Threshold-Based Access Control for Compute Resource Scheduling

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In compute environments, existing grid and cluster resource management systems face inefficiencies due to multiple layers of schedulers, heterogeneous resources, and differing performance goals among users, leading to challenges in allocating resources and achieving optimal workload distribution.

Innovation Solution

Implementing a threshold-based access system that associates a required service level with compute resources and requestors, allowing access only if the requestor's service level meets the specified threshold, enabling flexible resource management and load balancing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple layers of schedulers are used to manage compute resources, then resource allocation flexibility is improved, but system complexity increases

Engineering Contradiction:

Improveresource allocation flexibilityVSAvoidsystem complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system segments the scheduling function into multiple independent scheduler layers (grid scheduler, cluster schedulers, resource managers) that operate hierarchically. Each layer manages specific aspects of resource allocation independently, allowing flexibility in resource distribution while maintaining manageable complexity through clear separation of concerns.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces intermediary components such as the access control list and service level threshold mechanism that mediate between resource requests and resource allocation. These intermediaries translate high-level service level requirements into concrete resource allocation decisions, reducing the complexity burden on individual scheduler layers.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If local schedulers are given direct access to resources, then scheduling efficiency is improved, but centralized control is reduced

Engineering Contradiction:

Improvescheduling efficiencyVSAvoidcentralized control

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The scheduling authority is segmented between centralized grid schedulers and decentralized local schedulers. Local schedulers (resource managers) handle day-to-day resource allocation efficiently, while grid schedulers maintain centralized policy control through access control lists. This segmentation allows both efficiency and control to coexist.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary action by establishing access control lists and service level thresholds before resource allocation occurs. These pre-defined policies enable local schedulers to make efficient decisions autonomously within bounded constraints, eliminating the need for continuous centralized micromanagement while preserving overall control.

Inventive Principle:

Principle #10Preliminary action

3Productivity

If resources are shared among multiple users and projects, then resource utilization is improved, but performance predictability deteriorates

Engineering Contradiction:

Improveresource utilizationVSAvoidperformance predictability

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system changes the parameter of resource allocation from static to dynamic by introducing service level thresholds that can be adjusted based on current system state and user requirements. This allows the system to maintain high resource utilization while adapting performance guarantees to match actual workload conditions, improving predictability without sacrificing utilization.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent implements feedback mechanisms where the system continuously monitors resource usage, queue lengths, and service level compliance. This feedback information is used to dynamically adjust scheduling decisions and access control policies, ensuring that performance predictability is maintained even as resource utilization fluctuates due to shared access.

Inventive Principle:

Principle #23Feedback

4Reliability

If service level thresholds are enforced for resource access, then service quality is improved, but access flexibility is reduced

Engineering Contradiction:

Improveservice qualityVSAvoidaccess flexibility

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The service level thresholds are implemented as dynamic rather than static constraints. The system can adjust thresholds in real-time based on workload conditions, user priorities, and resource availability. This dynamics allows the system to enforce service quality when needed while maintaining access flexibility when conditions permit, resolving the contradiction between quality and flexibility.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS8370898B1System and method for providing threshold-based access to compute resources

Publication Date: 2013.02.05 III HOLDINGS 12 LLC

AI summary

The invention relates to systems, methods and computer-readable media for controlling access to compute resources in a compute environment such as a cluster or a grid. The method of providing conditional access to a compute environment comprises associating a required service level threshold with a compute environment, associating a service level with a requestor, receiving a request for access to the compute environment from the requestor; and, if the service level of the requestor meets the specified service level threshold, then allowing access to the compute resources. The threshold-based access may be enforced by reservations, policies or some other method.