Network Processor Scheduling Circuit for Load Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network processors face issues with non-optimal scheduling, leading to degraded performance and functionality due to speculative processing, single long packet flows stalling other flows, and sub-optimal use of processing resources, as well as the need for cycle-costly software locking protocols for atomic access.
Innovation Solution
A scheduling circuit that conducts work conserving, load balancing, and atomic scheduling among processing resources, eliminating the need for software locking protocols by detecting critical sequences and assigning them to a single processor for atomic execution, while cycling through flows to prevent single long flows from stalling others and ensuring efficient resource utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If speculative scheduling is used to improve resource utilization, then processor productivity increases, but system reliability deteriorates due to unpredictable performance when packets are not available
Solution Approach 1:
The scheduling circuit automatically detects critical sequences and performs atomic scheduling without external intervention. The system self-manages the detection of critical sequences, assignment to processors, and enforcement of atomic execution, eliminating the need for external coordination mechanisms while maintaining reliable performance.
2Reliability
If a single long flow is assigned to one processor to ensure atomic execution, then reliability improves, but productivity deteriorates as other flows are stalled
Solution Approach 1:
The scheduling circuit segments flows by detecting critical sequences and applying atomic scheduling only to those specific segments rather than entire flows. This allows non-critical portions of flows to be processed in parallel by multiple processors, maintaining atomicity where needed while maximizing overall throughput.
3Reliability
If software locking protocols are used to ensure atomic access, then reliability improves, but productivity deteriorates due to cycle-costly overhead
Solution Approach 1:
The patent replaces software-based locking protocols with a hardware scheduling circuit that performs atomic scheduling. This substitution eliminates the cycle-costly overhead of software locking by implementing critical sequence detection and atomic assignment in hardware, maintaining reliability while significantly improving processing speed.
4Productivity
If load balancing is implemented to distribute flows evenly, then productivity improves, but reliability deteriorates when critical sequences are fragmented across processors
Solution Approach 1:
The scheduling circuit applies different scheduling qualities to different parts of flows based on their criticality. Critical sequences receive atomic scheduling with dedicated processor assignment, while non-critical portions benefit from load balancing. This local differentiation maintains atomic execution integrity where required while maximizing overall productivity.
Data Source
AI summary
A system and method are described for work conserving, load balancing, and scheduling by a network processor. For example, one embodiment of a system includes a plurality of processing cores, including a scheduling circuit, at least one source processing core that generates at least one task and at least one destination processing core that receives and processes the at least one task, and generates a response. The scheduling circuit of the exemplary system receives the at least one task and conducts a load balancing to select the at least one destination processing core. In an embodiment, the scheduling circuit further detects a critical sequences of tasks, schedules those tasks to be processed by a single destination processing core, and, upon completion of the critical sequence, conducts another load balancing to potentially select a different processing core to process more tasks.