Workload Allocation Planning Tool for Service Classes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems lack effective tools for intelligently determining the proper partitioning of a consumer's workload demands across different classes of service (COS) in resource pools, leading to unclear allocation of capacity and potential degradation in quality of service (QoS).
Innovation Solution
A planning tool that evaluates consumer workload demands, QoS desires, and resource access QoS commitments to determine a suitable breakpoint for partitioning between COSs, allowing for guaranteed and non-guaranteed access, while considering permitted degraded performance and time-limited constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If capacity is allocated to satisfy all consumer demands, then QoS is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent segments consumer demands into different classes of service (COS) based on priority levels. High-priority demands are allocated guaranteed capacity while low-priority demands share remaining capacity probabilistically. This segmentation allows the system to satisfy critical QoS requirements for essential services while efficiently utilizing resources for less critical services, resolving the contradiction between QoS satisfaction and resource utilization efficiency.
Solution Approach 2:
The patent changes the allocation parameter from binary (all or nothing) to probabilistic based on consumer priority levels. By introducing probability parameters that vary with service class, the system can optimize the balance between guaranteeing QoS for high-priority services and maximizing overall resource utilization efficiency through statistical multiplexing of lower-priority services.
2Reliability
If guaranteed access is provided to all consumers, then QoS reliability is improved, but system complexity increases
Solution Approach 1:
The patent simplifies the allocation system by segmenting consumers into discrete priority classes rather than handling continuous priority levels. This segmentation reduces the complexity of managing individual consumer requirements while still providing guaranteed access where needed, as the system only needs to track aggregate statistics for each class rather than individual consumer states.
Solution Approach 2:
The patent implements self-service through statistical multiplexing where lower-priority consumers automatically share remaining capacity based on historical demand patterns and probability parameters. This eliminates the need for complex centralized scheduling algorithms to manage every possible allocation scenario, as the system uses pre-computed statistical models to handle capacity sharing automatically.
3Productivity
If resource allocation is optimized for cost-effectiveness, then productivity is improved, but QoS degradation occurs
Solution Approach 1:
The patent segments the resource allocation problem into two independent parts: guaranteed capacity allocation for high-priority services and probabilistic capacity allocation for low-priority services. This segmentation allows the system to optimize cost-effectiveness by fully utilizing statistical multiplexing for low-priority services while maintaining QoS for high-priority services, as the two segments can be optimized independently without compromising either objective.
Solution Approach 2:
The patent introduces probability parameters that change with service class to model the trade-off between cost-effectiveness and QoS. By using these parameters to calculate optimal allocation breakpoints, the system can determine the precise threshold where guaranteed access transitions to probabilistic access, maximizing cost-effectiveness while ensuring QoS for critical services remain satisfied.
Data Source
AI summary
A method comprises receiving into a planning tool a representative workload for a consumer. The method further comprises receiving into the planning tool quality of service desires of the consumer which define permitted degraded performance. In certain embodiments, the permitted degraded performance is time-limited wherein demands of the representative workload may exceed a pre-defined utilization constraint for at least one resource servicing the demands for no more than a pre-defined amount of contiguous time. The planning tool determines an allocation of demand of the consumer for each of a plurality of different classes of service (COSs). In certain embodiments, a first COS provides guaranteed resource access for servicing demand allocated thereto, and a second COS provides non-guaranteed resource access for servicing demand allocated thereto. In certain embodiments, the allocation of demand to the different COSs may be determined for both a normal mode and a failure mode of operation.


