【摘要】：以2010年為界限，我們可以把過去分為個人電腦時代和互聯(lián)網(wǎng)時代，，而未來就是云計算、物聯(lián)網(wǎng)和三網(wǎng)融合的時代。云計算這種新的計算模式的誕生，使ICT界發(fā)生了革命性的變化。云計算一出現(xiàn)就成為了業(yè)界關(guān)注的熱點，它的設(shè)計理念就是用戶獲得各種服務(wù)就像使用水電一樣，按需使用，付費使用。這種計算模式讓每一個人都能以極低的成本接觸到頂尖的ICT技術(shù)。ICT界的商業(yè)巨頭微軟、谷歌、IBM、亞馬遜、Salesforce等都紛紛推出了自己的云計算產(chǎn)品，并把云計算作為其未來發(fā)展的主要戰(zhàn)略目標之一。 如何對資源進行有效的分配成為云計算所要解決的重要問題，也是云計算研究的熱點問題。為了更好的合理利用資源和提高資源的利用率，本文提出了一種基于改進粒子群優(yōu)化算法的資源分配和調(diào)價策略。這種改進的粒子群算法叫做粒子數(shù)量動態(tài)可變的粒子群優(yōu)化算法。因為在云計算環(huán)境下資源是海量的，數(shù)以萬計的用戶需求，如何在短時間內(nèi)合理的分配資源是我們所期望的。通過與傳統(tǒng)的粒子群算法進行比較，粒子數(shù)量動態(tài)可變的粒子群算法在時間復雜度上要遠優(yōu)于傳統(tǒng)的粒子群算法，這樣就大大節(jié)省了云計算環(huán)境下對資源分配的時間。根據(jù)負載的特性，構(gòu)造了體現(xiàn)其對所獲資源的滿意度和符合自身經(jīng)濟利潤的效用函數(shù)，利用資源代理不斷調(diào)價的方法來達到合理利用資源的目的，同時也最大化了每個負載的效益。通過仿真實驗并證明了改進粒子群優(yōu)化算法在云計算環(huán)境下資源分配的有效性、可行性、魯棒性以及大大縮短了計算時間。 云計算數(shù)據(jù)中心是由大量的異構(gòu)的計算資源池、存儲資源池以及網(wǎng)絡(luò)資源池等構(gòu)成。當所有的服務(wù)器都啟動時，會消耗大量的電能，產(chǎn)生大量的廢氣污染物給我們的環(huán)境造成極大的污染。因此，我們需要綠色節(jié)能的解決方法，不但能夠最小化執(zhí)行的代價還減小了對環(huán)境的影響。本文根據(jù)以上問題提出了資源分配的高效節(jié)能的算法。這種算法的基本思想是通過支持VMs的運轉(zhuǎn)在不同的物理節(jié)點之間，然后根據(jù)執(zhí)行的需求，動態(tài)的遷移VMs。當VMs不利用主機提供給的資源的時候，它就會動態(tài)的改變或整合到最小數(shù)目的物理節(jié)點上，而閑置的物理節(jié)點就會被關(guān)掉，消除了閑置機器能量的消耗，最終達到減小總能量的消耗。最后通過和其他策略的資源分配算法進行比較，得出本文提出的算法是有效的，在節(jié)能方面是最優(yōu)的。 不管是在哪一種計算模式下，資源分配都是需要亟待解決的問題，在云計算環(huán)境下也不例外。本文做了兩方面關(guān)于資源分配的研究，第一方面是基于市場機制的資源分配策略，從服務(wù)提供商和服務(wù)Qos以及均衡負載多方面考慮，提出了一套行之有效的資源分配方法。第二方面是從綠色節(jié)能的角度出發(fā)，從現(xiàn)實生活中得到啟發(fā)，提出了如何整合資源節(jié)約能量，最終達到環(huán)保的目的。
[Abstract]:By 2010, we can divide the past into the personal computer era and the Internet era, and the future is cloud computing, the Internet of things and the three networks of the age. Cloud computing, a new computing model, has revolutionized the ICT world. Cloud computing has become the focus of attention of the industry as soon as it appears. Its design philosophy is that users can obtain various services just as they use hydropower, use them on demand, and pay for their use. This computing model gives everyone access to the top ICT technology, ICT business giants Microsoft, Google, IBM, Amazon Salesforce, and so on, at very low cost, all of which have launched their own cloud computing products. And cloud computing as its future development of one of the main strategic objectives. How to allocate resources effectively becomes an important problem in cloud computing, and it is also a hot issue in cloud computing research. In order to make better use of resources and improve the utilization of resources, this paper proposes a strategy of resource allocation and price adjustment based on improved particle swarm optimization algorithm. This improved particle swarm optimization algorithm is called particle swarm optimization with variable particle number. Because in the cloud computing environment the resources are massive, tens of thousands of user needs, how to allocate resources reasonably in a short time is what we expect. Compared with traditional particle swarm optimization (PSO), the time complexity of PSO with variable particle number is much better than that of PSO, which greatly saves the time of resource allocation in cloud computing environment. According to the characteristics of the load, the utility function, which reflects its satisfaction with the acquired resources and conforms to its own economic profit, is constructed, and the method of continuously adjusting the price of the resource agent is used to achieve the purpose of making rational use of the resources. It also maximizes the benefits of each load. The simulation results show that the improved particle swarm optimization algorithm is effective, feasible, robust and greatly shortens the computing time in cloud computing environment. Cloud computing data center is composed of a large number of heterogeneous computing resource pool, storage resource pool and network resource pool. When all servers start up, it will consume a lot of electric energy and produce a large amount of exhaust gas pollutants to cause great pollution to our environment. Therefore, we need green energy saving solutions, which not only minimize the cost of implementation, but also reduce the impact on the environment. In this paper, an efficient energy-saving algorithm for resource allocation is proposed. The basic idea of this algorithm is to dynamically migrate VMs according to the requirements of execution by supporting the operation of VMs between different physical nodes. When VMs does not take advantage of the resources provided by the host, it dynamically changes or integrates into a minimum number of physical nodes, and the idle physical nodes are shut down, eliminating the energy consumption of idle machines. Finally, the total energy consumption is reduced. Finally, by comparing with the resource allocation algorithm of other strategies, it is concluded that the algorithm proposed in this paper is effective and optimal in the aspect of energy saving. No matter in any computing mode, resource allocation is a problem that needs to be solved urgently, and it is no exception in cloud computing environment. In this paper, two aspects of resource allocation are studied. The first is the resource allocation strategy based on market mechanism. Considering service provider and service Qos and load balancing, a set of effective resource allocation method is proposed. The second part is from the point of view of green energy saving, from the real life inspiration, put forward how to integrate the resources to save energy, and finally achieve the goal of environmental protection.
【學位授予單位】：山東師范大學
【學位級別】：碩士
【學位授予年份】：2012
【分類號】：TP3

【引證文獻】

相關(guān)碩士學位論文 前1條

1 陳盈盈;具有云服務(wù)特征的遠程在線考試系統(tǒng)的設(shè)計與實現(xiàn)[D];北京郵電大學;2013年

本文編號：2192712