Policies for Self Tuning Home Networks

A home network (HN) is usually managed by a user who does not possess knowledge and skills required to perform management tasks. When abnormalities are detected, it is desirable to let the network tune itself under the direction of certain policies. However, self tuning tasks usually require coordination between several network components and most of the network management policies can only specify local tasks. In this paper, we propose a state machine based policy framework to address the problem of fault and performance management in the context of HN. Policies can be specified for complex management tasks as global state machines which incorporate global system behavior monitoring and reactions. We demonstrate the policy framework through a case study in which policies are specified for dynamic selection of frequency channel in order to improve wireless link quality in the presence of RF interference.