Event Enrichment via Data Correlation in Datacenters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large-scale datacenters with thousands of servers face challenges in monitoring and dynamically reconfiguring systems to meet changing demands, leading to overwhelming numbers of event messages that are not sufficiently specific for problem diagnosis, and require complex service models for event prioritization and correlation.
Innovation Solution
A method for enriching events using data correlation, which involves storing event messages and system metric data, determining the degree of correlation between system and service metrics, and modifying event messages based on this correlation to enhance their specificity and priority, thereby reducing the need for complex service models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If event correlation systems use computed correlation coefficients to set or modify event attributes, then event prioritization is improved, but the number of event messages presented to operators remains overwhelming and not sufficiently specific
Solution Approach 1:
The patent introduces service models as an intermediary layer between system metrics and event messages. These service models contain pre-defined relationships between system components and service-level objectives, acting as a mediator that translates raw metric correlations into meaningful service-level event attributes, thereby improving both prioritization accuracy and event specificity simultaneously
2Measurement precision
If service models are created and maintained to mirror datacenter components and relationships, then event correlation accuracy is improved, but system complexity and overhead increase
Solution Approach 1:
The patent implements self-service through automatic service model generation that uses observed correlations between system metrics and service-level objectives to dynamically create and update service models without manual intervention. The system automatically discovers relationships, generates appropriate service models, and maintains them through continuous monitoring, eliminating the need for manual service model creation and reduction of maintenance overhead
3Productivity
If pre-defined dependencies between system components are used to identify events for correlation, then processing efficiency is improved, but adaptability to changing service relationships deteriorates
Solution Approach 1:
The patent implements dynamic service models that are continuously updated based on observed correlations between system metrics and service-level objectives. Rather than relying on static pre-defined dependencies, the system dynamically adjusts service models to reflect changing service relationships and component interactions, maintaining both processing efficiency through structured models and adaptability through continuous updates
Data Source
AI summary
Systems and methods for enriching events using data correlation are described herein. At least some embodiments include a method for enriching events reflecting the state of a plurality of computer systems, the method including storing a plurality of event messages and system metric data that includes service metric data, determining a degree of correlation between a system metric and at least one of a plurality of service metrics, and enriching an event message of the plurality of event messages based at least in part on the degree of correlation. At least one system metric data value triggers the event message. The degree of correlation is based at least in part on the system metric data and the service metric data.