Zombie Server Detection via Utility Productivity Utilization Gap
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for identifying zombie servers, which are no longer used but still maintained, face challenges due to low processor utilization similarities with active servers, making it difficult to distinguish between utilized and unused servers.
Innovation Solution
The approach involves labeling processes as utility or productivity software, calculating their utilization, and comparing it to server utilization, identifying a server as a zombie if the difference is below a threshold, using a system that samples process IDs, CPU run times, and network traffic analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If processor utilization tracking is used to identify zombie servers, then identification simplicity is improved, but identification accuracy deteriorates because management software contributes to processor utilization without customer utilization
Solution Approach 1:
The patent segments processor utilization into two distinct categories: management software utilization and productivity application utilization. By separating these components, the system can accurately identify zombie servers by detecting when management software utilization exists without corresponding productivity application utilization, thus resolving the accuracy issue while maintaining operational simplicity.
Solution Approach 2:
The patent introduces an intermediary analysis layer that examines the relationship between management software processes and productivity application processes. This intermediary mechanism determines whether management software utilization is accompanied by actual customer utilization, thereby improving identification accuracy without complicating the overall identification process.
2Measurement precision
If low processor utilization threshold is used to identify zombie servers, then detection sensitivity is improved, but false positive rate increases because non-zombie servers used for productivity applications also have low utilization
Solution Approach 1:
The patent segments the analysis into comparing management software utilization against productivity application utilization separately, rather than using a single aggregate threshold. This segmentation allows the system to maintain high detection sensitivity by identifying servers where management software runs without productivity applications, while avoiding false positives by verifying the absence of productivity utilization rather than relying solely on low overall utilization thresholds.
3Measurement precision
If comprehensive process analysis is performed to distinguish zombie servers, then identification accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent extracts and focuses analysis only on the critical distinction between management software processes and productivity application processes, rather than analyzing all system parameters. By taking out and isolating these specific process types for comparison, the system achieves high identification accuracy while minimizing computational complexity by ignoring irrelevant system attributes.
Solution Approach 2:
The system uses existing process information already available in the server environment, allowing processes to essentially self-report their utilization through standard monitoring mechanisms. This self-service approach enables comprehensive process analysis without requiring additional complex computational resources, as the system leverages existing data infrastructure.
Data Source
AI summary
A zombie server can be detected. Detecting a zombie server can include labeling a plurality of processes as utility software, calculating a utilization of utility software on the plurality of processes executed in one or more processing resources during an interval of time, and calculating a server utilization of the one or more processing resources during the interval of time. Detecting the zombie server can also include determining whether a difference between the utilization of utility software and the server utilization is greater than a threshold, and identifying a server that hosts the processing resource as a zombie server based on a determination that the difference is smaller than the threshold.