Service-Type Network Pipelines for Bandwidth-Weighted Traffic Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network flow control methods struggle to balance the performance requirements of diverse service types in a network environment, leading to mutual interference and inefficient bandwidth utilization, particularly in large-scale networks with varying service demands.
Innovation Solution
A network flow control method that allocates packet flows to different service pipelines based on service types, assigning bandwidth weights to each pipeline, and performs scheduling based on these weights to avoid interference and meet specific performance requirements of each service type.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If high bandwidth and differentiated service manner are used, then network simplicity and ease of operation are improved, but bandwidth utilization and network performance deteriorate
Solution Approach 1:
The patent segments the network service pipeline into multiple independent service pipelines, each dedicated to a specific service type. This segmentation allows different service types to be transmitted through separate pipelines, avoiding mutual interference while maintaining simple differentiated service handling at each pipeline level. The segmentation resolves the contradiction by enabling both operational simplicity (through clear service-type-based routing) and improved bandwidth utilization (through dedicated pipelines that prevent congestion spillover).
Solution Approach 2:
The patent introduces a new dimension of service type classification alongside traditional routing dimensions. By adding service type as an additional dimension for pipeline selection, the system can differentiate services without complex multi-parameter algorithms. This dimensional addition enables simple yet effective service differentiation while improving overall network utilization through specialized pipelines optimized for different service characteristics.
2Productivity
If centralized flow control with dynamic flow matrix is implemented, then network performance and utilization are improved, but real-time decision making and expansibility deteriorate
Solution Approach 1:
The patent segments the centralized control function into distributed service pipeline controllers, each managing a specific service type independently. This segmentation eliminates the need for a single complex centralized controller that must make real-time decisions for all traffic. Each pipeline controller operates semi-independently, reducing real-time decision complexity while maintaining high network utilization through coordinated multi-pipeline operation.
Solution Approach 2:
The patent introduces service pipeline selection mechanisms as intermediaries between traffic sources and the core network infrastructure. These intermediaries classify traffic by service type and route to appropriate pipelines, absorbing the complexity of flow control decisions at the edge rather than requiring complex real-time processing in the core network. This intermediary layer enables simple high-speed forwarding in the core while maintaining optimized utilization through service-type-based routing.
3Adaptability or versatility
If multiple service types share the same network pipeline, then network resource sharing is improved, but mutual interference between service flows increases
Solution Approach 1:
The patent physically segments the network infrastructure into multiple service pipelines, each dedicated to a specific service type. This segmentation eliminates mutual interference between different service flows by providing dedicated transmission paths. At the same time, resource sharing is maintained at the infrastructure level through the coordinated operation of multiple pipelines, allowing the network to adapt to different service demands while preventing harmful interference through physical separation.
Solution Approach 2:
The patent applies local quality by optimizing each service pipeline for its specific service type characteristics. Each pipeline can be tailored with appropriate buffering, scheduling, and transmission parameters suited to its designated service type (e.g., low-latency optimization for real-time services, high-throughput optimization for bulk data services). This local optimization eliminates interference between service types while maintaining versatile resource utilization across the multi-pipeline system.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
Embodiments of this application provide a network flow control method and a network device. The method includes: receiving a packet flow; determining, based on a service type of the packet flow, a service pipeline used for transmitting the packet flow, where service types of all packet flows in the service pipeline are the same; and based on a bandwidth weight allocated to the service type, transferring the packet flow in the service pipeline to a physical port. In the embodiments of this application, packet flows are allocated to different service pipelines based on a service type, and bandwidth weights are allocated, in a centralized manner, to service pipelines that carry a same service type. Therefore, each service pipeline can schedule packet flows in the service pipeline based on a bandwidth weight allocated to a specific service type, so as to avoid mutual interference generated between packet flows of different service types, and meet performance requirements of service flows of different service types.