Workload Policy Repository for Dynamic Container Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing workload management systems are complex and error-prone, requiring manual configuration file modifications and daemon restarts, which complicates defining and updating workload management policies across multiple computing containers.
Innovation Solution
A configuration repository with policy objects that allows workload management policies to be defined independently of computing containers, enabling dynamic resource allocation and automatic updates without administrator intervention, using a central management station with a GUI and CLI for policy management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If workload management policies are configured through manual configuration file modifications and daemon restarts, then policy updates can be implemented, but the system complexity and error-proneness increase significantly
Solution Approach 1:
The patent introduces a configuration repository as an intermediary layer between administrators and workload management daemons. This repository stores policy definitions in a structured format and automatically distributes them to relevant daemons, eliminating the need for manual configuration file edits and daemon restarts. The intermediary handles policy validation, version control, and automatic propagation, thereby reducing errors and simplifying the update process.
Solution Approach 2:
The workload management system implements self-service capabilities where the configuration repository automatically detects policy changes, validates them against predefined rules, and propagates updates to the appropriate daemons without requiring administrator intervention for each change. The system self-manages the complexity of policy distribution and coordination, reducing the burden on administrators while maintaining reliability.
2Ease of operation
If workload management daemons are restarted to apply policy updates, then new policies take effect, but service interruption and loss of time occur
Solution Approach 1:
The patent implements dynamic policy updates where the configuration repository can modify workload management policies in real-time without requiring daemon restarts. The system dynamically loads new policy definitions from the repository and applies them to running daemons, allowing policies to be updated while services continue to operate. This dynamic approach eliminates service interruptions and reduces the time loss associated with restart operations.
Solution Approach 2:
The configuration repository performs preliminary validation and preparation of policy updates before they are applied to daemons. Policies are validated against syntax rules and consistency constraints in advance, ensuring that updates can be applied smoothly without requiring restarts to resolve configuration errors. This preliminary action prevents failed updates that would otherwise require service interruption.
3Productivity
If multiple computing containers share common server platforms to improve utilization rates, then resource efficiency increases, but coordination complexity and administration costs increase
Solution Approach 1:
The patent creates a universal configuration repository that serves multiple computing containers and workload management daemons simultaneously. This single repository provides a centralized location for storing and managing policies across the entire platform, eliminating the need for separate configuration management for each container. The universal repository handles policy distribution, validation, and version control for all containers, reducing coordination complexity while enabling efficient resource sharing and improving overall utilization rates.
Data Source
AI summary
In one embodiment, a system comprises a plurality of computing containers having processing resources for executing software workloads, a plurality of management processes for controlling access to the processing resources according to workload policy definitions, a data container for storing a plurality of workload policy definitions and associations between the plurality of workload policy definitions and the plurality of computing containers, and a configuration process for communicating workload policy definitions from the data container to the plurality of management processes.