Network Usage Profiling via Partial Packet Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network traffic inspection methods are limited by high computational and storage resource requirements and may slow network traffic, leading to incomplete data analysis, as they often only inspect header and payload information, missing crucial device-level data and failing to account for inactive devices.
Innovation Solution
Generating a network usage profile that estimates the set of customer devices available for communication with a customer-side network node, including active and inactive devices, by processing device data and creating devices-by-node profiles, which can be used to affect various network services such as content delivery and security.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep-packet inspection is implemented to fully inspect each data packet, then measurement precision is improved, but productivity deteriorates due to high computational resources and network traffic slowdown
Solution Approach 1:
The patent applies partial inspection by selecting only certain data packets for deep inspection based on probability sampling rather than inspecting every packet. This reduces computational overhead and maintains network speed while still obtaining statistically significant insights into network traffic patterns and user behavior.
Solution Approach 2:
The system performs preliminary classification of data packets into inspection and non-inspection groups before actual inspection occurs. By pre-selecting which packets to analyze based on sampling criteria, the system prepares the inspection process in advance, reducing real-time processing demands and preventing network slowdowns.
2Productivity
If header and payload information only is inspected, then productivity is maintained, but measurement precision deteriorates as device-level data is missed
Solution Approach 1:
The patent implements partial deep-packet inspection by selecting specific packets for thorough analysis including device-level information, rather than inspecting all packets at full depth. This selective approach recovers device-level data for a representative sample while maintaining overall network processing efficiency.
Data Source
AI summary
Methods, systems, devices, and software are disclosed for generating a network usage profile. Certain embodiments of the network usage profile include a devices-by-node profile, indicating the set of customer devices available for use in communicating with a customer-side network node located at a customer side of an access network over a period of time, where some of the customer devices are not in operative communication with the customer-side network node during a portion of that time. Other embodiments associate the network usage profile with customer information to generate device-by-customer profiles. Still other embodiments associate the network usage profile with network traffic information to generate traffic-by-device profiles. Even other embodiments associate the multiple sources and types of information to generate traffic-by-customer profiles and/or traffic-by-device-by-customer profiles. Any of the profiles may then be accessed by one or more parties for use in affecting various network services, including targeting content delivery.