Switching Device Dynamic Queue Allocation for IaaS Congestion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In Infrastructure as a Service (IaaS) environments, existing technologies face challenges in maintaining high communication performance and fairness in network resource allocation, leading to congestion and decreased communication quality due to limited queue allocation and fixed scheduling, which can result in unfair resource sharing and increased costs.
Innovation Solution
A switching device with a plurality of queues that acquires and stores packets, groups flows based on transmission/reception information, and schedules output based on priority to ensure fair allocation and improved communication performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the number of queues is increased to allocate dedicated queues to all flows, then communication performance is improved, but device complexity and cost increase
Solution Approach 1:
The patent implements dynamic queue allocation where the number and configuration of queues are not fixed but adapt based on current network conditions, flow characteristics, and congestion states. The control unit dynamically creates, modifies, or merges queues to match actual traffic patterns, allowing the system to maintain high communication performance without requiring a large fixed number of queues.
Solution Approach 2:
The patent makes queues multi-functional by allowing a single queue to serve multiple flows through the flow group mechanism. Instead of dedicating one queue per flow, multiple flows are grouped together and share queues dynamically, enabling the queueing system to handle diverse traffic patterns with fewer queues while maintaining performance.
2Device complexity
If fixed scheduling is used to simplify queue management, then device complexity is reduced, but communication performance and fairness deteriorate due to congestion
Solution Approach 1:
The patent implements feedback mechanisms where the control unit continuously monitors network conditions, queue states, and flow characteristics. Based on this feedback, the system dynamically adjusts queue allocations, scheduling policies, and flow groupings to prevent congestion and maintain fairness. The feedback loop enables adaptive response to changing traffic patterns without requiring complex manual configuration.
Solution Approach 2:
The system performs self-service by automatically detecting congestion conditions and adjusting its own queue management without external intervention. The control unit autonomously modifies queue configurations, reallocates resources, and adjusts scheduling parameters based on observed network states, reducing the need for complex external control mechanisms while maintaining performance.
3Reliability
If band control is performed for each flow to ensure fairness, then communication performance is improved, but device complexity increases
Solution Approach 1:
The patent segments the fairness control problem by introducing flow groups as an intermediate layer between individual flows and queues. Instead of controlling each flow separately, the system groups related flows together and applies band control at the group level. This segmentation reduces the complexity of per-flow control while maintaining fairness through group-level resource allocation and scheduling.
4Reliability
If VMs are disposed with sufficient band to avoid congestion, then communication performance is maintained, but resource use efficiency decreases and cost increases
Solution Approach 1:
The patent implements dynamic resource allocation where bandwidth and queue resources are allocated based on actual traffic demands rather than static pre-allocations. The control unit continuously monitors usage patterns and adjusts resource distribution in real-time, allowing the system to provide sufficient bandwidth during high-demand periods while consolidating resources during low-demand periods, thereby maintaining performance while improving efficiency.
Solution Approach 2:
The patent merges multiple flows into flow groups that share common queue resources. By combining traffic from multiple virtual machines into shared queues with dynamic scheduling, the system achieves better resource utilization compared to dedicated queues. This merging allows underutilized resources from one flow to be allocated to other flows in the same group, improving overall efficiency while maintaining quality of service.
Data Source
AI summary
A plurality of internal queues that receives and stores therein packets and that outputs each of the stored packets is included. Then, a transmission side meter counter and a reception side meter counter acquire transmission/reception information on the packets that are stored in each of the internal queues. A control unit creates, on the basis of the transmission/reception information, flow groups to each of which one or a plurality of flows belongs and allocates the internal queues to the flow groups. A scheduler determines, on the basis of the transmission/reception information and previously determined priority of the internal queues, the order in which each of the internal queues is output.