ISP-Level Web Page Ranking via DNS Session Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing page ranking techniques, such as PageRank and BlockRank, are inefficient, unreliable, and fail to accurately reflect the relevance and popularity of web pages due to their reliance on static link structures, lack of consideration for traffic flow, and susceptibility to manipulation, leading to outdated and biased rankings that do not account for user behavior or geographic location.
Innovation Solution
A method and system for ranking pages and hosts at the ISP level using DNS data to create session maps and traffic flow models, which provide timely, relevant, and geographically specific search results by analyzing user requests and traffic patterns, eliminating reliance on global search engines and ad content providers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If PageRank algorithm is used to rank web pages, then page popularity can be measured based on link structure, but the ranking becomes outdated and computationally expensive
Solution Approach 1:
The patent pre-computes and stores the normalized hyperlink matrix H and its dominant eigenvector π during an offline phase. This preliminary action allows the system to avoid performing expensive iterative PageRank computations at query time, instead using the pre-computed eigenvector for rapid page ranking decisions.
Solution Approach 2:
The patent separates the ranking system into two distinct phases: an offline phase for computing the normalized hyperlink matrix and its eigenvector, and an online query phase for using pre-computed data. This segmentation allows computationally intensive operations to be performed only once rather than repeatedly for each query.
2Reliability
If static link structure is used for ranking, then page importance can be determined, but user behavior and traffic flow information is lost
Solution Approach 1:
The patent merges two previously separate ranking approaches: the static link-based PageRank method and the dynamic traffic-based TrafficRank method. By combining the normalized hyperlink matrix H with the traffic flow matrix T, the system produces a unified ranking that reflects both structural importance and actual user behavior patterns.
Solution Approach 2:
The patent creates a composite ranking model that integrates multiple data sources (hyperlink structure and traffic flow) into a unified framework. This composite approach uses the strengths of both methods: the structural stability of PageRank and the behavioral relevance of TrafficRank, producing more comprehensive and accurate rankings.
3Adaptability or versatility
If global search engines are used for ranking, then comprehensive results can be provided, but geographic and user-specific relevance is reduced
Solution Approach 1:
The patent implements location-aware ranking by incorporating the geographic location of users into the ranking process. The system adjusts page rankings based on the user's location, prioritizing locally relevant results while maintaining access to global content. This creates different ranking qualities for different user locations rather than a single uniform ranking.
Data Source
AI summary
Systems and methods for ranking pages and/or hosts in a faster and more relevant manner are provided. Systems and methods for ranking pages and/or hosts based on session data and/or traffic data are also provided. According to the invention, session maps can be created using DNS and/or ISP data. Systems and methods for ranking pages and/or hosts for the purpose of doing business are also provided.