完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.author | 柯承孝 | en_US |
dc.contributor.author | 易志偉 | en_US |
dc.contributor.author | 邵家健 | en_US |
dc.date.accessioned | 2014-12-12T01:51:56Z | - |
dc.date.available | 2014-12-12T01:51:56Z | - |
dc.date.issued | 2011 | en_US |
dc.identifier.uri | http://140.113.39.130/cdrfb3/record/nctu/#GT079855501 | en_US |
dc.identifier.uri | http://hdl.handle.net/11536/48236 | - |
dc.description.abstract | 在現今的移動通訊網路中,網路服務已經非常普遍,幾乎每位手機用戶都會使用到。但隨著用戶數量迅速的成長,硬體設備漸漸無法滿足大眾的需求,導致網路傳輸速率緩慢、應用程式服務品質下降,因此網路頻寬的不足已經成為一個全世界電信業者共通的課題。若是要增加基地台數目及線路數量,成本將會十分昂貴且緩不濟急,應該從另一個方面著手,試著降低多餘的資料傳輸量,並且尋找有效率的方式應用有限的頻寬資源。 在本篇論文中,將代理伺服器此一元件引進現今的網路架構,設置於手機及網際網路中。手機端的代理伺服器擁有快取的功能,可避免重複的資料傳輸,同時能控管與網路之間的所有連線,並進行流量排程的工作。網路端的代理伺服器會定期向各個應用程式的伺服器抓取新資料,儲存在它的快取中,同時在資料更新時,發送一短訊息告知手機端。所有從手機端送出的資料要求訊息都須經過網路端的代理伺服器,若是代理伺服器中的快取已經過期,才再向應用程式伺服器發送要求。 有了手機端及網路端的代理伺服器的支援後,本文中提出了基於此系統架構下的三個目標: (1)避免重複的資料傳輸造成頻寬浪費。(2)減少網路流量超出負荷及壅塞的情況。(3)針對每個使用者的需求以求將資料效益最大化。第一個目標可由代理伺服器提供的快取功能獲得實現。第二及第三個目標則可由手機端的代理伺服器對資料流量進行適當的排程來達到目的。因此,我們設計出一個以效能函數為基準的排程演算法。此效能函數將以服務品質和個人偏好兩方面做考量來計算其值。演算法會調整每個連線的傳輸順序及各自的流量,以即時的方式做排程,期使在一段時間內所獲得的效能函數值最大化。 實驗部分是以模擬的方式進行,我們基於現在的行動網路架構選擇一個合理的網路流量模型,並且加入了代理伺服器的支援。模擬所得到的數據結果,呈現出在有代理伺服器進行快取和流量排程的情形下,能有效的降低網路流量、減少壅塞的發生以及增加資料效益。未來則可考慮實際將代理伺服器部署於移動通訊網路中,佐以良好的演算法,則能有效減少增加硬體的成本。 | zh_TW |
dc.description.abstract | In recent years, a variety of services have been available to mobile subscribers. Because the number of subscribers grows too fast, the current infrastructure cannot give services to fulfill everyone’s need, which causes reduced data rates and poor service performances. To solve this problem, we propose Application Proxies into current network design. Mobile Proxy is set at the cell phone end, which has a local cache and manages all sessions so that traffic scheduling can be achieved. Cloud Proxy on the other hand is sited at the external data network namely Internet, which does periodicity crawling to third party application servers. When the content of an application has updated, an short message would be sent to the Mobile Proxy. With the support of Application Proxies, we may attain the following objectives: (1) eliminate redundant data transmission. (2) reduce network congestion and overloaded cases. (3) maximize data values for individual subscribers. First objective can be realized by content caching at Mobile Proxy, while second and third objectives will be achieved by sophisticated traffic scheduling mechanism. Thus an utility-based scheduling algorithm, which takes service quality and user preferences into consideration is proposed. The simulation environment is based on mobile network architecture and corresponding traffic models. The result shows that with suitable caching and traffic scheduling method, our goals can be truly achieved. | en_US |
dc.language.iso | zh_TW | en_US |
dc.subject | 代理伺服器 | zh_TW |
dc.subject | 效能函數 | zh_TW |
dc.subject | 行動網路 | zh_TW |
dc.subject | 線上排程演算法 | zh_TW |
dc.subject | Application Proxy | en_US |
dc.subject | Utility Function | en_US |
dc.subject | Mobile Network | en_US |
dc.subject | Scheduling Algorithm | en_US |
dc.title | 以行動應用程式為對象的效能導向流量排程演算法 | zh_TW |
dc.title | Utility-Based Traffic Scheduling for Mobile Applications | en_US |
dc.type | Thesis | en_US |
dc.contributor.department | 資訊科學與工程研究所 | zh_TW |
顯示於類別: | 畢業論文 |