| dc.contributor | 資訊科學系 | |
| dc.creator (作者) | 蔡子傑;張宏慶 | zh_TW |
| dc.date (日期) | 2013 | |
| dc.date.accessioned | 20-四月-2016 17:01:40 (UTC+8) | - |
| dc.date.available | 20-四月-2016 17:01:40 (UTC+8) | - |
| dc.date.issued (上傳時間) | 20-四月-2016 17:01:40 (UTC+8) | - |
| dc.identifier.uri (URI) | http://nccur.lib.nccu.edu.tw/handle/140.119/85803 | - |
| dc.description.abstract (摘要) | 雲端運算是當下極受矚目的技術。做為在網路上提供服務的一個新選擇，雲 端平台擁有很高的彈性與以使用量付費的機制，讓使用者不再需要提前做審慎的 硬體採購規畫，可以隨時在需求增加時，再添購更多服務。 然而，在雲端環境中提供服務時，若以固定數量的虛擬主機配置，在面對少 數使用者時，將造成許多虛擬主機的閒置，而在面對數量遠大於預期的使用者 時，將面臨無法有效提供服務的問題。在現有的雲端平台中，會以 Auto Scale動 態的方式增加虛擬主機，提供水平擴充，以處理意外的負載，以及動態減少虛擬 主機，以節省成本。但是目前的做法，大都必須仰賴使用者自行提供的 threshold 設定，而且沒有預測未來的 workload 變化，所以動態擴充的效果可能會不如預 期。[11] 因此本研究的議題如下：針對雲端環境，提出一個動態式的資源服務。由於 增加或減少虛擬主機有一定的時間成本，如果在某個時間內將要處理大量的使用 者需求，透過 Queuing Model 與 Support Vector Regression 預測系統未來 workload 之效能影響與需求 request rate 的變化，設計一套動態式的資源服務， 根據所預測的結果，系望能在系統負載過高之前，就能進行水平擴充。而如果在 某個時間內使用者數量不多時，系統能自動減少虛擬主機，以有效使用雲端服務 之資源。 | |
| dc.description.abstract (摘要) | Cloud computing is one of the most important technologies today. As a new choice for hosting and delivering services over the Internet, users do not have to carefully plan ahead for hardware due to cloud computing features with high elasticity and "pay as you go" service. More computing service is purchased when demand increases. However, when hosting a service in the cloud environment, using a fixed number virtual machine configuration may fail to deliver good quality of service when the number of incoming requests rapidly rises; may leave many virtual machines idle when the number of users is very small. Cloud computing services in the market provides "auto scale" services, allowing dynamic creation of virtual machines to handle unexpected workload, and shut down unnecessary virtual machines to lower the cost. However, most of the existing services rely on threshold setting by users themselves. Thus, without the capability of predicted future workload change, the effect of dynamic auto scale services is not very prominent. The research issue of this research is the dynamic auto scale service provisioning in the cloud environment. There is some boot time required to start up a virtual machine. In order to effectively handle a large number of incoming requests, we hope that the system can be auto scaled up before the system is overwhelmed. We propose to use Queuing Model and Support Vector Machine to predict system response time and system workload. Based on the estimates, we will develop a novel auto scale strategy. Therefore, the resource will be effectively and economically utilized. | |
| dc.format.extent | 974661 bytes | - |
| dc.format.mimetype | application/pdf | - |
| dc.relation (關聯) | 計畫編號 NSC 102-2221-E004-003 | |
| dc.subject (關鍵詞) | 雲端運算;Auto Scale;Queuing Theory;Support Vector Machine | |
| dc.subject (關鍵詞) | Cloud computing;Auto Scale;Queueing Theory;Support Vector Machine | |
| dc.title (題名) | 適用於雲端環境下的動態資源服務 | zh_TW |
| dc.title.alternative (其他題名) | Dynamic Resource Provisioning in the Cloud Environment | |
| dc.type (資料類型) | report | |
