Variable Analysis Windows for Shared-IP Download Session Counting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for counting unique podcast downloads inaccurately account for shared IP addresses and user agents, leading to undercounting or overcounting due to fixed lookback windows.
Innovation Solution
An algorithm that dynamically adjusts lookback windows based on the degree of IP overlap to accurately count unique downloads by analyzing server-side data and adjusting lookback windows for each unique device.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a fixed 24-hour lookback window is used to count downloads, then the counting process is simple, but the accuracy of unique download counts deteriorates when IP addresses are shared among multiple devices
Solution Approach 1:
The patent applies dynamics by transitioning from a fixed 24-hour lookback window to a dynamic, adjustable lookback window that adapts based on IP address overlap analysis. The system calculates the actual time span between first and last downloads for each IP address and adjusts the window accordingly, allowing the counting parameter to change based on observed behavior patterns rather than remaining static.
Solution Approach 2:
The patent implements parameter changes by modifying the lookback window duration based on calculated IP overlap factors. Instead of using a constant time parameter, the system varies the window size dynamically according to the specific download patterns and IP address sharing characteristics of each device group, optimizing the counting accuracy for different scenarios.
2Quantity of substance
If a longer lookback window is used to capture all downloads from shared IP addresses, then more downloads are captured, but incorrectly split sessions increase
Solution Approach 1:
The patent applies feedback by using the calculated IP overlap information to adjust the lookback window parameters. The system analyzes download patterns, determines the degree of IP address sharing, and feeds this information back into the counting process by modifying the window size, thereby creating a closed-loop system that continuously optimizes for accuracy based on observed data.
Solution Approach 2:
The patent implements preliminary action by performing IP overlap analysis and calculating appropriate lookback window durations before conducting the actual download counting. This pre-processing step prepares the optimal counting parameters in advance, allowing the system to accurately distinguish between multiple devices sharing an IP address without requiring complex real-time adjustments during the counting process.
3Measurement precision
If a shorter lookback window is used to reduce incorrectly split sessions, then session accuracy improves, but downloads from shared IP addresses are undercounted
Solution Approach 1:
The patent resolves this contradiction through dynamics by making the lookback window adaptive rather than fixed. The system dynamically determines the appropriate window duration for each IP address based on the time span of observed downloads and the calculated overlap factor, allowing the window to expand or contract as needed to capture all relevant downloads while maintaining session integrity.
Data Source
AI summary
The present invention teaches a method for determining an accurate measure of unique download sessions in an environment where IP addresses are shared among multiple devices. According to a first preferred embodiment, the system of the present invention may analyze requested server log data to determine the number of unique client devices sharing the same IP addresses. The system may then dynamically adjust the length of lookback windows for each unique device based on observed download behavior and the number of user agents associated with each overlapping IP address.