On-Chip Interconnect Dynamic QoS via Priority Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current on-chip interconnect technologies face challenges in providing end-to-end Quality-of-Service (QoS) due to their static nature, leading to inefficiencies in bandwidth management and increased costs, particularly in complex Systems-on-Chip (SoCs), where virtual channels fail to manage performance within classes of service and introduce additional costs.
Innovation Solution
A method and system that implement a dynamic and configurable QoS mechanism using a dedicated priority network and priority virtual path to manage bandwidth and prioritize data transmission based on urgency, allowing for efficient end-to-end QoS across different classes of service, independent of network configuration, and enabling flexible performance regulation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If virtual channels are used to implement QoS, then separation of service classes is improved, but performance management within the same class deteriorates and costs increase
Solution Approach 1:
The patent extracts the QoS management function from the network infrastructure and relocates it to the initiator components. Each initiator is equipped with a QoS manager that independently manages its own traffic flows and performance, removing the need for complex network-wide virtual channel infrastructure while maintaining QoS capabilities.
Solution Approach 2:
Initiators are empowered with self-service QoS management capabilities through integrated QoS managers that automatically monitor, control, and optimize their own traffic flows. This self-managing approach eliminates the need for centralized network control and complex infrastructure duplication.
2Ease of manufacture
If static QoS approach is used, then implementation simplicity is improved, but end-to-end QoS performance deteriorates
Solution Approach 1:
The patent implements dynamic QoS management at the initiator level, where QoS parameters and performance targets are continuously monitored and adjusted based on actual traffic conditions. This dynamic approach maintains implementation simplicity while achieving superior end-to-end QoS performance compared to static network-wide approaches.
Solution Approach 2:
The system allows flexible configuration and modification of QoS parameters at the initiator level, enabling dynamic adjustment of performance targets, traffic patterns, and service requirements without requiring changes to the underlying network infrastructure.
3Reliability
If initiators wait for interconnect to supply data, then data consistency is improved, but productivity deteriorates
Solution Approach 1:
The QoS manager at each initiator performs preliminary actions by pre-managing traffic flows, prioritizing data packets, and preparing transmission queues before data requests are issued. This preliminary organization enables initiators to maintain consistency requirements while minimizing waiting time and maximizing throughput.
Data Source
AI summary
In a method for making an on-chip interconnect for conveying between a set of initiators and a set of targets in which traffic is organized in classes of service, priority values representing the classes of service are associated with the traffic. The method further includes propagating the priority values towards the points of the network where an arbitration is performed between two classes of service of the traffic, and providing arbitration as a function of the priority values.