Application data flow management in an IP network

摘要

A method for serving an aggregate flow in a communication network node includes a plurality of individual flows. The method includes identifying in the aggregate flow, based on serving resources allocated to the network node, individual flows that may be served without substantial detriment to perceived performance, and serving the identified individual flows with priority with respect to the remaining individual flows in the aggregate flow. The method allows the presence of individual flows that may not be served without substantial detriment to perceived performance due to shortage of serving resources to be notified to an external control entity.

主权项

1. A method for serving an aggregate flow in a communication network node having allocated serving resources, the aggregate flow comprising a plurality of individual flows having associated serving resource needs and serving priorities, comprising:
identifying in the aggregate flow, based on the serving resources allocated to the network node, a set of individual flows that may be served according to the serving resource needs of individual flows within the set and placing the set of individual flows in a serving queue; identifying in the aggregate flow a second individual flow from the remaining individual flows other than the identified set and placing the second individual flow in a staging queue; automatically updating the composition of the set of individual flows, based on changes to the aggregate flow; serving the updated set of individual flows based on the associated serving priorities and on a serving priority criterion, by sharing the serving resources allocated to the network node such that individual flows are served by meeting the serving resource needs of individual flows within the updated set; serving the second individual flow in the staging queue using remaining serving resources allocated to the network node; determining whether data loss occurs in the staging queue; and when it is determined that no data loss occurs, upgrading the second individual flow from the staging queue to the serving queue.