Network Request Classification via Multi-Dimensional Key Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for classifying network requests using blacklist and whitelist classifiers based solely on IP addresses are ineffective, as they either overly restrict legitimate traffic or fail to protect against malicious requests from infected devices.
Innovation Solution
Implementing a system that classifies network requests using both IP and non-IP keys, such as URI, user, HTTP, and location keys, with dynamic classifiers that can be configured as blacklist or whitelist classifiers to determine whether to allow or deny requests based on predefined rules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a blacklist classifier is used to deny all network requests from an IP address identified by the blacklist rule, then network protection against malicious software is improved, but legitimate user operations are restricted
Solution Approach 1:
The patent segments the classification criteria from a single IP address into multiple dimensions including IP address, URI, user identifier, device identifier, location, and timestamp. This multi-dimensional segmentation allows the system to distinguish between malicious and legitimate requests more accurately, preventing over-blocking of legitimate user operations while maintaining network protection.
Solution Approach 2:
The patent implements dynamic classification rules that can be configured as whitelist, blacklist, or score-based rules with adjustable thresholds. The classifier dynamically adapts its behavior based on the combination of multiple keys and their matching results, rather than statically denying all traffic from a single IP address. This dynamic approach resolves the contradiction by allowing legitimate operations to pass through when other dimensions indicate normal behavior.
2Ease of operation
If a whitelist classifier is used to allow all network requests from an IP address identified by the whitelist rule, then user access is improved, but network protection against malicious requests is reduced
Solution Approach 1:
The patent segments the security verification into multiple independent key dimensions (IP address, URI, user identifier, device identifier, location, timestamp) rather than relying solely on IP address whitelisting. This segmentation allows the system to verify multiple aspects of a request simultaneously, maintaining user access convenience while enhancing network protection through comprehensive validation.
Solution Approach 2:
The patent merges multiple classification keys and rules into a unified classification system that combines whitelist, blacklist, and score-based rules. By merging these different rule types and evaluating them together based on configurable weights and thresholds, the system achieves both ease of access for legitimate users and robust network protection against malicious requests.
3Device complexity
If classification is based solely on IP addresses, then implementation simplicity is improved, but classification accuracy is reduced
Solution Approach 1:
The patent creates a universal classification framework that can handle multiple types of keys (IP address, URI, user identifier, device identifier, location, timestamp) through a unified rule structure. This multi-functional design maintains implementation simplicity by using a consistent classification mechanism for all key types while significantly improving classification accuracy through the additional dimensions of analysis.
Data Source
AI summary
Methods, systems and programming for classifying network requests. In one example, a network request for content to be fetched by a content server is received from a client device. At least one non-internet protocol (IP) key is obtained based on the network request. Whether to deny or allow the network request is determined based on the at least one non-IP key.