Research On Policy Based QoS And Dynamic Resource Management In Mobile IP Network

With the development of wireless technology and the large numbers of requirements for the multimedia application, carrying multimedia services, such as video, voice and data, has become an inevitable trend in wireless networks. However, because of the users'mobility and the lack of resources, it is more complex to guarantee QoS of mobile users in the mobile environment. Policy-based Network Management (PBNM) is a new kind of network management scheme and has more flexibility. It emphasizes mamagement automatization and uses the configuration and execution of policies to realize management target. So it can better meet the demands of variability and real-time processing of the mobile IP network and it has a very active role to ensure the mobile users QoS, get a rational allocation of resources by applying PBNM technology in QoS and resource management of mobile IP network.Combining to the project of National Natural Science Foundation of China: Research on Policy-based Dynamic Resource Allocation and Management Mechanism in Mobile IP Handoff(serial number:60573128), This dissertation focuses on correlative problems of applying the policy mechanism to guarantee QoS, implement dynamic resource management. To enable the mobile node to obtain guaranteed QoS, it needs a mechanism to fulfill QoS negotiation When the location or QoS requirements of mobile nodes change. Currently, PBNM does not consider QoS negotiation for QoS guarantee.The dissertation presents a policy-based QoS two-level negotiation framework, and expands COPS as a negotiation protocol between the entities to support mobile node obtaining necessary resources by policy-based QoS negotiation. It provides the basis for following works about QoS guarantee and dynamic resource management in mobile IP network. Then, for low efficiency of the traditional policy configuration and over-reliance of PEP(Policy Enforcement Point) on PDP (Policy Decision Point), it advances meta-policy to improve the policy configuration efficiency and flexibility, and to enhance PEP's intelligence. Based on the meta-policy, it achievs adaptive bandwidth reservation and allocation control in mobile IP network. Although the meta-policy has improved the traditional policy systems, but it is still manually defined by manager. As the mobility and real-time needs of the mobile IP network, if the policy system can automatically generate policy or offer policy parameters, it will meet the requirements of dynamic management of mobile IP network better, for which this dissertation set up a dynamic resource allocation policy model for mobile IP handoff, puts forward the concept of dynamic policy and designs a dynamic policy generating algorithm. The algorithm can automatically generate dynamic policy for adjusting bandwidth allocation based on the mobility prediction information and network status. However, because the mobility prediction information is uncertain, the dynamic policy is also uncertain. it brings forward a adaptive trigger mechanism for the dynamic policy to reduce this uncertainty. It makes the triggered dynamic policy fit actual needs. Finally, according to the problem that the action of traditional policy can not be changed flexibly in the policy execution process, the dissertation puts forward a policy's action updating methods based on the network status during the execution of the policy. This method achieves a dynamic resource allocation during policy implementation and better meets the needs of dynamic resource management in mobile IP network.The main research contents and conclusions in this dissertation can be summarized as follows:1.QoS negotiation is the foundation for mobile node to get necessary resources and obtain guaranteed QoS. Because there is no corresponding negotiation mechanism for policy based QoS management, a policy-based QoS two-level negotiation control framework is put forward to realize two level negotiation in local wireless network and in management domain(current management domain and between the different management domain). It makes a detailed design for the three functional entities:the global policy manager, local policy manager and policy enforcement point, defines the form of negotiation policy and then analyzes the state of the client and server on the basis of the standards document RFC2748 to expand COPS as negotiation protocol between entities. Finally, the experiment is done to simulate the negotiation process. policy-based QoS two-level negotiation prvides a effective way to making mobile QoS management and QoS negotiation mechanism independent with each other. According to the location of mobile handoff nodes, it selects different level negotiation to carry out QoS recovery on the part of the path or QoS rebuilding on end-to-end path. This mechanism increases negotiation flexibility and offers fundamental guarantee for the following works about policy based QoS and resource management in mobile IP network.2.Because of node's mobility, it needs a dynamic resource allocation based on change of network status to guarantee user's QoS.In traditional policy based management, PEP in network devices can obtain the network status information, while PDP gets these information from PEP and then follows management policy to issue the allocation resources decisions to PEP. For PDP intervening any change of the network state, it reducs the implementation efficiency and increases communication between PDP and PEP. To improve traditional scheme, meta-policy is put forward to download corresponding guidance policy about how to deal with the network events from PDP to PEP.The meta-policy is created by administrator and is sent to the PEP from the PDP. When the network events defined in meta-policies occur, PEP needn't communicate with PDP and automatically execute meta-policies to decision the operation (action) according to the current network state (conditions).There are parameters in the condition and action, so it has more flexibility. Further, meta-policies are applied in the process of resource management in mobile IP handoff. The arrival of mobile prediction triggers bandwidth reservation and the arrival of call triggers bandwidth allocation.Using meta-policies, it achieve adaptive bandwidth reservation and allocation in PEP to provide better QoS guarantee by minimizing new call blocking rate under the premise of ensuring the certain handoff dropping probability.3.As the mobile user's random remotion, the requirement to network resources is dynamic changeable. If the policy system can automatically generate policy based on the changes of the external environment, it will meet the requirements of dynamic management better. So a dynamic policy scheme of resource management for mobile IP handoff is proposed, and the concept and reference syntax of dynamic policy is defined. Then, it focuses on the dynamic policy generation mechanism and designs the generation algorithm. On the basis of the prediction information of the mobile node, the mechanism produces dynamic policies automatically to do admission control and prepare resources for handoff before it happens. When the resources of access domain can't afford mobile nodes'requisition, it adopts"step back"arithmetic to adjust agilely resource possession according to the priority under assuring the service credibility. By analysis and simulations, it is shown to ensure the QoS of applications with higher priority and improve the success rate of handoff, and then the management scheme is more flexible than traditional methods.4. The generation algorithm of dynamic policy presented in previous segment generates dyanmic policy based on the mobile prediction information. Due to the prediction information is uncertain and dynamic policy is also uncertain, so the prediction accuracy has a significant impact on the validity of dynamic policy. A adaptive trigger mechanism of dynamic policy is proposed to reduce this uncertainty. It adopts high-level trigger policies and process functions to select dynamic policies meeting the requirements. Dynamic policy choices aims at deal with the uncertainty of the predicted next cell and the expected handoff time. It determines what policies can be triggered and when they will execute using the approach of adjusting reservation bandwidth according to handoff probability priority and resizing adaptive trigger time, and reduces the uncertainty influence. It employs less resources reservation to guarantee handoff call QoS, balances the handoff dropping probability and new call blocking probability, and increases resource utilization rate.5.In traditional execution way of policy, policies can not be modified in the period of its execution and removing existed policy is only by sending a new policy. As to the dynamic policy, although it is created automatically by system, its implementation is not changed. However, the implementation of policy is to achieve many services degrading and promotion operations in dynamic resource allocation process, and the new policy is different with old policy only in action parameters. A dynamic policy updating scheme is proposed to modify some of the parameters of actions based on network status during policy running. In this way, the amount of resources occupied by the ongoing services is flexiblly change in accordance with the network load. When available resources is abundant, the user can as far as posiible to obtain the best services and the other way round, the services is degraded in reasonable degree. It enhances the flexibility of management, and effectively improves the utilization rate of the network resources.