Network cell classification for QoS policy optimization
Wireless network cells are classified into groups based on radio and network conditions in the cells. Quality of Service, QoS, policy optimization is then performed per each classified cell group, by selecting more or less stringent predetermined rules until a target Quality of Experience, QoE, is met, at least for the highest priority traffic. To classify the cells into groups, the radio and network parameters collected ( 110 ) are ranked ( 112 ) by their influence on QoE. Based on this ranking, only a few of the most impactful parameter are selected ( 114 ). The range of values of each parameter is then divided ( 116 ) into a plurality of ranges, for example by the ability of the parameter, when within each range, to impact service quality. Cells are then classified ( 118 ) into groups based on the same parameters within the cells falling into the same value ranges. QoE targets for at least high priority traffic classes are obtained ( 120 ). QoS policy is then optimized ( 122 ), separately for each classified cell group, to meet the QoE targets. The optimized QoS policy is applied ( 124 ) to a new UE entering a cell.
1 . A method of classifying cells in a wireless communication network for Quality of Service (QoS) policy optimization, the method characterized by:
obtaining, per cell, measurements of radio network and traffic load parameters, and estimates of user Quality of Experience (QoE);
ranking the radio network and traffic load parameters by their influence on QoE;
selecting a subset of the highest ranked parameters;
determining a plurality of value ranges for each of the selected parameters; and
classifying cells into groups based on the value ranges of the selected parameters.
2 . The method of claim 1 further characterized by:
obtaining QoE targets for at least high priority traffic classes; and
optimizing QoS policy for the high priority traffic classes, separately for each classified cell group, to meet the QoE targets.
3 . The method of claim 2 , further characterized by implementing the optimized QoS policy for each cell in each classified cell group through a network Policy Charging Function.
4 . The method of claim 3 , further characterized by applying a cell's optimized QoS policy to a User Equipment entering the cell from a cell having a different QoS policy.
5 . The method of claim 2 wherein optimizing QoS policy for the high priority traffic classes separately for each classified cell group to meet the QoE targets comprises, separately for each classified cell group:
if the estimated QoE for the high priority traffic classes is less than the target QoE, applying a more stringent QoS Indicator; and
if the estimated QoE for the high priority traffic classes exceeds the target QoE by a predetermined threshold or more, applying a less stringent QoS Indicator.
6 . The method of claim 2 further characterized by merging two categorized cell groups if the determined polices in each group are the same over a predetermined duration.
7 . The method of claim 1 wherein ranking the radio network and traffic load parameters by their influence on QoE comprises:
for each parameter, calculating a ratio of a change in QoE per change in parameter value over a normalized value range of the parameter values;
comparing the ratios; and
ranking each parameter according to its associated ratio.
8 . The method of claim 1 wherein determining the plurality of value ranges for each of the selected parameters comprises determining three value ranges, wherein:
the parameter values over a first range cause a QoS degradation;
the parameter values over a second range cannot cause a QoS degradation; and
the parameter values over a third range lie between the first and second ranges, and the parameter values over the third range could cause a QoS degradation.
9 . The method of claim 1 wherein determining the plurality of value ranges for each of the selected parameters comprises one of:
applying predetermined thresholds to the parameter values; and
ranking a plurality of cells by the measured values of a parameter and dividing the parameter values into ranges according to predetermined percentile values of the cell ranking.
10 . The method of claim 1 wherein determining the plurality of value ranges for an identified one of the selected parameters comprises:
determining the plurality of value ranges for selected parameters other than the identified selected parameter;
determining a threshold value for the first range of the identified parameter by fitting a QoE curve as a function of the identified parameter wherein all other selected parameter values are in the second range and wherein the threshold value for the first range of the identified parameter is the value at which the fitted QoE curve equals a target QoE; and
determining a threshold value for the third range of the identified parameter by fitting a QoE curve as a function of the identified parameter using all other selected parameter values and wherein the threshold value for the third range of the identified parameter is the value at which the fitted QoE curve equals the target QoE.
11 . A network node operative in a wireless communication network in which Quality of Service (QoS) policies are enforced, characterized by:
communication circuitry configured to communicate with other network nodes; and
processing circuitry operatively connected to the communication circuitry and configured to:
obtain, per cell, measurements of radio network and traffic load parameters, and estimates of user Quality of Experience (QoE);
rank the radio network and traffic load parameters by their influence on QoE;
select a subset of the highest ranked parameters;
determine a plurality of value ranges for each of the selected parameters; and
classify cells into groups based on the value ranges of the selected parameters.
12 . The network node of claim 11 , wherein the processing circuitry is further configured to:
obtain QoE targets for at least high priority traffic classes; and
optimize QoS policy for the high priority traffic classes, separately for each classified cell group, to meet the QoE targets.
13 . The network node of claim 12 , wherein the processing circuitry is further configured to implement the optimized QoS policy for each cell in each classified cell group through a network Policy Charging Function.
14 . The network node of claim 13 , wherein the processing circuitry is further configured to apply a cell's optimized QoS policy to a User Equipment entering the cell from a cell having a different QoS policy.
15 . The network node of claim 12 wherein the processing circuitry is configured to optimize QoS policy for the high priority traffic classes separately for each classified cell group to meet the QoE targets by, separately for each classified cell group:
if the estimated QoE for the high priority traffic classes is less than the target QoE, applying a more stringent QoS Indicator; and
if the estimated QoE for the high priority traffic classes exceeds the target QoE by a predetermined threshold or more, applying a less stringent QoS Indicator.
16 . The network node of claim 12 wherein the processing circuitry is further configured to merge two categorized cell groups if the determined QoS polices in each group are the same over a predetermined duration.
17 . The network node of claim 11 , wherein the processing circuitry is configured to rank the radio network and traffic load parameters by their influence on QoE by:
for each parameter, calculating a ratio of a change in QoE per change in parameter value over a normalized value range of the parameter values;
comparing the ratios; and
ranking each parameter according to its associated ratio.
18 . The network node of claim 11 wherein the processing circuitry is configured to determine the plurality of value ranges for each of the selected parameters by determining three value ranges, wherein:
the parameter values over a first range cause a QoS degradation;
the parameter values over a second range cannot cause a QoS degradation; and
the parameter values over a third range lie between the first and second ranges, and the parameter values over the third range could cause a QoS degradation.
19 . The network node of claim 11 wherein the processing circuitry is configured to determine the plurality of value ranges for each of the selected parameters by one of:
applying predetermined thresholds to the parameter values; and
ranking a plurality of cells by the measured values of a parameter and dividing the parameter values into ranges according to predetermined percentile values of the cell ranking.
20 . The network node of claim 11 wherein the processing circuitry is configured to determine the plurality of value ranges for an identified one of the selected parameters by:
determining the plurality of value ranges for selected parameters other than the identified selected parameter;
determining a threshold value for the first range of the identified parameter by fitting a QoE curve as a function of the identified parameter wherein all other selected parameter values are in the second range and wherein the threshold value for the first range of the identified parameter is the value at which the fitted QoE curve equals a target QoE; and
determining a threshold value for the third range of the identified parameter by fitting a QoE curve as a function of the identified parameter using all other selected parameter values and wherein the threshold value for the third range of the identified parameter is the value at which the fitted QoE curve equals the target QoE.