Workload Manager Policy Execution for Host Event Response
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional workload managers in distributed computer systems require manual intervention by human operators to change policies in response to host-level events, which can lead to suboptimal resource allocation and increased operational complexity.
Innovation Solution
A system that automatically detects host-level events and adjusts policies by reallocating processing resources, enhancing processing capabilities, and notifying operators, using a workload management program that monitors the system and invokes actions based on predefined event responses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual intervention by human operators is used to change policies in response to host-level events, then system control and policy changes can be made with human judgment, but response time increases and operational complexity increases
Solution Approach 1:
The patent pre-defines multiple policies with associated host-level events and responses before runtime. When a host-level event occurs, the workload manager automatically matches the event to a predefined policy and executes the corresponding action without requiring human intervention. This preliminary configuration enables rapid automated response while maintaining policy quality through careful pre-planning.
2Adaptability or versatility
If manual intervention by human operators is used to change policies, then complex policy decisions can be made with expert knowledge, but operational complexity and the need for skilled personnel increase
Solution Approach 1:
The workload manager is designed to automatically detect host-level events, retrieve corresponding predefined policies, and execute appropriate actions without human intervention. The system serves itself by having the workload manager directly interact with the operating system to implement policy changes, eliminating the need for operators to manually analyze events and select policies, thus reducing operational complexity while maintaining adaptability.
3Productivity
If predefined policies are executed automatically in response to host-level events, then response speed and efficiency improve, but the ability to handle unexpected or complex situations may be reduced
Solution Approach 1:
The patent implements a dynamic policy execution framework where multiple predefined policies with different priorities can be associated with the same host-level event. The workload manager evaluates policies in priority order and can execute multiple policies if appropriate. This dynamic approach allows the system to handle unexpected situations by having pre-planned responses for various event scenarios, while maintaining high response efficiency through automated execution.
Data Source
AI summary
Policies, which specify event-responses for reacting to events, may be configured on host computer systems. An event-response may be executed when a particular event is detected on a host. The events may include a software change, a host machine processes change, a resources key change, and/or a critical failure. One or more event-responses may be defined for a policy. When an event is detected, the active policy is examined for event-responses corresponding to the event. The event-responses may also have defined a permitted number of executions. Thus, an event-response may be automatically executed only a permitted number of times. After the permitted number is reached, the event-response no longer executes in response to an event. The counters corresponding to event-responses may be reset when a new policy is activated.