Distributed inter-domain SLA negotiation using Reinforcement Learning

QoS-demanding applications (e.g. telepresence, cloud computing, etc.) are about to become mainstream. One obstacle remains to be tackled: network operators must be able to negotiate end-to-end QoS and configure accordingly their networks in an automated manner. But, other important factors must be consider: QoS delivery must be attractive enough for them and confidentialy on resources states must be respected. To satisfy these requirements, this paper describes distributed algorithms that will allow network operators to negotiate endto-end Service Level Agreements (SLAs, QoS contracts) and optimize globally their revenues while treating requests in a realtime manner.