, such as the client-centric approach (CCA) scheme, the GDB clients have smaller waiting time. Extending GDB, this Dissertation designs a reverse GDB (RGDB) scheme to achieve small buffering spaces as well as low bandwidths at the client. We further mathematically prove that RGDB still guarantees on-time video delivery at small client bandwidths. A formula is derived for the maximum number of segments buffered by an RGDB client. Finally, an analysis shows that RGDB has 33%-50% smaller client buffer requirements than GDB in most situations.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	531	View/Open

社群 sharing

Loading...