Downstream Scheduler with Transitioning Hold Queues for Fair Bandwidth Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current DOCSIS network schedulers face challenges in managing downstream communication bandwidth efficiently, leading to queuing delays and packet losses due to congestion, and fail to fairly allocate bandwidth among service flows, potentially violating Quality of Service (QoS) requirements.
Innovation Solution
A downstream scheduler with a queuing module and QoS policies that utilize transitioning hold queues to manage service flow bursts and TCP flows, ensuring fair bandwidth allocation and rate limiting without additional complexity, by dynamically adjusting queue priorities and service credits based on loading conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a downstream scheduler implements complicated queuing algorithms to ensure adequate delivery rates during network congestion, then the delivery rate reliability is improved, but the device complexity increases
Solution Approach 1:
The patent segments the single transmit queue into multiple hold queues (first hold queue, second hold queue, etc.), each dedicated to specific service flows. This segmentation allows different queuing strategies to be applied to different queue segments, simplifying the overall scheduling logic while maintaining reliable delivery rates through structured organization of traffic streams.
Solution Approach 2:
The patent implements dynamic queue state transitions where hold queues can be promoted to or demoted from the transmit queue based on loading conditions and service flow requirements. This dynamic adjustment allows the scheduler to adapt to changing network conditions without requiring complex static algorithms, maintaining reliability while managing complexity through flexible, condition-based transitions.
2Productivity
If the downstream scheduler manages large bursts of downstream network traffic, then the traffic handling capacity is improved, but queuing delays and packet losses occur due to congestion
Solution Approach 1:
By dividing the transmit queue into multiple hold queues, the patent prevents single-queue congestion from affecting all service flows simultaneously. Each hold queue can be managed independently, allowing traffic to flow through alternative paths when congestion occurs in one queue, thereby reducing overall queuing delays while maintaining high traffic handling capacity.
Solution Approach 2:
The patent performs preliminary classification and placement of packets into appropriate hold queues before transmission. By pre-organizing traffic into service-flow-specific queues based on loading conditions and QoS requirements, the system avoids last-minute scheduling decisions that cause delays, enabling smoother traffic flow and reduced queuing time during congestion events.
3Productivity
If the downstream scheduler allocates bandwidth to large traffic flows, then the throughput is improved, but smaller traffic flows experience excessive queuing latency
Solution Approach 1:
The patent segments service flows into different hold queues based on their characteristics and QoS requirements. Large traffic flows can be directed to one hold queue while smaller, latency-sensitive flows are directed to another, allowing simultaneous optimization of throughput for large flows and latency for small flows without one adversely affecting the other.
Solution Approach 2:
Different queuing and scheduling characteristics are applied to different hold queues based on local requirements. Queues containing latency-sensitive small flows can use different transmission priorities or transition rules compared to queues containing throughput-optimized large flows, enabling each segment to be optimized for its specific traffic characteristics while maintaining overall system efficiency.
4Reliability
If the downstream scheduler implements rate limiting mechanisms to meet DOCSIS requirements, then the QoS compliance is improved, but the device complexity increases
Solution Approach 1:
The patent combines rate limiting functionality with the hold queue structure itself, rather than implementing rate limiting as a separate mechanism. The hold queues inherently enforce rate limiting by controlling the flow of packets from service flows into the transmit queue, merging classification, queuing, and rate control functions into a unified structure that meets DOCSIS requirements without adding separate complex rate-limiting components.
Data Source
AI summary
A network device utilizes fair bandwidth allocation techniques based on configurable service classes. The network device comprises a scheduler that receives a packet associated with a service flow for delivery to a destination. The scheduler identifies a service credit associated with the service flow that represents a bandwidth allocation available for consumption by an associated service flow. Based on the identified service credit, the downstream scheduler assigns the packet to one of a plurality of hold queues. The number of hold queues may remain static during the allocation of bandwidth, which reduces the complexity of the scheduler and permits fair bandwidth allocation to occur. Such a queuing architecture also allows the downstream scheduler to achieve rate limiting and minimum bandwidth guarantee without using separate algorithms and/or architectures.