Autonomic Computing Measurement via Segmented Detection Intervals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for measuring autonomic computing capabilities, such as disturbance injection, fail to provide accurate and repeatable measurements for partially autonomic systems and lack control over detection and recovery timing in fully autonomic systems.
Innovation Solution
Implementing a separate adjustable detection interval and recovery initiation interval during fault injection to simulate the time required for problem detection and recovery procedure initiation, respectively, allowing for quantitative measurement of autonomic capabilities across different automation levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fixed detection interval is used in traditional fault injection methods, then the measurement process is simple, but the measurement accuracy and repeatability are poor for partially autonomic systems
Solution Approach 1:
The detection interval is segmented into two separate adjustable parameters: detection_interval (time from disturbance injection to problem detection) and recovery_initiation_interval (time from problem detection to recovery procedure initiation). This segmentation allows independent tuning of each phase to match the specific characteristics of different autonomic system types, thereby improving measurement accuracy without excessive complexity
Solution Approach 2:
The measurement methodology transitions from a static fixed detection interval to dynamic adjustable intervals that can be configured based on the autonomic system type being tested. The system allows runtime configuration of detection_interval and recovery_initiation_interval to match real-world detection and response times, improving measurement repeatability
2Adaptability or versatility
If a single fixed detection interval is used, then the testing methodology is simple, but control over detection and recovery timing is lost in fully autonomic systems
Solution Approach 1:
The enhanced fault injection methodology is designed to universally test all three types of autonomic systems (non-autonomic, partially autonomic, and fully autonomic) using the same framework. By providing adjustable detection_interval and recovery_initiation_interval parameters, the methodology can adapt to different automation levels: setting recovery_initiation_interval to 0 for fully autonomic systems, using fixed values for non-autonomic systems, and allowing variable configuration for partially autonomic systems
3Measurement precision
If the detection interval is fixed based on MTTR, then the methodology is straightforward, but it cannot accurately measure partially autonomic systems that provide alerts and messages to operators
Solution Approach 1:
The detection process is segmented into automatic detection time (detection_interval) and human operator response time (recovery_initiation_interval). This allows the measurement to capture the full timeline of partially autonomic systems where alerts are automatically generated but recovery requires manual intervention, improving measurement precision and repeatability
Solution Approach 2:
The methodology changes from using a single fixed parameter (detection interval based on MTTR) to using two separate adjustable parameters (detection_interval and recovery_initiation_interval). This parameter change enables accurate representation of partially autonomic system behavior where automatic detection occurs quickly but manual recovery takes additional time
Data Source
AI summary
The present invention is directed to the quantitative measurement of the autonomic capabilities of computing systems. A method in accordance with an embodiment of the present invention includes: subjecting the computing system to a workload; injecting a disturbance into the computing system; providing a notification that the computing system has detected a problem in response to the injected disturbance; determining an amount of time required to initiate a recovery procedure to address the detected problem; and determining an amount of time required to execute the recovery procedure.