Platform Interconnect QoS Management via Credit-Based Traffic Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In compute devices, the platform interconnect can become saturated with messages from one workload, adversely affecting the quality of service for other workloads, leading to overprovisioning and excess costs to prevent saturation.
Innovation Solution
Implementing a system where each compute device manages quality of service by obtaining class of service data for workloads, determining credits for components, and selectively allowing or delaying message transmission through the platform interconnect based on available credits, establishing different classes of service for communication channels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the platform interconnect is overprovisioned to prevent saturation, then the reliability of service for all workloads is improved, but the cost increases due to excess data transmission capacity
Solution Approach 1:
The patent segments the platform interconnect capacity into multiple virtual channels, each dedicated to specific workloads or workload categories. This segmentation allows the system to provide guaranteed bandwidth to critical workloads without requiring the entire interconnect to be overprovisioned, thereby reducing overall capacity requirements while maintaining service reliability.
Solution Approach 2:
The patent implements dynamic bandwidth allocation where the available capacity in virtual channels is adjusted based on real-time workload demands and quality of service requirements. This dynamic approach allows the system to optimize capacity utilization across different workloads, preventing saturation in critical paths without requiring excessive static overprovisioning of the entire interconnect.
2Reliability
If the platform interconnect is overprovisioned with more data transmission circuitry, then the quality of service for workloads is improved, but the device complexity increases
Solution Approach 1:
The patent employs a universal virtual channel management mechanism that handles multiple workload types and quality of service requirements through a single integrated control framework. This multi-functional approach consolidates the complexity of managing different interconnect configurations into a unified system, reducing overall device complexity while maintaining high quality of service across diverse workloads.
3Reliability
If separate virtual channels are established for different workloads, then the quality of service for each workload is improved, but the device complexity increases
Solution Approach 1:
The patent implements self-service mechanisms where virtual channels automatically manage their own bandwidth allocation and traffic prioritization based on pre-configured quality of service parameters. This autonomous operation eliminates the need for complex centralized management systems, reducing device complexity while maintaining distinct quality of service for different workloads through distributed control.
Data Source
AI summary
A compute device includes one or more processors, one or more resources capable of being utilized by the one or more processors, and a platform interconnect to facilitate communication of messages between the one or more processors and the one or more resources. The compute device is to obtain class of service data for one or more workloads to be executed by the compute device. The class of service data is indicative of a capacity of one or more of the resources to be utilized in the execution of each corresponding workload. The compute device is also to execute the one or more workloads and manage the amount of traffic transmitted through the platform interconnect for each corresponding workload as a function of the class of service data as the one or more workloads are executed.