Edge Connection Resource Loaning for Balanced Queue Depths
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-site cloud systems face inefficiencies in resource utilization and performance degradation due to unequal queue depths and idle connections, leading to resource wastage and potential QFULL conditions.
Innovation Solution
A connection management system that loans resources between hosts in a multi-site target system, optimizing queue depths and sharing idle connections to improve resource utilization and performance by generating a superior equilibrium.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If connections are maintained in multi-site target system, then host connectivity and data access are ensured, but resource utilization deteriorates due to idle connections and unequal queue depths
Solution Approach 1:
The system identifies idle connections that are not currently needed by hosts and transfers their associated resources (command buffers, queue depth) to other hosts that do have need. This recovers previously wasted resources and makes them available for active connections, thereby improving overall resource utilization while maintaining connectivity through the connection management system.
Solution Approach 2:
The connection management system continuously monitors queue depths and connection states across multiple sites, using this feedback to dynamically reallocate resources. When idle connections are detected, the system uses the feedback mechanism to transfer resources to hosts with deeper queue needs, creating a self-regulating system that optimizes resource distribution in real-time.
2Speed
If queue depth is increased for all hosts, then data access performance is improved, but network congestion and QFULL conditions increase due to resource wastage
Solution Approach 1:
Instead of uniformly increasing queue depth for all hosts, the system applies local quality by selectively allocating queue depth resources only to hosts that actually need them. Each host receives the appropriate queue depth based on its current workload and connection state, preventing both under-provisioning (which limits performance) and over-provisioning (which causes congestion).
Solution Approach 2:
The system dynamically changes the queue depth parameter on a per-host basis rather than maintaining fixed allocations. When hosts become idle, their queue depth is reduced and resources are freed; when hosts become active, queue depth is increased. This dynamic parameter adjustment allows the system to optimize performance without causing sustained network congestion.
3Adaptability or versatility
If connection resources are allocated to all hosts, then connectivity availability is ensured, but resource wastage occurs due to idle connections
Solution Approach 1:
The system transitions from static connection resource allocation to dynamic allocation. Connections and their associated resources (command buffers, queue depth, memory) can be transferred between hosts based on real-time needs. This dynamic approach ensures that connectivity resources are available when needed while eliminating wastage during idle periods, as resources are reassigned to active connections.
Data Source
AI summary
Provided are a method, system, and computer program product in which a connection management system is provided in multi-site target system. Resources for connections not needed at a point in time are loaned, from one entity to another in the connection management system.