Cognitive packet network for QoS adaptation of asymmetric connections

Abstract

The Cognitive Packet Network (CPN) is a software defined self-aware and self-organised network system that flexibly offers best-effort Quality of Service (QoS) optimisation. Many network accesses request downloads from some web site, with an Uplink which represents the request and a Downlink for the download, and each of these two connections have distinct minimal QoS needs. More generally pairs of nodes communicating across a network may have distinct QoS goals for each of the directions of communication. Thus this article presents a scheme to introduce into CPN the option of bilateral QoS differentiation between communicating nodes. In the proposed approach, each CPN edge or user node is a source and a destination at the same time, managing Uplink user originated traffic, and Downlink traffic sent back in response to the Uplink. Bilateral communication is implemented with four distinct QoS objectives that can be met between sender nodes (original source or destination). Traffic volume asymmetry between received and sent data is used as a QoS triggering condition, when the lower traffic rate requires short delay QoS, while the higher traffic rate requires loss minimisation. The approach is implemented and illustrated with measurements that show its effectiveness.