Systems and methods for robustly determining time series relationships in wireless networks
Quality of service problems in wireless networks can be identified, and in some cases resolved, autonomously using an optimization technique that adjusts parameters of the wireless network based on relationships between key quality indicators (KQIs) and key performance indicators (KPIs). A KQI may be used to gauge the quality of service/experience collectively observed by users/devices when communicating a particular type of traffic in a wireless network or wireless network area. A KPI may be any specific performance metric of a wireless network tending to have a causal or correlative relationship with a KQI. Different types of relationships between a KQI and KPIs may be used to evaluate a quality of service problem, including correlation coefficients, slopes of linear regression, hit-ratios, and others.
1. A method for adjusting parameters of a wireless network, the method comprising:
receiving, by a controller in the wireless network, a key quality indicator (KQI) and two or more key performance indicators (KPIs) associated with wireless transmissions in the wireless network during a first period;
identifying, by the controller, one or more subsets of discrete time intervals in the first period in which the KQI fails to satisfy quality criteria;
calculating, by the controller, relationship indicators between the KQI and the two or more KPIs during the one or more subsets of discrete time intervals, the relationship indicators include hit-ratio correlation coefficients indicating degrees to which a difference between the KQI and a quality threshold correlates to a difference between a corresponding one of the two or more KPIs and a corresponding performance threshold; and
adjusting, by the controller, configuration parameters that are to be used to operate the wireless network during a second or subsequent period based on the relationships between the KQI and the two or more KPIs.
2. The method of claim 1 , wherein the relationship indicators include correlation coefficients indicating degrees to which a value of the KQI is related to values of corresponding ones of the two or more KPIs.
3. The method of claim 1 , wherein the relationship indicators include slopes of linear regression indicating degrees in which a change in a value of the KQI is attributable to changes in values of the two or more KPIs.
4. The method of claim 1 , wherein the relationship indicators include hit-ratios indicating a ratio of a number of instances in which both a value of the KQI exceeds a quality threshold and a value of a corresponding one of the two or more KPIs exceeds a corresponding performance threshold to a total number of instances in which the value of the KQI exceeds the quality threshold.
5. The method of claim 1 , wherein the one or more subsets of discrete time intervals includes at least a first subset of discrete time intervals in which the KQI fails to satisfy a lower quality criteria, and a second subset of discrete time intervals in which the KQI fails to satisfy a higher quality criteria, the first subset of discrete time intervals including at least one time interval excluded from the second subset of discrete time intervals.
6. The method of claim 5 , wherein determining relationships between the KQI and the two or more KPIs during the one or more subsets of discrete time intervals comprises:
calculating a first set of relationship indicators between the KQI and the two or more KPIs during the first subset of discrete time intervals;
calculating a second set of relationship indicators between the KQI and the two or more KPIs during the second subset of discrete time intervals; and
determining a trend between at least the first set of relationship indicators and the second set of relationship indicators.
7. The method of claim 1 , wherein adjusting the configuration parameters of the wireless network in accordance with the relationships between the KQI and the two or more KPIs comprises:
ranking the two or more KPIs based on the relationship indicators; and
adjusting the configuration parameters to improve two or more of the higher-ranked KPIs.
8. The method of claim 1 , wherein determining the relationships between the KQI and the two or more KPIs comprises:
identifying a first subset of discrete time intervals in the first period in which both a value of the KQI exceeds a quality threshold and a value of a first KPI exceeds a first performance threshold;
identifying a second subset of discrete time intervals during the first period in which both the value of the KQI exceeds the quality threshold and a value of a second KPI exceeds a second performance threshold; and
calculating a first relationship indicator between the KQI and the first KPI during the first subset of discrete time intervals and a second relationship indicator between the KQI and the second KPI during the second subset of discrete time intervals.
9. The method of claim 8 , wherein adjusting the configuration parameters of the wireless network in accordance with the relationships between the KQI and the two or more KPIs comprises:
adjusting the configuration parameters to improve the first KPI when the first relationship indicator exceeds the second relationship indicator; and
adjusting the configuration parameters to improve the second KPI when the second relationship indicator exceeds the first relationship indicator.
10. The method of claim 1 , wherein determining the relationships between the KQI and the two or more KPIs comprises:
identifying a first subset of discrete time intervals in the first period in which both a value of the KQI exceeds a first quality threshold and a value of a first KPI exceeds a first performance threshold;
identifying a second subset of discrete time intervals in the first period in which both the value of the KQI exceeds a second quality threshold and the value of the first KPI exceeds a second performance threshold, the second quality threshold being different than the first quality threshold, and the second performance threshold being different than the first performance threshold; and
calculating a first relationship indicator between the KQI and the first KPI during the first subset of discrete time intervals and a second relationship indicator between the KQI and the first KPI during the second subset of discrete time intervals.
11. The method of claim 10 , wherein the first quality threshold and the first performance threshold correspond to a same percentile in values of the KQI and the first KPI, respectively, and
wherein the second quality threshold and the second performance threshold correspond to the same percentile in values of the KQI and the first KPI, respectively.
12. The method of claim 1 , wherein determining the relationships between the KQI and the two or more KPIs comprises:
preprocessing the KQI wherein preprocessing includes normalization; and
preprocessing the two or more KPIs wherein preprocessing includes normalization, and a direction adjustment of 1 minus a value of a normalized value of a KPI if the KPI is not along a same direction as the KQI when the wireless network is in a normal state, wherein the KPI not along the same direction as the KQI corresponds to a decrease in the KQI in response to an increase in the KPI or to an increase in the KQI in response to a decrease in the KPI.
13. An apparatus in a wireless network comprising:
a non-transitory memory storage comprising instructions; and
a processor in communication with the non-transitory memory storage, wherein the processor executes the instructions to:
receive a key quality indicator (KQI) and two or more key performance indicators (KPIs) associated with wireless transmissions in the wireless network during a first period;
identify one or more subsets of discrete time intervals in the first period in which the KQI fails to satisfy quality criteria;
calculate relationship indicators between the KQI and the two or more KPIs during the one or more subsets of discrete time intervals, the relationship indicators include hit-ratios indicating a ratio of a number of instances in which both a value of the KQI exceeds a quality threshold and a value of a corresponding one of the two or more KPIs exceeds a corresponding performance threshold to a total number of instances in which the value of the KQI exceeds the quality threshold; and
adjust configuration parameters that are to be used to operate the wireless network during a second or subsequent period based on the relationships between the KQI and the two or more KPIs.
14. The apparatus of claim 13 , wherein the relationship indicators include correlation coefficients indicating degrees to which a value of the KQI is related to values of corresponding ones of the two or more KPIs.
15. The apparatus of claim 13 , wherein the relationship indicators include slopes of linear regression indicating degrees in which a change in a value of the KQI is attributable to changes in values of the two or more KPIs.
16. The apparatus of claim 13 , wherein the relationship indicators include hit-ratio correlation coefficients indicating degrees to which a difference between the KQI and a quality threshold correlates to a difference between a corresponding one of the two or more KPIs and a corresponding performance threshold.