Latency-Based Packet Forwarding with Distributed Policy Queuing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current communication networks lack the capability to specify quantifiable latency Service Level Objectives (SLOs) for packet transmission, leading to unpredictable delays that can impact applications requiring stringent latency guarantees, such as Virtual Reality, Tactile Internet, and industrial control systems.
Innovation Solution
The implementation of a distributed algorithm within network nodes that processes packets based on measured delays, remaining delay budgets, and predicted downstream delays to ensure packets are transmitted within specified latency bounds, using a forwarding header that includes accumulated delay, minimum latency, and latency policies to prioritize and queue packets effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If traditional networks deliver packets as quickly as possible without latency policies, then packet transmission speed is improved, but latency control precision deteriorates
Solution Approach 1:
The patent changes the parameter of packet forwarding from simple speed optimization to latency-budget-based forwarding. Each packet carries a latency policy parameter that specifies maximum acceptable latency, and network nodes use this parameter to determine forwarding decisions, transforming the network from a speed-optimized system to a latency-controlled system.
Solution Approach 2:
The patent implements feedback by having receiving nodes measure actual packet latency and send feedback messages to sending nodes. The sending nodes use this feedback to adjust their forwarding decisions and latency budgets, creating a closed-loop control system that continuously optimizes latency performance based on actual network conditions.
2Measurement precision
If centralized coordination is used to manage latency SLOs, then latency control precision is improved, but device complexity and controller dependence worsen
Solution Approach 1:
The patent enables self-service by embedding latency policies directly in packet headers and enabling each network node to autonomously make forwarding decisions based on these policies. Nodes independently calculate latency budgets and make queue selection decisions without centralized coordination, allowing the network to self-manage latency SLOs distributed across all nodes.
Solution Approach 2:
The patent segments the centralized latency control function into distributed decision-making at each network node. Instead of one central controller managing all latency decisions, each node independently handles its own forwarding decisions based on packet latency policies, dividing the control function across multiple autonomous units.
3Reliability
If larger buffer sizes are used to handle varying latencies, then reliability is improved, but loss of time worsens due to increased queuing delay
Solution Approach 1:
The patent applies dynamics by making buffer allocation dynamic rather than static. Instead of allocating fixed large buffers to all packets, the system dynamically allocates buffer space based on each packet's specific latency budget and priority, allowing packets with tighter latency requirements to use smaller buffers while packets with more flexible deadlines can use larger buffers.
Solution Approach 2:
The patent implements local quality by providing different buffering characteristics to different packets based on their individual latency requirements. High-priority packets with strict latency SLOs receive preferential buffer allocation and faster service, while lower-priority packets with more flexible deadlines receive standard buffering, creating localized quality differences in buffer management.
4Measurement precision
If packets are prioritized based on destination policies, then latency control precision is improved, but device complexity worsens due to multiple queue management
Solution Approach 1:
The patent applies partial action by implementing prioritization only for packets that carry latency policy information, rather than prioritizing all packets. Best-effort packets without latency policies are handled by standard queueing mechanisms, while latency-sensitive packets receive enhanced prioritized handling, providing partial prioritization action only where needed.
Data Source
AI summary
Latency Based Forwarding (LBF) techniques are presented for the management of the latencies, or delays, of packets forwarded over nodes, such as routers, of a network. In addition to a network header indicating a destination node for receiving the packet, a packet also includes an LBF header indicating the packets accumulated delay since leaving the sender, a maximum latency for the entire journey from the sender to the receiver and a minimum latency for the entire journey from the sender to the receiver. When a packet is received at a node, based on the accumulated delay, the maximum latency, and the minimum latency, the node places the packet in a forwarding queue to manage the delays between the sender and the receiver. The LBF can also indicate a policy for the forwarding node to used when determining the enqueueing of the packet.