Dynamic Multi-Fabric Load Balancing for Network Resource Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current multi-fabric network environments in cloud computing fail to effectively prioritize critical applications with high-bandwidth needs, leading to inefficient bandwidth utilization and increased costs due to the need for high-port-count switches, which are costly and power-intensive.
Innovation Solution
Implementing a dynamic multi-fabric load balancing system that reassigned network fabrics based on the detection of critical applications, allowing critical applications to exclusively use one fabric while non-critical applications share the other, thereby optimizing bandwidth allocation and reducing the need for high-port-count switches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If high-port-count switches are deployed to support all applications on all network fabrics, then network availability and bandwidth access are improved, but device cost and power consumption increase
Solution Approach 1:
The patent implements dynamic fabric assignment where network fabrics are reassigned based on application criticality and bandwidth needs. The system continuously monitors application performance and dynamically adjusts fabric allocation, transitioning from static to dynamic resource management. This allows the network to adapt to changing conditions without requiring permanent high-capacity connections for all applications.
Solution Approach 2:
The patent applies different quality levels of network fabric assignment to different applications. Critical applications receive exclusive or prioritized fabric assignment with guaranteed bandwidth, while non-critical applications share fabrics dynamically. This creates local quality differentiation where each application receives appropriate network resources based on its specific needs rather than uniform allocation.
2Productivity
If high-port-count switches are deployed to support all applications on all network fabrics, then network bandwidth access is improved, but device cost increases
Solution Approach 1:
The system dynamically assigns network fabrics to applications based on detected bandwidth needs and criticality levels. When critical applications require high bandwidth, they are assigned exclusive fabric access. When bandwidth demands are lower, fabrics are shared among multiple applications. This dynamic allocation maximizes bandwidth utilization efficiency without requiring expensive high-port-count switches to be permanently configured for maximum capacity.
Solution Approach 2:
The patent makes network fabrics universal resources that can be assigned to different applications based on needs. Rather than dedicating specific fabrics to specific applications or using high-port-count switches for all applications, the same fabrics serve multiple applications at different times and with different priority levels, increasing overall resource utilization efficiency.
3Ease of operation
If network fabrics are statically assigned to applications, then configuration simplicity is improved, but bandwidth utilization efficiency deteriorates
Solution Approach 1:
The system implements self-service through automated fabric assignment and reassignment based on application performance monitoring. The load balancing manager automatically detects critical applications, monitors bandwidth usage patterns, and reassigns fabrics without manual intervention. This maintains operational simplicity while achieving high bandwidth utilization efficiency through automated adaptive resource management.
Solution Approach 2:
The patent incorporates continuous feedback loops where the system monitors application bandwidth usage, detects critical applications, and uses this information to dynamically adjust fabric assignments. The load balancing manager receives feedback about network conditions and application performance, then makes real-time adjustments to optimize bandwidth utilization while maintaining simple operation through automated decision-making.
Data Source
AI summary
An example apparatus to manage network resources includes a load balancing detector to determine to reassign first and second network fabrics; and a network fabric configurator to, in response to the detecting to reassign the first and second network fabrics, configuring a virtual network distributed switch to: assign the first network fabric to ones of the first applications previously assigned to the second network fabric; and assign the second network fabric to the second application.