中国科学院软件研究所机构知识库
Advanced  
ISCAS OpenIR  > 中科院软件所  > 中科院软件所
题名:
基于移动Agent的网络管理系统的研究与实现
作者: 张普含
答辩日期: 2001
专业: 计算机软件与理论
授予单位: 中国科学院软件研究所
授予地点: 中国科学院软件研究所
学位: 博士
关键词: 移动agent ; 网络管理 ; 网络故障检测 ; 负载平衡
其他题名: Research on and Implementation of Network Management System based on Mobile Agents
摘要: 随着网络规模和服务的飞速发展,传统的集中式管理模式越来越暴露其在性能、可扩展性、灵活性等方面固有的局限性.为此该文对基于移动agent的网络管理系统进行了深入的研究和探讨,设计并实现了一个基于移动agent的网络故障检测系统;首次提出一个基于移动agent的网络管理系统性能分析的定量模型;给出一种新的基于SI和一种基于multi-swarm的网络负载平衡算法.该文首次提出了一个系统性能分析的定量模型,从理论上分析比较了两种模式的流量和响应时间以及影响系统性能的主要参数,并用模拟实验进行了验证.该文提出了一种新的基于SI的网络负载平衡算法-ABLB算法.该文就算法的性能在模拟实验中进行了实验分析,并与AntNet和MultipleRoundTripRouting两种算法进行了比较.进一步提出了另一种基于milti-swarm的改进算法-mABLB算法,此算法中多组具有SI特性的移动agent组同时工作,组与组之间相互协调.实验表明,此算法对ABLB算法又有进一步的优化.
英文摘要: Typically, most of existing network management applications are based on static, centralized and non-scalable client/server architectures, where some central entity processes large amounts of raw data gathered from each network element. By this reason, the management applications are not flexible, have problems of scalability and produce too much traffic in the network. Mobile agents have been proposed as a potential solution to the problems. But current research on network management system based on mobile agents is surely not enough, and many related problems need to be further researched. This thesis concentrates on three of them: we propose a network fault detection system based on mobile agents, and we present a quantitative model to analyze the performance(traffic and response time) of mobile agent design paradigm against the traditional client-server architecture for network management, in addition, a new algorithm based on swarm intelligence and other one based on multi-swarm for load balancing in communication networks are presented. As networks become larger and more complex, the need for advanced fault management capabilities becomes critical. Faults are unavoidable in large and complex communication networks, but quick detection and identification can significantly improve network reliability. Packet monitoring has become a standard technique in network fault detection, but when applied to a large-scale network it yields a high volume of packets. To overcome this problem, some techniques are proposed. However, the proposed techniques are based on popular SNMP agent, RJV1ON technology, which are also characterized by centralization, they do not provide scalability and strong local processing ability. In this thesis, we present an intelligent mobile agent-based architecture to deal with the problem of network fault detection using ICMP packet monitoring technique and implemented it by using java technology and' aglet. Experiment results show the proposed architecture is very effective in network fault detection. Mobile agents based network management system is a potential solution to the problems of traditional network management system, but it is difficult to identify precisely under which condition design exploiting mobile agent is preferable over a traditional one. In this thesis, we present a quantitative model to analyze the traffic and the response time of mobile agent design paradigm against the traditional client server architecture for network management. We compare mobile agent performance with SNMP one in several simulations. Analysis and simulations results show that the mobile agent performs better than the SNMP when the number of managed network elements exceeds a value related to the overhead of several retrievals of the SNMP. Results show that the mobile agent based system is more influenced by the task to be performed and is less sensitive to the bandwidth of the bottleneck link that connects the management station to the remote LAN. Moreover, a significant improvement of the mobile agent performance is obtained when the agent returns to the management station after visiting a fixed number of nodes. Network load balancing aims at evenly distributing the load over a network, thus leaving no routers idle and preventing overloads for others. It is becoming increasingly difficult due to the increasing size, rapidly changing topology, and complexity of communication networks. A new class of algorithms, inspired by swarm intelligence, is currently being developed that can potentially solve numerous problems of load balancing in modern communications networks. The paper presents a new algorithm based on swarm intelligence for load balancing in communication networks. This algorithm is an adaptive, distributed, mobile-agents-based algorithm which extends from Multiple Round Trip Routing algorithm. It includes two classes of mobile agents: strategy agents and load agents. As a further extended algorithm, we present another new algorithm based on multi-swarm for load balancing in communication networks. We have designed a network simulator to validate our algorithms and to compare them with AnteNet and Multiple Round Trip Routing. The good performance is proved by experiments.
语种: 中文
内容类型: 学位论文
URI标识: http://ir.iscas.ac.cn/handle/311060/7138
Appears in Collections:中科院软件所

Files in This Item:
File Name/ File Size Content Type Version Access License
LW008622.pdf(2600KB)----限制开放-- 联系获取全文

Recommended Citation:
张普含. 基于移动Agent的网络管理系统的研究与实现[D]. 中国科学院软件研究所. 中国科学院软件研究所. 2001-01-01.
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[张普含]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[张普含]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit
所有评论 (0)
暂无评论
 
评注功能仅针对注册用户开放,请您登录
您对该条目有什么异议,请填写以下表单,管理员会尽快联系您。
内 容:
Email:  *
单位:
验证码:   刷新
您在IR的使用过程中有什么好的想法或者建议可以反馈给我们。
标 题:
 *
内 容:
Email:  *
验证码:   刷新

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.

 

 

Valid XHTML 1.0!
Copyright © 2007-2017  中国科学院软件研究所 - Feedback
Powered by CSpace