Cloud Resource Allocation Sequence Classification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Cloud computing platforms face challenges in detecting malicious resource allocation operations, as existing methods lack effective mechanisms to differentiate between benign and malicious sequences of resource allocation, making them vulnerable to hacking and unauthorized access.

Innovation Solution

A machine learning-based approach is employed to generate a sequence classification model by training on known malicious and benign resource allocation sequences, using honeypot cloud resources to identify malicious intent and classify resource allocation operations as malicious or non-malicious.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional monitoring methods are used to track resource allocation operations, then system simplicity is maintained, but the ability to detect malicious sequences is insufficient

Engineering Contradiction:
Improvedetection accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

A machine learning model is introduced as an intermediary component between the resource allocation monitoring system and the detection process. The model receives resource allocation sequences as input and outputs classification results, enabling sophisticated detection without directly modifying the core monitoring infrastructure.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

Traditional rule-based detection mechanisms are replaced with a machine learning-based classification system. The patent uses supervised learning models that automatically learn patterns from training data, substituting manual rule creation and maintenance with automated statistical learning.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If honeypot cloud resources are deployed to collect malicious sequences, then training data quality improves, but security vulnerability increases

Engineering Contradiction:
Improvetraining data reliabilityVSAvoidsecurity risk
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent converts the security risk of deploying honeypots into a benefit by using them specifically for collecting malicious resource allocation sequences. The potential vulnerability is transformed into a controlled data collection mechanism that actively attracts and captures malicious behavior patterns for training purposes.

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

Honeypot resources are deployed in advance to proactively collect malicious sequences before they can affect production systems. This preliminary data collection enables the machine learning model to be trained on authentic malicious patterns, improving detection capability before real threats reach critical systems.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If machine learning models are trained on extensive resource allocation sequences, then classification accuracy improves, but processing time increases

Engineering Contradiction:
Improveclassification accuracyVSAvoidtraining time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The machine learning model is trained in advance on comprehensive datasets containing both benign and malicious resource allocation sequences. This preliminary training phase occurs offline, allowing extensive processing without impacting real-time detection performance. The trained model is then deployed for rapid online classification.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The training process is segmented into distinct phases: data collection from honeypots and legitimate sources, data preprocessing and labeling, model training, and validation. This segmentation allows parallel processing and optimization of each stage, reducing overall training time while maintaining accuracy.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3834392B1Malicious cloud-based resource allocation detection
Publication Date: 2023.08.02 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP3834392B1 patent drawingFigure 1
  • EP3834392B1 patent drawingFigure 2A
  • EP3834392B1 patent drawingFigure 2B

AI summary

Methods, systems, and computer program products are described herein for detecting malicious cloud-based resource allocations. Such detection may be achieved using machine learning-based techniques that analyze sequences of cloud-based resource allocations to determine whether such sequences are performed with a malicious intent. For instance, a sequence classification model may be generated by training a machine learning-based algorithm on both resource allocation sequences that are known to be used for malicious purposes and resource allocation sequences that are known to be used for non-malicious or benign purposes. Using these sequences, the machine learning-based algorithm learns what constitutes a malicious resource allocation sequence and generates the sequence classification model. The sequence classification model is used to classify any sequence of resource allocation operations performed via a valid user's cloud services subscription provided thereto as being a malicious sequence or a non-malicious sequence.