DSL Profile Management via User Activity Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dynamic Line Management (DLM) techniques face challenges in accurately assessing DSL connection stability due to misleading metrics, particularly the number of re-synchs, which can be caused by user actions rather than line instability, leading to frequent profile changes and service disruptions.
Innovation Solution
A method that aggregates connection data over multiple monitoring periods, discounting performance indications during periods of user inactivity to estimate the actual performance of DSL connections, allowing for more reliable profile selection and minimizing unnecessary re-synchronizations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of re-synchs is used as a metric to monitor line performance, then line instability can be detected, but user actions cause false positives leading to unnecessary profile changes
Solution Approach 1:
The patent introduces user activity status as an intermediary variable that mediates between the re-synch metric and profile change decision. By checking whether the user is active or inactive before acting on re-synch events, the system filters out false positives caused by user actions during inactive periods, thereby improving both measurement precision and profile change reliability
Solution Approach 2:
Instead of directly responding to re-synch events as indicators of line instability, the patent inverts the logic by first checking user activity status and only treating re-synchs as instability indicators when the user is active. This inversion prevents misinterpretation of re-synch causes and improves the reliability of profile changes
2Reliability
If profile changes are made frequently to respond to detected instability, then connection stability can be improved, but service disruptions increase due to re-synchs
Solution Approach 1:
The patent performs preliminary checking of user activity status before initiating profile changes. By anticipating whether a re-synch event is user-caused or line-instability-caused before acting, the system avoids unnecessary profile changes and their associated service disruptions, thereby reducing time loss while maintaining connection stability for genuine issues
Solution Approach 2:
The patent converts potentially harmful false positive profile changes into beneficial actions by using user activity status as a filter. Re-synch events during inactive periods, which would normally trigger unnecessary profile changes and service disruptions, are now recognized as non-problematic and ignored, turning a harmful metric into a beneficial filtering mechanism
3Reliability
If monitoring is performed continuously to detect instability early, then re-synchs can be prevented, but false alarms increase during user inactivity
Solution Approach 1:
The patent applies local quality by treating performance monitoring differently based on user activity state. During active periods, re-synch events are treated as significant instability indicators warranting profile changes, while during inactive periods, the same events are treated as non-significant. This localized treatment based on user state filters out false alarms while maintaining sensitivity to genuine issues
Solution Approach 2:
Instead of assuming all re-synch events indicate line instability, the patent inverts the assumption by presumes user inactivity may cause re-synchs and only treats them as instability indicators when the user is active. This inversion eliminates false performance indications during inactive periods while maintaining accurate detection during active periods
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods and apparatus are disclosed for operating an access network including data connections (19) between end-user devices (12, 14, 16, 18) and an aggregation transceiver (20) at which connections (19) are aggregated for onward connection through the network. The method comprises, for each connection, obtaining connection data for each of a plurality of monitoring periods, each comprising shorter periods, the connection data comprising performance and user activity indications for each shorter period; monitoring the performance of the connection in order to estimate a measure of its performance during a particular monitoring period that discounts performance indications obtained in respect of shorter periods within that monitoring period when the user was inactive unless activity indications in respect of corresponding shorter periods within other monitoring periods indicate that such user inactivity is anomalous; and applying a profile to the connection that is dependent on the measure of performance of the data connection.