Method for introducing network congestion predictions in policy decision

Abstract

There is provided a Congestion Prediction Engine 'CPE' of a Policy and Charging Control 'PCX'' architecture. The PCC architecture also has a congestion database, a Policy and Charging Enforcement Function “PCEF” device and a Policy and Charging Roles Function “PCRF” node. The CPE includes a processing unit of the CPE which defines a currently applicable congestion prediction. The CPE includes a network; interface unit of the CPE which defines from the congestion database actual congestion data for a latest time period. The processing unit of the CPE determines a subsequent congestion prediction for a subsequent time period by comparing the currently applicable congestion prediction with the actual congestion data and by applying machine learning techniques. The CPE includes a memory of the CPE in which the currently applicable congestion prediction is replaced with the subsequent congestion prediction. The network interface unit of the CPE sends to a network entity the currently applicable congestion prediction so the PCRF can make a policy decision with quality of service “QoS” based on the currently applicable congestion prediction, upon establishment or modification of an IP Connectivity Access Network “IP-CAN” session for a user. A Policy and Charging Rules Function “PCRF” node of a Policy and Charging Control “PCC” architecture for making policy decisions based on network congestion. There is provided a method of a Congestion Prediction Engine “CPE” of a Policy and Charging Control “PCC” architecture and a method of making policy decisions at a Policy and Charging Control “PCC” architecture based on network congestion.