完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.author | 周奕佑 | en_US |
dc.contributor.author | Chou, Yi-Yu | en_US |
dc.contributor.author | 陳耀宗 | en_US |
dc.contributor.author | Chen, Yaw-Chung | en_US |
dc.date.accessioned | 2015-11-26T00:56:59Z | - |
dc.date.available | 2015-11-26T00:56:59Z | - |
dc.date.issued | 2015 | en_US |
dc.identifier.uri | http://140.113.39.130/cdrfb3/record/nctu/#GT070256523 | en_US |
dc.identifier.uri | http://hdl.handle.net/11536/126818 | - |
dc.description.abstract | 隨著Skype、Line等網路電話應用程式的普及, VoIP 這項科技越來越被重視。人們使用這些應用程式在網路上交談,若有過長的延遲,將會導致使用者對於此服務的體驗變差,更甚可能導致無法正常對話。現今有很多改善VoIP QoS的機制,但大部分都是區分各種服務的優先權的作法,現今網路隨時都存在著大量的VoIP資料流,一些VoIP資料流若在同一路由器相遇,且要往同一的地方去的話,此時應該要把VoIP資料流區分得更細,因為有些VoIP資料流已經花了很多時間到達這個路由器,又或者有些VoIP流接下來可能要花很久的時間才能到達目的地,而有些則沒有上面這些困擾,而不是只是簡單的區分各個服務的優先權。 因此,我們提出了一個方法去更加細分VoIP資料流的優先權。正常來說非固定的延遲通常都是發生在封包要經過的交換器與路由器上,而其中又以排隊延遲造成的非固定延遲最大,影響排隊延遲的最大因素為輸出埠頻寬的使用率,所以我們想用VoIP流的跳躍數與它所經過的交換器之輸出埠頻寬使用率來區分優先權,為了輕鬆拿到VoIP資料流的跳躍數與它所經過的交換器之頻寬使用率這兩樣資訊,所以我們使用軟體定義網路,而非一般的傳統網路,因為軟體定義網路的控制器能輕鬆的拿到它所管理之網路的各項資訊。 | zh_TW |
dc.description.abstract | Voice over IP becomes (VoIP) one of the most popular services in networking nowadays. Due to the wide usage of VoIP services such as Skype and Line, people are able to make conversation with each other over the Internet with low cost. However, VoIP is a critical real-time application because large end-to-end delay may be encountered and it leads to poor communication experience of VoIP users. There are several mechanisms to improve VoIP QoS and most of them are based on the traffic classes. It means that all VoIP flows will be assigned to the same class and hence the same queue. But there may be a huge number of VoIP flows over the Internet at any time. Some VoIP flows may come across in a switch or a router simultaneously. Some of these flows take a long time from source to this switch or will take a long time from this switch to the destination, but some of them are not. In other words, some of VoIP flows may be able to wait longer than others in an intermediate switch. In this thesis, we propose a novel mechanism that assigns different VoIP flows into different queues based on the path selection by SDN. Typically the biggest variable delay in IP voice communication systems is the delay introduced by switches or routers along the path, and most of time the biggest one is the queuing delay which is influenced by the link utilization. Thus, we use hop-count and link utilization to determine which queue the incoming VoIP flows should go through. To ease the task of calculating hop-count of flows and link utilization in the switches, we use software-defined network (SDN) instead of traditional network to perform the task of routing path selection. As the simulation results showed, our proposed mechanism can improve QoS of VoIP significantly. | en_US |
dc.language.iso | en_US | en_US |
dc.subject | 網路電話 | zh_TW |
dc.subject | 服務品質 | zh_TW |
dc.subject | 流量分類 | zh_TW |
dc.subject | 軟體定義網路 | zh_TW |
dc.subject | VoIP | en_US |
dc.subject | Quality of service | en_US |
dc.subject | Traffic classification | en_US |
dc.subject | software-defined network | en_US |
dc.title | 在軟體定義網路下基於跳躍數實現有限制的網路電話端對端延遲 | zh_TW |
dc.title | A hop-count based mechanism for achieving bounded end-to-end delay of VoIP over SDN | en_US |
dc.type | Thesis | en_US |
dc.contributor.department | 網路工程研究所 | zh_TW |
顯示於類別: | 畢業論文 |