Full metadata record
DC FieldValueLanguage
dc.contributor.author郭宇軒en_US
dc.contributor.authorYu-Hsuang Guoen_US
dc.contributor.author邵家健en_US
dc.contributor.authorJohn Kar-Kin Zaoen_US
dc.date.accessioned2014-12-12T02:54:57Z-
dc.date.available2014-12-12T02:54:57Z-
dc.date.issued2005en_US
dc.identifier.urihttp://140.113.39.130/cdrfb3/record/nctu/#GT009317508en_US
dc.identifier.urihttp://hdl.handle.net/11536/78720-
dc.description.abstract我們的目的是設計出一個有彈性的應用層多點傳播架構,來提供即時多媒體串流服務。我們主要是針對應用層多點傳播的兩個問題。第一、組成應用層多點傳播的用戶隨時都會動態的加入或離開,因此資料傳輸並不可靠。第二、使用非對稱網路連線的用戶,它們上傳的頻寬遠遠少於下載頻寬,上傳的頻寬不足將會是應用層多點傳播的瓶頸。 我們提出三種方法的結合來產生出一個有彈性的應用層多點傳播機制,並且解決上傳頻寬不足的問題。這三種方法是:一、information dispersal algorithm,二、multiple stripes/trees,三、helper。我們增強彈性的方法是藉由保證訂閱戶就算遺失一些封包仍然可以得到完整資料,以及任一條網路連線中斷將不會有訂閱戶收不到任何封包。我們解決上傳頻寬不足的方法是充分利用所有用戶的上傳頻寬,以及藉由helper的加入來增加上傳頻寬的總量。 我們從模擬實驗中觀察到幾項結果:一、每個訂閱戶的訊息延遲時間是穩定的,而且訂閱戶之間的訊息延遲時間差距很少,二、就算對上傳頻寬不足的非對稱網路連線而言,每條網路連線的平均頻寬消耗是少的,三、結果顯示就算用戶有機率會發生錯誤時,訂閱戶仍然有良好的訊息成功還原率。zh_TW
dc.description.abstractOur purpose is to devise a new application layer multicast scheme to provide real time multimedia streaming service. There are two challenges for application layer multicast that we focus on. First challenge is peers that form the application layer multicast service may dynamically join or leave at any time. Data transmission is not reliable. Second one is peers with asymmetric connectivity that upstream bandwidth is much less than downstream bandwidth. Insufficient upstream bandwidth will be the bottleneck of application layer multicast. We propose the combination of three approaches to provide a resilient application layer multicast mechanism and solve the issue of peers with insufficient upstream bandwidth. The three approaches are: (1) information dispersal algorithm, (2) multiple stripes/trees, (3) helper. We improve resilience by promising subscribers can tolerate some packets loss without losing data completeness and when any link break, none of subscribers can’t receive any packets. We solve the insufficient upstream bandwidth issue by fully utilizing the upstream bandwidth of all peers and increasing the total amount of upstream bandwidth by the participation of helpers. We observe several results from the simulation: (1) the delay of message restoration for each subscriber is stable and the difference of delay between subscribers is small, (2) the average bandwidth consumption of one link is low, even for insufficient upstream bandwidth links, (3) it shows subscribers have good successful probability of message restoration even if peers have failure probability.en_US
dc.language.isoen_USen_US
dc.subject應用層多點傳播zh_TW
dc.subject多點傳播zh_TW
dc.subjectApplication Layer Multicasten_US
dc.subjectMulticasten_US
dc.title適用於非對稱網路連線之動態用戶的彈性應用層多點傳播zh_TW
dc.titleResilient Application Layer Multicast Tailored for Dynamic Peers with Asymmetric Connectivityen_US
dc.typeThesisen_US
dc.contributor.department資訊科學與工程研究所zh_TW
Appears in Collections:Thesis


Files in This Item:

  1. 750801.pdf

If it is a zip file, please download the file and unzip it, then open index.html in a browser to view the full text content.