本文選題：云計算　切入點：P2P網(wǎng)絡(luò)　出處：《上海理工大學(xué)》2014年碩士論文　論文類型：學(xué)位論文

【摘要】：近年來,云計算技術(shù)的興起革新了傳統(tǒng)的計算模式,以互聯(lián)網(wǎng)為依托開辟出了計算資源利用的新模式。云計算的核心是在虛擬化技術(shù)的基礎(chǔ)上,通過互聯(lián)網(wǎng)技術(shù)為用戶提供動態(tài)易擴展的計算資源。從本質(zhì)上來講,云計算就是一種通過互聯(lián)網(wǎng)以服務(wù)的方式提供動態(tài)虛擬化資源的計算模式。如何高效地組織和控制網(wǎng)絡(luò)上大量的虛擬計算資源,并以計算資源池的方式向用戶按需提供計算資源以及協(xié)調(diào)用戶需求是云計算應(yīng)用研究的重點。利用中心服務(wù)器的計算模式來管控網(wǎng)絡(luò)上大量云資源使得中心服務(wù)器成為整個系統(tǒng)的瓶頸,不利于云計算的大規(guī)模應(yīng)用,因此提出使用對等網(wǎng)技術(shù)構(gòu)建分布式的云資源索引存儲和查詢系統(tǒng)。具體思路是用SHA-1散列算法為云節(jié)點生成一個唯一的標(biāo)識符,并且該標(biāo)識符被網(wǎng)絡(luò)拓撲結(jié)構(gòu)、索引存儲和節(jié)點查詢共同使用,從而生成一個基于結(jié)構(gòu)化P2P網(wǎng)絡(luò)的分布式云計算平臺。對等網(wǎng)絡(luò)是一個自組織的網(wǎng)絡(luò),被廣泛地應(yīng)用于資源共享和資源搜索領(lǐng)域。相比其它的網(wǎng)絡(luò)模型,P2P網(wǎng)絡(luò)具有分散化、可擴展性、健壯性以及高性能等顯著特點。與非結(jié)構(gòu)化P2P網(wǎng)絡(luò)相比,基于DHT的結(jié)構(gòu)化P2P網(wǎng)絡(luò),在穩(wěn)定性和擴展性方面都更有優(yōu)勢。因此,使用對等網(wǎng)絡(luò)技術(shù)來解決云計算中資源共享和資源查詢的問題。在對等網(wǎng)絡(luò)技術(shù)的基礎(chǔ)上解決云計算應(yīng)用中面臨的云資源查詢問題是課題研究的重點。雖然基于DHT的P2P網(wǎng)絡(luò)在單維度查詢的效率很高,但其不支持云資源多維查詢。為了解決云資源多維查詢這個問題,在結(jié)構(gòu)化P2P網(wǎng)絡(luò)的基礎(chǔ)上,根據(jù)資源信息相關(guān)度和網(wǎng)絡(luò)分組機制建立適合多維查詢的覆蓋網(wǎng)絡(luò);然后按照不同方式對資源屬性和屬性值進行編碼,結(jié)合多維度云資源存儲和發(fā)布策略實現(xiàn)了一種適用于云資源多維查詢的方法。在設(shè)計的結(jié)構(gòu)化云對等網(wǎng)絡(luò)模擬器上進行實驗,實驗結(jié)果表明云資源查詢維度個數(shù)的變化對網(wǎng)絡(luò)的平均查詢時延很小,查詢時延并沒有隨著查詢維度數(shù)的增加而增加,而是隨著查詢維度的增加而趨于穩(wěn)定。此外,與傳統(tǒng)的對等網(wǎng)絡(luò)Chord相比,本文所提出的多維度查詢算法是穩(wěn)定高效的,不會隨著網(wǎng)絡(luò)節(jié)點數(shù)的增加而產(chǎn)生較大的查詢時延,適合云計算這種網(wǎng)絡(luò)節(jié)點數(shù)較大的計算模型,拓展了P2P網(wǎng)絡(luò)技術(shù)在云計算服務(wù)中的應(yīng)用。
[Abstract]:In recent years, the rise of cloud computing technology has revolutionized the traditional computing model and opened up a new model of computing resource utilization based on the Internet. The core of cloud computing is based on virtualization technology. Provide users with dynamic and scalable computing resources through Internet technology. Essentially, Cloud computing is a computing model that provides dynamic virtualization resources through the Internet as a service. How to efficiently organize and control a large number of virtual computing resources on the network, It is the focus of cloud computing application research to provide computing resources to users on demand by computing resource pool and to coordinate user needs. The central server computing model is used to control a large number of cloud resources on the network to make the central services. The device becomes the bottleneck of the whole system, Therefore, a distributed cloud resource index storage and query system based on peer-to-peer network technology is proposed. The idea is to generate a unique identifier for cloud nodes by using SHA-1 hash algorithm. Moreover, the identifier is used by network topology, index storage and node query to generate a distributed cloud computing platform based on structured P2P network. Peer-to-peer network is a self-organized network. It is widely used in the field of resource sharing and resource searching. Compared with other network models, P2P networks are characterized by decentralization, scalability, robustness and high performance. Structured P2P networks based on DHT have more advantages in stability and scalability. P2P network technology is used to solve the problem of resource sharing and resource query in cloud computing. Solving the problem of cloud resource query in cloud computing application on the basis of peer-to-peer network technology is the focus of the research. DHT P2P network is very efficient in single dimension query. In order to solve the problem of multi-dimensional query of cloud resource, based on structured P2P network, the overlay network suitable for multi-dimensional query is established according to the correlation of resource information and network grouping mechanism. Then the resource attributes and attribute values are coded in different ways, and a multi-dimensional cloud resource storage and publishing strategy is used to implement a method for multi-dimensional query of cloud resources. The experiments are carried out on a structured cloud peer-to-peer network simulator. The experimental results show that the change of the number of query dimensions of cloud resources has little average query delay, and the query delay does not increase with the increase of query dimension, but tends to be stable with the increase of query dimension. Compared with the traditional peer-to-peer network Chord, the multi-dimensional query algorithm proposed in this paper is stable and efficient, and will not cause a large query delay with the increase of the number of network nodes, so it is suitable for cloud computing, which has a large number of network nodes. The application of P2P network technology in cloud computing services is expanded.
【學(xué)位授予單位】：上海理工大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TP393.02
，

本文編號：1593626