Bandwidth Allocation for Delay-Sensitive Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In best-effort networks, the lack of prioritization for real-time packets leads to delays and potential packet drops due to bottlenecks, as they are not differentiated from non-time-sensitive traffic, causing queuing issues at service provider routers.
Innovation Solution
Implementing a system where bandwidth is allocated dynamically between classes of traffic, with proactive management by client edge routers to assign specific percentages to time-sensitive and non-time-sensitive data, and adjusting nominal excess capacity to prevent queue buildup, using time probes to monitor and adjust bandwidth allocation based on queuing delay variations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If best-effort delivery is used without prioritization, then network simplicity is maintained, but real-time packets experience queuing delays and potential drops at bottleneck links
Solution Approach 1:
The patent segments traffic into different classes (real-time and non-real-time) and applies different handling policies to each class. The bottleneck router categorizes packets based on their delay sensitivity and processes them through different queues, allowing time-sensitive packets to receive priority treatment while maintaining overall network simplicity through automated classification rules.
Solution Approach 2:
The patent changes the parameter of packet classification by examining packet headers and metadata to identify real-time traffic. By detecting specific packet characteristics (such as protocol type, source/destination information), the system dynamically adjusts routing decisions to prioritize time-sensitive packets without requiring complex manual configuration.
2Productivity
If bandwidth is allocated to ensure bottleneck link utilization, then throughput is improved, but queuing delay increases for time-sensitive traffic
Solution Approach 1:
The patent implements dynamic bandwidth allocation where the bottleneck router continuously monitors queue conditions and adjusts packet scheduling in real-time. When real-time packets are detected in the queue, the router dynamically prioritizes their transmission over non-real-time traffic, ensuring that bandwidth is allocated based on current traffic conditions rather than static provisions.
Solution Approach 2:
The patent employs feedback mechanisms where the bottleneck router monitors queue depth and packet delay metrics, then adjusts its packet scheduling decisions accordingly. When queuing delay for real-time packets exceeds thresholds, the router increases priority handling for subsequent real-time packets, creating a closed-loop control system that optimizes both utilization and delay performance.
3Loss of time
If priority queuing is implemented for real-time packets, then queuing delay is reduced, but device complexity increases at the router
Solution Approach 1:
The patent implements self-service mechanisms where packets essentially identify their own priority requirements through header information and metadata. The bottleneck router performs automated classification based on predefined rules examining packet characteristics, eliminating the need for complex manual configuration or intelligent agents at each router node.
Solution Approach 2:
The patent simplifies router complexity by changing the classification parameters to rely on easily extractable packet header fields and metadata rather than deep packet inspection. By using shallow classification based on protocol type, source/destination addresses, and other readily available packet attributes, the router achieves priority queuing with minimal processing overhead.
Data Source
AI summary
In one embodiment, a system and method include determining bandwidth of a link that connects a local modem to a remote router. A first percentage of the bandwidth is assigned to a first class of data and a second percentage of bandwidth is assigned to a second class of data. The remaining percentage of the bandwidth is assigned for nominal excess capacity. The flow of first class of data and second class of data are controlled to below respective percentages of the bandwidth.


