Network ASIC Interrupt Propagation via Service Level Manager
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multiple-interrupt propagation schemes in network ASICs lack an efficient mechanism for categorizing and managing interrupts based on service levels, leading to inconsistent and potentially delayed processing of interrupts across different destinations.
Innovation Solution
A multiple-interrupt propagation scheme that categorizes interrupts into different service levels using a manager and user-defined register specification, allowing instantaneous or delayed propagation to destinations based on priority, with a script generating registers for each interrupt to manage and forward them accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If interrupts are propagated immediately without categorization, then processing speed is maintained, but service level differentiation and resource allocation efficiency deteriorate
Solution Approach 1:
The patent segments interrupts into different service levels (e.g., high priority, low priority) using categorization logic in the manager module. This segmentation allows the system to differentiate between critical and non-critical interrupts, enabling efficient resource allocation by directing high-priority interrupts to immediate processing while deferring low-priority ones, thus resolving the contradiction between maintaining propagation speed and improving resource allocation efficiency.
Solution Approach 2:
The patent changes the parameter of interrupt propagation timing based on service level classification. High-priority interrupts maintain immediate propagation (speed preserved), while low-priority interrupts undergo delayed propagation (resource allocation improved). This parameter change approach allows the system to optimize both speed and productivity by adapting propagation behavior to interrupt criticality.
2Device complexity
If all interrupts are processed with equal priority, then simplicity is maintained, but system responsiveness to critical events deteriorates
Solution Approach 1:
The patent applies segmentation by dividing interrupts into distinct service levels with different processing priorities. The manager module categorizes incoming interrupts based on their service level, enabling the system to respond reliably to critical events while maintaining manageable complexity through structured classification rather than ad-hoc handling.
Solution Approach 2:
The patent introduces an intermediary manager module that sits between interrupt sources and processing destinations. This intermediary categorizes and routes interrupts based on service levels, shielding the rest of the system from complexity while ensuring critical interrupts receive immediate attention, thus improving system responsiveness without proportionally increasing overall complexity.
3Productivity
If interrupts are delayed for batching, then resource utilization improves, but processing time for individual interrupts worsens
Solution Approach 1:
The patent applies local quality by treating different service level interrupts differently in terms of propagation timing. High-priority interrupts receive immediate local processing without batching delays, while low-priority interrupts are eligible for batching and delayed propagation to improve resource utilization. This local quality approach ensures that time-critical interrupts are not penalized by batching while still achieving improved resource utilization for non-critical interrupts.
Solution Approach 2:
The patent changes the propagation timing parameter based on service level classification. High-priority interrupts maintain zero delay (processing time preserved), while low-priority interrupts accept delayed propagation for batching (resource utilization improved). This selective parameter change resolves the contradiction by applying different timing strategies to different interrupt categories.
Data Source
AI summary
Embodiments of the present invention are directed to a multiple-interrupt propagation scheme, which is an automated mechanism for the specification and creation of interrupts. Interrupts originating at leaf nodes of a network chip are categorized into different service levels according to their interrupt types and are propagated to a master of the network chip via a manager. For each interrupt, depending on its service level, the manager either instantaneously propagates the interrupt or delays propagation of the interrupt to the master. The master forwards the interrupts to different destinations. A destination can be a processing element that is located on the network chip or externally on a different chip.