Workload Manager Using Request Classification Identifiers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In complex enterprise-class computer server environments, real-time system health monitoring and resource management are challenging due to the fast-changing nature of these systems, making traditional methods inadequate for preventing unintended system interruptions and ensuring optimal application performance.
Innovation Solution
A system and method for collecting and surfacing request metrics via Request Classification Identifiers (RCIDs) to enable Quality-of-Service (QoS) and workload management, allowing for classification of requests into performance classes and prioritization based on business objectives, with metrics collected and aggregated across multiple dimensions to optimize resource allocation and performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional system monitoring methods are used, then system health can be monitored, but real-time management and prevention of system interruptions becomes difficult in fast-changing environments
Solution Approach 1:
The patent implements dynamic workload management by continuously monitoring system metrics and automatically adjusting resource allocation in real-time. The workload manager dynamically shifts resources between performance classes based on current system state and business objectives, enabling the system to adapt to fast-changing conditions without manual intervention and maintaining both reliability and real-time responsiveness.
Solution Approach 2:
The patent establishes a feedback loop where system metrics are continuously collected, analyzed, and used to adjust workload distribution. The workload manager receives feedback on system performance and automatically modifies resource allocation to meet business objectives, creating a closed-loop system that continuously improves its own performance in real-time.
2Ease of operation
If resources are allocated equally to all requests, then system simplicity is maintained, but performance objectives for critical requests cannot be prioritized
Solution Approach 1:
The patent segments requests into different performance classes based on business objectives and criticality. The workload manager then applies different resource allocation strategies to each segment, allowing critical requests to receive prioritized resources while non-critical requests receive standard allocation. This segmentation enables differentiated service levels without overwhelming complexity.
Solution Approach 2:
The patent applies local quality by treating different request types differently based on their importance to business objectives. Each performance class receives customized resource allocation and management strategies tailored to its specific requirements, allowing optimal performance for critical requests while maintaining appropriate service for less critical requests.
3Reliability
If detailed request classification and metrics collection is implemented, then quality-of-service management improves, but system complexity increases
Solution Approach 1:
The patent implements a universal workload manager that handles multiple functions: request classification, metrics collection, resource allocation, and performance monitoring. By consolidating these functions into a single integrated system, the patent reduces overall complexity compared to having separate systems for each function, while still providing comprehensive quality-of-service management.
Solution Approach 2:
The workload manager acts as an intermediary layer between requesters and system resources. It sits in the middle of the system, collecting metrics, classifying requests, and making intelligent resource allocation decisions. This intermediary approach simplifies the overall system architecture by centralizing management logic rather than requiring complex distributed intelligence throughout the entire system.
Data Source
AI summary
Described herein are systems and methods for collecting and surfacing metrics with respect to their classification; and the use of the metrics by a workload manager and other application monitoring tools to provide quality-of-service and workload management. Each request is classified, either by the application server or another process. A request classification identifier (RCID) is associated with each request, and thereafter flows with that request as it is being processed. The RCID value is used by data collectors at various points in the system to aggregate the metrics, and a workload manager collects the metrics. The collected metrics are then processed by a rules engine at the workload manager, which analyzes the metrics and generates adjustment recommendations to provide quality-of-service and workload management.