HTTP Request Classifier for Cloud Action Detection Under API Change
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud service access management systems struggle with accurately classifying actions due to frequent changes in back-end APIs, leading to inefficiencies and increased risks of data loss, as they often rely on outdated blacklists or hard-coded rules that fail to adapt to service updates.
Innovation Solution
A machine learning-based multiclass classifier system that analyzes HTTP network requests to identify actions such as upload or download, allowing for real-time classification and adaptive policy enforcement, enabling precise control over cloud service usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If blacklisted cloud services or hard-coded rules are used to block requests, then access control is simplified, but the system cannot adapt to API changes and becomes outdated
Solution Approach 1:
The patent replaces mechanical blacklisting systems with a machine learning-based classification system. Instead of maintaining static rulebooks that break when APIs change, the system uses ML models that automatically learn from traffic patterns and adapt to new API versions without manual intervention.
Solution Approach 2:
The machine learning system performs self-learning and self-updating functionality. The model continuously processes new traffic data, automatically identifies patterns, and updates its classification capabilities without requiring administrators to manually discover and update rules for changing APIs.
2Reliability
If blacklists are frequently updated to track cloud service changes, then blocking accuracy is improved, but the process becomes costly and inefficient
Solution Approach 1:
The machine learning system automatically learns from traffic patterns and updates its models without human intervention. The system processes traffic data in real-time, identifies new API patterns, and retrain models as needed, eliminating the manual process of discovering and updating blacklists.
Solution Approach 2:
The system continuously monitors traffic and uses feedback loops to refine its classifications. By analyzing actual traffic patterns and outcomes, the model automatically adjusts its understanding of cloud service APIs, improving accuracy without requiring manual updates.
3Reliability
If all requests to blacklisted services are blocked, then security is improved, but legitimate feature access is prevented
Solution Approach 1:
Instead of blocking entire services uniformly, the patent applies granular classification at the request level. The machine learning model analyzes individual request characteristics and applies different policies to different features or endpoints, allowing selective access to specific functionalities while blocking others.
Solution Approach 2:
The system changes the parameter of classification precision from binary (block/all-or-nothing) to multi-class classification. By predicting specific action types and service contexts, the system can apply differentiated policies based on the specific parameters of each request, enabling fine-grained control over feature access.
Data Source
AI summary
Embodiments include an activity monitoring machine learning model method. One embodiment the method includes transforming HTTP network requests into feature vectors, each feature vector representing selected features from a corresponding HTTP network request and an action selected from a plurality of actions to be monitored and inputting the feature vectors into a machine learning model to train the machine learning model to classify new HTTP requests according to the plurality of actions, wherein the plurality of actions include an upload action and a download action.