| Carl Friedrich Gauß Faculty | Department of Computer Science

Policy-based Management - An Approach With Adaptive Control Theories

Student(anonymous, Login required)
SupervisorXiaoyuan Gu
ProfessorProf. Dr.-Ing. Lars Wolf
IBR GroupCM (Prof. Wolf)
TypeDiploma Thesis

Motivation: Policy-based Management (PBM) has been the subject of extensive research for over a decade. The IETF has been investigating Policy-based Networking as a means of managing IP-based multi-service networks with quality of service guarantees. More recently, PBM has risen to the forefront of the telecom industry. With many vendors announcing plans and introducing products. This is driven by the fact that policy has been well recognized as a solution to manage complexity, and to guide the behaviour of a network or distributed system through high-level user-oriented abstractions. By definition, policy refers to a global goal statement or constraint. A policy statement does not specify the internal implementation of the policy. Moreover, a policy statement should closely match with the goals of the policy maker to reduce the chance of entry error.

Adaptive control theories are widely recognized as an effective way in handling system uncertainties, in particular in environment with variable operating conditions that make the traditional closed-loop control based on constant gain and linear feedback virtually invalid. An adaptive control system is able to couple the design of controller with on-line and recursive estimation of system variables. Or in other words, the controller is able to automatically adjust its parameters and behaviors in response to changes in process and disturbance dynamics. To a large extent, adaptive control theories offer valuable inputs on how proactive monitoring and control in Autonomic Communications can be achieved.

Tasks: In this project, the student will design a monitoring and control model within OPNET Modeler for policy-based management. This is done through to have a good understanding of policy-based management and adaptive control theories. And, based on this, to study the COPS-PR framework and the simulation workflow in OPNET Modeler. A focus of the work will be on the QoS parameter optimization. Analysis of critical control parameters to AutoComm system performance for QoS provisioning, and applying adaptive control theories to COPS-PR are at the core of this work. Performance evaluation of PBM with such flavors will be compared with that of the conventional paradigm, to validate the gain of monitoring and control with the novel design.

Prerequisite: Good knowledge of computer networks, and programming skills in C/C++

last changed 2006-07-29, 12:07 by Xiaoyuan Gu