Dynamic Policy-Based Resource Allocation for Storage Arrays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional allocation methods for system resources, such as storage arrays, are either inefficient due to lack of consideration for resource characteristics or inflexible due to hard-coded ad hoc algorithms, failing to provide dynamic and policy-based management.
Innovation Solution
A method and system that maintain statistics for system resources to select the best resource based on class and policy, allowing for dynamic allocation while avoiding rigidity, using class-based and total resource allocation statistics, and allowing for adjustments and client preferences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If naive allocation algorithms are used, then device complexity is reduced, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent changes the parameters of the allocation algorithm from static to dynamic by introducing policies that can be modified without changing the underlying algorithm structure. The system maintains statistics about resource characteristics (bus capacity, RAID group status, LU capacity) and uses these parameters to dynamically select allocation strategies, thereby improving resource utilization without increasing algorithmic complexity.
Solution Approach 2:
The patent introduces dynamic policy-based allocation that adapts to changing system conditions. Instead of using fixed naive algorithms, the system continuously updates resource statistics and adjusts allocation decisions based on current system state, making the allocation process dynamic rather than static.
2Productivity
If ad hoc allocation algorithms are used, then resource utilization efficiency is improved, but adaptability deteriorates
Solution Approach 1:
The patent segments the allocation system into distinct components: statistics collection, policy definition, and allocation decision-making. This segmentation allows the policy layer to be independently modified without affecting the underlying allocation algorithm, providing adaptability while maintaining efficient resource utilization.
Solution Approach 2:
The patent creates a universal policy-based framework that can handle multiple allocation scenarios and resource types through a single adaptable system. The policy mechanism serves multiple functions including resource selection, constraint enforcement, and optimization, replacing the need for multiple specialized ad hoc algorithms.
3Productivity
If statistics-based policy allocation is implemented, then resource utilization efficiency is improved, but device complexity increases
Solution Approach 1:
The patent introduces an intermediary policy layer that sits between the allocation request and the resource selection process. This intermediary collects and processes statistics about resource characteristics, then applies policies to make allocation decisions, thereby improving efficiency without directly complicating the core allocation mechanism.
Solution Approach 2:
The system implements self-service by automatically collecting resource statistics and applying allocation policies without requiring manual intervention or complex external control systems. The allocation mechanism serves itself by using its own collected data to make intelligent decisions.
Data Source
AI summary
Systems, methods, and computer readable medium for optimizing storage allocations based on system resources are disclosed. According to one aspect, the subject matter described herein includes a method for dynamic, policy-based allocation of system resources. The method includes maintaining statistics in connection with like system resources. Each of the system resources having associated statistics that are suitable for describing a relationship with an allocatable class of resource. A request is received for a resource. In response to receiving the request, one of the like system resources is selected, based on the statistics and a policy, as a best system resource in connection with the allocatable class of resource.