Beside IPv6 migration, the greatest challange of internet service providers nowadays is controlling peer-to-peer traffic. Handling this kind of traffic is a remarkable problem for the providers either we are talking about financial or legal aspects. At the moment there is not any suitable and economical way of handling traffic generated by peer-to-peer applications. Some of the methods are requering too much computing capacity, or there are legal anxieties against using them. Furthermore ISPs can not presume on voluntary cooperation of customers.
In my thesis I will present a framework, which internet service providers can use to reduce costs rose from relaying peer-to-peer traffic without affecting satisfaction level of customers. The framework, which is based on traffic behaviour and patterns, can make real-time routing decisisions based on passing flows, and apply these decisions automatically on the network appliances in the infrastructure of the provider. The framework also provides an easy-to-use interface for network administrators of the providers, where they can easily configure the required routing decisions. Moreover the framework is scalable, and it can be improved easily. Consequently results of new research projects can be implemented with ease.