Global Finish-Time Packet Scheduling for Network Latency Guarantees
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network latency guarantee technologies face challenges in managing state information for each flow, leading to complexity and inefficiencies, particularly in core nodes of networks, which affects the ability to maintain service order and ensure upper limits of latency.
Innovation Solution
A network latency guarantee system that derives packet finish times at a first node and uses metadata to schedule packets based on global finish time information, allowing for simple updates in downstream nodes without managing flow state information, using a packet processing unit, scheduling unit, and packet output unit to maintain service order and ensure latency limits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fair queuing scheduler is used to guarantee network latency, then service order and latency upper limit are ensured, but complexity of managing state information for each flow increases significantly
Solution Approach 1:
The patent extracts the finish time calculation from individual node schedulers and performs it centrally at the ingress node. Only the calculated finish time metadata is extracted and attached to packets, while the complex state information management is removed from core nodes. This allows latency guarantee through finish time-based scheduling without the complexity of managing flow states at each node.
Solution Approach 2:
The patent introduces finish time metadata as an intermediary that carries scheduling information from the ingress node to egress nodes. This metadata acts as a mediator that enables coordinated scheduling across the network without requiring core nodes to maintain complex flow state information, thus guaranteeing latency while reducing device complexity.
2Ease of operation
If detailed flow state information is managed at each core node, then precise scheduling control is achieved, but scalability and operational efficiency deteriorate
Solution Approach 1:
The patent extracts flow state management from core nodes and concentrates it at the ingress node. The finish time metadata extracted and attached to packets enables precise scheduling control at egress nodes without requiring core nodes to maintain detailed flow state information, thus improving processing efficiency while maintaining scheduling precision.
3Productivity
If maximum burst size is increased to handle large data transmissions, then throughput is improved, but upper limit of latency becomes difficult to guarantee
Solution Approach 1:
The patent performs preliminary finish time calculation at the ingress node before packets enter the network. By calculating and attaching finish time metadata in advance, the system can handle large burst transmissions while ensuring that latency upper limits are guaranteed, as the finish time scheduling decision is made beforehand rather than reactively at each node.
Data Source
AI summary
A network latency guarantee system includes a packet processing unit configured to generate new metadata by storing arrived packets and calculating finish times of stored packets, a scheduling unit configured to extract a packet having a minimum finish time by comparing the finish times of the stored packets, and a packet output unit configured to output scheduled packets through output ports.