P2P Contact Recommendation via Distributed Hash Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In large peer-to-peer (P2P) communication systems, finding and recommending mutual contacts becomes resource-intensive and unfeasible due to the increasing number of users and shared usernames, particularly when conventional Friends of Friends search methods require significant processing power and central database operations.
Innovation Solution
A method that offloads search operations to individual P2P nodes, allowing parallel comparison of contact lists using hash values to identify common contacts without revealing true identities, thus reducing resource consumption and maintaining privacy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional Friends of Friends search is applied to large P2P systems, then contact recommendations can be provided, but processing power requirements become prohibitively high
Solution Approach 1:
The patent divides the contact recommendation task into segments distributed across multiple P2P nodes. Instead of one node performing all database operations, each node handles a portion of the search by comparing contact lists locally, then sharing results. This segmentation reduces the processing burden on any single node while maintaining system-wide recommendation capability.
Solution Approach 2:
The patent introduces hash values as intermediaries to represent contact identifiers. These hash values are transmitted between nodes to enable comparison of contact lists without transferring actual contact information. The hash values act as mediators that allow nodes to identify common contacts while preserving privacy and reducing data transmission requirements.
2Measurement precision
If contact lists are compared to find mutual friends, then relevant contact recommendations are found, but the number of database operations increases with system size
Solution Approach 1:
The patent extracts only the essential information needed for comparison - hash values representing contact identifiers - from the full contact lists. By working with these extracted hash values rather than complete contact data, the system reduces the volume of data that must be stored, transmitted, and processed in database operations.
Solution Approach 2:
The contact comparison operation is segmented and distributed across multiple nodes. Each node performs local comparison of contact lists to identify mutual friends, then shares these localized results. This divides the total database operations into smaller, manageable tasks distributed throughout the network rather than requiring a single node to process all comparisons.
3Measurement precision
If usernames are made unique in large P2P systems, then identification accuracy is maintained, but the number of users with non-real names increases
Solution Approach 1:
The patent uses hash values as intermediaries to represent usernames and contact identifiers. These hash values maintain uniqueness for identification purposes while allowing users to choose any username they desire. The hash values serve as the technical identifier that ensures uniqueness, decoupling this requirement from the visible username displayed to users.
Data Source
AI summary
In one embodiment, a method of transmitting contact data in a peer to peer communications network includes transmitting a list of contacts from a first contact store to a first compare means. At the first compare means, the list of contacts from the first contact store is compared with a list of contacts from a second contact store to identify common contacts between the lists of contacts from the first and second contact stores, the first compare means outputting a contacts result based on the identification of the common contacts. At a second compare means, the contacts result output from the first compare means is compared with a list of contacts from a third contact store. An output of the second compare means is used to identify at least one contact to recommend as a new contact to a recommendee user in the network. The first contact store is at a node of the network and the first compare means is at another node of the network, and the list of contacts transmitted from the first contact store is transmitted over the network to the first compare means.