結合虛擬佇列和分配演算法於雲端運算中的任務管理

Combining Virtualized Queues and Dispatched Algorithms for Task Management in Cloud Computing

張運晟(Yun-Cheng Chang)

指導教授 ： 林傳筆

朝陽科技大學/資訊學院/資訊與通訊系碩士班/碩士(2011年)

摘要

由於網路服務使用量快速增長，服務供應商所能夠提供的運算資源越來越不足，而為了維持服務的品質及提供更完整的網路應用服務，高擴展及強大運算能力的雲端環境將是目前趨勢；然而，在雲端環境下達到有效率的資源分配，以解決任務分配不均而造成壅塞及遺失的問題。因此，本文結合 Combined Input-Crosspoint Buffer (CICB)、Threshold 及 Flow Control，提出一個混合式 Cross Queue Capacity Virtualization Management (CCQVM) 的雲端基礎架構，以降低佇列等待時間及任務遺失率。
雖然提出的 CCQVM 能夠提供較高的服務品質保證於雲端運算中，但在網路服務商還是有兩個必須要去考慮的問題：Unfair 和 Unbalanced； Unfair 的問題是提供服務時沒有考慮到任務的屬性，Unbalanced 的問題是不考慮雲端分群的工作量去分配任務造成滿載的狀況，所以在本論文中將於上述的 CCQVM 中提出動態的雲端服務區塊選擇機制(Dynamic Cloud Cluster Election Mechanism)，讓各個雲端分群定期回報目前負載狀態，利用此方法不僅能讓每個工作都能被分配到適合的雲端分群，並且讓各個雲端分群的資源使用率達到負載平衡。

關鍵字

負載平衡 ； 服務品質保證 ； 雲端運算

並列摘要

As the uses of network service are growing rapidly, the service providers of computing resource have become inadequate increasingly. To keep Quality of Service and more complete network applications, the scalable and powerful computing capability of cloud environment will be a research trend. However, the cloud environment causes the problems of task loss and network congestion to achieve an efficient allocation of resources. Therefore, we propose a cloud infrastructure of hybrid Cross Queue Capacity Management (CCQVM) integrating Combined Input-Crosspoint Buffer (CICB), Threshold, and Flow control mechanisms to reduce the queue wait time and task loss rate.
Although the proposed CCQVM with algorithms can provide high QoS in Cloud Computing, there are two critical problems in the Internet service provider’s (ISP): one is unfair and the other is unbalanced. The unfair problem is caused by dispatching service requirement of user without considering service character; the latter problem is overloading as a number of service requirements are dispatched to the specific servers without considering the workload of servers. As a result, an efficient Dynamic Cloud Cluster Election Mechanism (DCCEM) is proposed to achieve load balance of resource deployment in this thesis. Therefore, the performance and load balance of resource deployment can be enhanced in this studies even if a number of service requirements exit, a cloud computing environment especially.

並列關鍵字

Cloud Computing ； Quality of Service ； Load Balance

參考文獻

[1]N. Yigitbasi, A. Iosup, D. Epema, and S. Ostermann, "C-Meter: A Framework for Performance Analysis of Computing Clouds," IEEE/ACM International Symposium on Cluster Computing and the Grid 2009, pp.472-477, 18-21 May 2009.

[3]A. di Costanzo, M. D. de Assuncao, and R. Buyya, "Harnessing Cloud Technologies for a Virtualized Distributed Computing Infrastructure," IEEE Journal of Internet Computing, vol.13, no.5, pp.24-33, September 2009.

[5]R. Prodan and S. Ostermann, "A Survey and Taxonomy of Infrastructure as a Service and Web Hosting Cloud Providers," IEEE/ACM International Conference on Grid Computing 2009, pp.17-25, October 2009.

[6]R. Maggiani, "Cloud Computing is Changing How We Communicate," IEEE International Conference on Professional Communication 2009, pp.1- 4, July 2009.

[7]W. Tian, S. Su, and G. Lu, "A Framework for Implementing and Managing Platform as a Service in a Virtual Cloud Computing Lab," 2th International Workshop on Education Technology and Computer Science 2010, pp.273-276, March 2010.