可信的自治式服务协同多租户租赁模型与云负载均衡算法的研究-计算机应用技术专业毕业论文.docxVIP

＼＼删渊浙江大学硕士学位论文 ＼＼删渊 浙江大学硕士学位论文 摘要 规范调控、策略驱动的自治式服务协同模型NGPD，可以解决因自治个体行 为难以预测、控制而导致的协同效应“可信’’危机【l】。然而，服务协同的社区应 用依托于众多的中小企业的参与，但协同模型NGPD部署、升级管理的复杂性难 以为中小企业所胜任，这一弊端势必成为NGPD发展和商业应用的瓶颈。本文结 合云计算中“软件即服务’’(Soft as a Service以下简称SaaS)的概念，采用托管 集成、集中部署、租户定制等方式，提出可信的自治式协同多租户租赁模型 NGPD．SaaS从而很好的解决该问题。同时，为应对大规模商业化的应用，本文针 对现有集群负载均衡算法得到弊端以及NGPD．SaaS模型的特点提出了云负载均 衡算法CLBA，将云能力动态平衡机制融入模型，通过本文模型大大增加NGPD 的实用价值和商业价值。 本文首先介绍了课题的相关背景，分析NGPD中可作为租赁的服务元素，并 给出了基于NGPD的服务分离、管理需要的策略，然后详细描述NGPD．SaaS模 型的业务流程和模型的详细软件设计架构，并在此基础上介绍云负载均衡算法 CBLA，文章的最后以小型会议安排实例分析了模型的执行效率和设计、部署方 法。 关键词：NGPD；服务对象分离；软件功能租赁；负载均衡；云计算；任务调度 浙江大学硕士学位论文 浙江大学硕士学位论文 Abstract Ab stract Norm-Govemed and Policy·Driven autonomic service cooperation model NGPD Can be resolved the‘trusted’crisis．However,the application of community service coordination relying on the participation of a large number of SMEs，but the collaborative model NGPD deployment，upgrade the management of complexity is difficult for small and medium enterprises are competent，this evil is bound to become the bottleneck of the NGPD development and commercial applications．In this paper， cloud computing”software as a service”(Soft as a Service SaaS)hereinafter referred to as the concept of a hosted integrated，centralized deployment，tenants custom to give a credible autonomous coilaborative multi．tenant rental the model NGPD．of SaaS makes it very good solution to the problem．The same time，in response to large—scale commercial application for the existing cluster load balancing algorithm ills and NGPD-SaaS model is characterized by the cloud load balancing algorithm CLBA， dynamic equilibrium mechanism of cloud capabilities into the model by the proposed model greatly increase the practical value and commercial value ofNGPD． This paper firstly introduces the topic of the relevant background，and then gives a definition for Server separate and management．Then it describes the server of NGPD-SaaS which base NGPD to separate and management，and the Architecture of NGPD-Saas，and base