System Call Attack Detection Using Instrumented Delimiters

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems struggle to effectively link kernel-level system call analysis to user-level program attacks, missing vulnerabilities and failing to accurately identify the program sections causing the attacks due to the difference between user and kernel levels.

Innovation Solution

Instrumentation is added to a program to include delimiters in system call streams, linking attacks to specific program portions by augmenting the system call stream with application-level data, using a delimiter to separate uninteresting from targeted program sections, and employing an attack model to detect malicious activity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If system call requests are analyzed independently at the kernel level, then anomaly detection can be performed, but the analysis misses vulnerabilities and fails to link attacks to the corresponding program sections

Engineering Contradiction:
Improveattack detection accuracyVSAvoidprogram section linkage information
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent introduces delimiters as intermediary markers that are inserted into the system call stream at specific program locations. These delimiters act as mediators that connect kernel-level system call observations back to user-level program sections, enabling precise linkage between detected anomalies and their originating code without requiring direct access to program execution context at the kernel level

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If instrumentation is added to link system calls to program portions, then attack linkage is improved, but system complexity increases

Engineering Contradiction:
Improveprogram section identificationVSAvoidsystem instrumentation complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the program into specific portions of interest and inserts delimiters only at the boundaries of these segments rather than instrumenting the entire program. This selective segmentation approach maintains precise program section identification while minimizing the complexity overhead by focusing instrumentation only on critical code regions

Inventive Principle:
Principle #1Segmentation

3Reliability

If static or dynamic analysis is performed on the program, then vulnerabilities can be detected, but the analysis is time-consuming and may miss runtime attacks

Engineering Contradiction:
Improvevulnerability detectionVSAvoidanalysis time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary action by inserting delimiters into the system call stream during program execution setup, before runtime attacks occur. This preliminary instrumentation enables real-time attack detection during execution without requiring time-consuming static analysis, while the delimiters are pre-positioned to enable rapid identification of attacked program sections

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS20260044595A1Attack detection with application-aware system call analysis
Publication Date: 2026.02.12 ORACLE INT CORP
  • US20260044595A1 patent drawing
  • US20260044595A1 patent drawing
  • US20260044595A1 patent drawing

AI summary

The operations include obtaining a system call stream from executing an instrumented program comprising an instrumented portion. The instrumented portion comprises instrumentation to add. The operations further include monitoring the system call stream for a start delimiter defined by instrumentation in the instrumented program, extracting a system call trace starting at the start delimiter, processing the system call trace through an attack model to obtain an attack probability, detecting an attack based on a comparison of the attack probability with an attack detection threshold, and generating an alert identifying the attack and the instrumented portion.