Dynamic Control Block Allocation for Distributed Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In distributed computer systems, the limited availability of control blocks can lead to degradation of high-priority work, starvation, and priority inversion due to task overloads, as each control block is exclusively associated with a single task, and the pool of available control blocks is constrained, limiting expansion and causing quotas to slow down tasks based on flow limits.
Innovation Solution
Implementing a system that applies limits to control block occupancy rates by determining an allowed occupancy rate for each service class, acquiring and releasing control blocks to ensure compliance with the actual tenancy rate, and maintaining a credit balance to manage the allocation and usage of control blocks fairly across multiple client tasks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a large fixed number of control blocks are allocated to support multiple client tasks, then system capacity and productivity are improved, but system complexity and resource management difficulty increase
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the number of control blocks allocated to each service class based on occupancy rates. The system monitors the actual number of tasks being serviced and adjusts control block allocation accordingly, transitioning from static to dynamic parameter management to resolve the contradiction between system capacity and management complexity
Solution Approach 2:
The patent implements feedback mechanisms by continuously monitoring control block occupancy rates and using this information to regulate task admission and control block allocation. The system measures actual occupancy against target occupancy rates and adjusts resource allocation in response, creating a closed-loop control system that manages complexity while maintaining high system capacity
2Reliability
If quotas are applied to control block classes to manage occupancy, then resource allocation fairness is improved, but task processing speed deteriorates
Solution Approach 1:
The patent applies dynamics by making control block quotas flexible rather than fixed. The system dynamically adjusts the number of control blocks available to each service class based on current occupancy rates and system conditions. This allows the system to maintain fairness while adapting to changing workloads, preventing permanent slowdowns that would result from rigid quotas
Solution Approach 2:
The patent implements preliminary action by pre-calculating target occupancy rates and control block allocation parameters before task admission. The system establishes admission control rules and occupancy targets in advance, allowing tasks to be processed at full speed when conditions permit while ensuring fairness constraints are met through pre-planned resource allocation
3Reliability
If control blocks are exclusively associated with single tasks, then task processing reliability is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent applies universality by making control blocks reusable across multiple tasks through a pool-based allocation system. Control blocks are released back to the pool upon task completion and can be reallocated to subsequent tasks. This multi-functional use of control blocks maintains reliability through proper task-control block association while dramatically improving resource utilization efficiency through shared access
Data Source
AI summary
Embodiments of the invention are provided to enable fair and balanced allocation of control blocks to support processing of requests received from a client machine. The server is configured with tools to manage an account balance of control block availability for each service class. The account balance is periodically adjusted based upon usage, tenancy, deficits, and passage of time. Processing of one or more tasks in a service class is support when the credit value in the service class account is equal to or greater than the entry cost estimated for the request.