Mainframe Logical Partition Capacity Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mainframe computer systems face challenges in optimizing processing capacity usage and cost control, particularly in managing workload performance and capacity limitations across logical partitions, leading to inefficient resource allocation and increased software costs due to lack of discrimination between workload importance classes and time-criticality levels.
Innovation Solution
A method for managing mainframe computer system usage involves establishing group definitions and policies with time criticality levels for workload tasks, monitoring performance, and automatically balancing or adjusting processing capacity limits between logical partitions to ensure optimal resource allocation and cost control, by reallocating capacity from non-time critical to time-critical tasks within defined thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual capacity limits (DC and GCL) are set for logical partitions, then processing capacity control is achieved, but resource allocation efficiency deteriorates due to lack of automatic adjustment
Solution Approach 1:
The system enables logical partitions to automatically adjust their own capacity limits based on monitored performance metrics and workload characteristics. The WLM system self-regulates capacity allocation without manual intervention, allowing each partition to serve itself according to actual needs rather than static manual configurations.
Solution Approach 2:
The system continuously monitors workload performance, capacity utilization, and service class metrics, then uses this feedback to dynamically adjust capacity limits. The feedback loop compares actual performance against targets and automatically modifies capacity allocations to optimize resource utilization while maintaining service levels.
2Ease of operation
If capacity limits are uniformly applied across all workload classes, then simplicity is maintained, but discrimination between important and non-important workloads deteriorates
Solution Approach 1:
The system applies different capacity management policies to different workload classes, service classes, and logical partitions based on their specific requirements. Each workload class can have customized capacity limits, performance targets, and adjustment parameters, allowing precise discrimination between critical and non-critical workloads while maintaining localized control.
Solution Approach 2:
The system segments workloads into multiple service classes and importance levels, applying distinct capacity management rules to each segment. This segmentation enables differentiated treatment of workloads based on business priority, performance requirements, and resource needs, allowing simultaneous management of multiple workload types with different characteristics.
3Stability of the object's composition
If static capacity limits are configured, then system stability is maintained, but adaptability to changing workload demands deteriorates
Solution Approach 1:
The system transitions from static capacity limits to dynamic, continuously adjustable capacity allocations. Capacity limits are automatically modified in response to changing workload patterns, performance metrics, and system conditions, enabling the system to adapt to varying demands while maintaining operational stability through controlled adjustment mechanisms.
Solution Approach 2:
The system dynamically changes capacity parameters based on monitored performance indicators and workload characteristics. By adjusting capacity limits as a variable parameter rather than a fixed value, the system can respond to changing conditions while maintaining stability through gradual, controlled parameter modifications that prevent abrupt transitions.
4Device complexity
If capacity is allocated without considering time criticality, then allocation simplicity is maintained, but workload performance for critical tasks deteriorates
Solution Approach 1:
The system pre-configures service classes and performance targets for different workload types, establishing priority frameworks before actual workload execution. By preliminarily defining which workloads are time-critical and their associated performance requirements, the system prepares capacity allocation rules in advance, enabling rapid response to critical workloads without complex real-time decision-making.
Solution Approach 2:
The system applies specialized capacity management policies specifically to time-critical workloads, distinguishing them from non-critical workloads through service class definitions. Critical workloads receive preferential treatment with guaranteed capacity allocations and performance targets, while non-critical workloads operate under different rules, allowing optimized performance for critical tasks without uniformly increasing complexity across all workloads.
Data Source
AI summary
In mainframe computer system, workload tasks are accomplished using a logically partitioned data processing system, where the partitioned data processing system is divided into multiple logical partitions. In a system and method managing such a computer system, each running workload tasks that can be classified based on time criticality, and groups of logical partitions can be freely defined. Processing capacity limits for the logical partitions in a group of logical partitions based upon defined processing capacity thresholds and upon an iterative determination of how much capacity is needed for time critical workload tasks. Workload can be balanced between logical partitions within a group, to prevent surplus processing capacity being used to run not time critical workload on one logical partition when another logical partition running only time critical workload tasks faces processing deficit.