Access Gateway Aggregated Shaping for Multi-Class Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing access gateways struggle to effectively prioritize and manage competing data and voice traffic in communication networks, particularly VoIP, video, and data traffic, due to differing bandwidth, latency, and jitter requirements, often relying on work-conserving schedulers or single bandwidth services.
Innovation Solution
The implementation of an access gateway with a classification module that sorts and prioritizes traffic into multiple classes of service flows, using queues and token buckets to manage bandwidth and latency, ensuring strict priority scheduling based on committed and excess information rates, and VLAN tagging for efficient transmission over Ethernet virtual circuits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If a work-conserving scheduler is used to prioritize traffic, then VoIP traffic can be prioritized with low latency and jitter, but the system cannot effectively manage multiple prioritized classes of service flows simultaneously
Solution Approach 1:
The patent segments traffic into multiple prioritized classes of service flows (e.g., voice, video, data) and manages each class separately using dedicated queues and shaping mechanisms. This allows VoIP traffic to receive priority treatment with low latency while simultaneously managing other traffic types with different requirements, resolving the contradiction between prioritizing single class and managing multiple classes.
Solution Approach 2:
The patent introduces a new dimension of service class management by implementing aggregated shaping that operates across multiple queues and service flows simultaneously. This multi-dimensional approach allows the system to prioritize traffic within individual classes while also coordinating across classes to manage overall bandwidth and resources, enabling both low latency for VoIP and effective multi-class management.
2Device complexity
If a single bandwidth controlled service is provided, then bandwidth management is simplified, but the system cannot accommodate different bandwidth, latency, and jitter requirements of various traffic types
Solution Approach 1:
The patent divides the single bandwidth-controlled service into multiple service classes, each with its own bandwidth parameters, shaping mechanisms, and quality of service requirements. This segmentation allows the system to accommodate different bandwidth, latency, and jitter requirements for voice, video, and data traffic while maintaining manageable complexity through standardized shaping protocols for each class.
Solution Approach 2:
The patent changes the bandwidth management parameters dynamically for different service classes based on traffic requirements. Each class can have customized CIR (committed information rate), EIR (excess information rate), and other shaping parameters that are adjusted according to the specific needs of voice, video, or data traffic, enabling flexible accommodation of diverse requirements without overwhelming complexity.
3Reliability
If multiple prioritized classes of service flows are managed separately, then each class can receive optimized service, but the overall system complexity increases
Solution Approach 1:
The patent merges the management of multiple prioritized classes of service flows into a unified aggregated shaping mechanism. Instead of managing each class entirely separately with independent complex control logic, the system combines them into a coordinated structure where classes are aggregated and shaped together, reducing overall system complexity while maintaining optimized service quality for each class through shared resources and coordinated scheduling.
Data Source
AI summary
A system coupled between at least one input port and at least one output port comprises at least one queue, each queue being identified by a QID and operable to receive and buffer data in at least one service flow from the at least one input port. The system further comprises a predetermined at least one token allocated to each queue, each token indicative whether a predetermined amount of data may be dequeued from a queue and transmitted to the output port. The system comprises at least one group of queues where each queue in the group has a subordinate QID identifying a subordinate queue in the group having a lower priority for reallocating unused tokens. The at least one output port receives at least one output flow comprising the dequeued data from the at least one queue.