Packet Routing Engine Allocation for Bandwidth and Compute Balance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Adaptive routing in network devices requires significant computational overhead, making it costly to replicate multiple engines, while limiting the number of engines reduces available bandwidth.
Innovation Solution
Implementing a combination of adaptive and random routing engines within a network device, where large packets use adaptive routing and small packets use random routing, based on packet size or type, to optimize bandwidth utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple adaptive routing engines are replicated within a device, then routing accuracy and bandwidth efficiency are improved, but device cost and complexity increase significantly
Solution Approach 1:
The patent segments the routing engine population into two distinct groups: adaptive routing engines for large packets and random routing engines for small packets. This segmentation allows each engine type to be optimized for its specific packet size category, achieving high routing accuracy where needed while keeping the overall system cost-effective through the use of simpler engines for the majority of small packets.
Solution Approach 2:
The patent applies local quality by assigning different routing methodologies to different packet sizes based on their specific requirements. Large packets receive the computationally intensive adaptive routing treatment locally, while small packets are handled by simpler random routing engines locally. This localized approach optimizes performance for each packet category without requiring all engines to be complex adaptive types.
2Device complexity
If the number of routing engines is limited, then device cost is reduced, but available bandwidth and packet processing capacity decrease
Solution Approach 1:
The patent segments the routing engine functionality into two tiers: a limited number of adaptive routing engines for large packets and a larger population of simple random routing engines for small packets. This segmentation enables the device to achieve high packet processing capacity through the numerous simple engines while maintaining routing accuracy for bandwidth-intensive large packets through the fewer sophisticated engines.
Solution Approach 2:
The patent uses copying by replicating multiple instances of the simple random routing engine architecture, which is inexpensive to copy, for handling the majority of small packets. Meanwhile, fewer copies of the complex adaptive routing engine are deployed for large packets. This copying strategy maximizes packet processing capacity while controlling device cost.
3Measurement precision
If adaptive routing is used for all packets, then routing accuracy is maximized, but computational overhead and processing time increase
Solution Approach 1:
The patent segments packet handling based on size, applying adaptive routing only to large packets where routing accuracy has the greatest impact on overall performance. Small packets, which constitute the majority of traffic, are routed using simple random routing that requires minimal computation time. This segmentation maximizes routing accuracy for bandwidth-intensive traffic while minimizing average processing time across all packets.
Solution Approach 2:
The patent applies partial action by using adaptive routing selectively only for large packets rather than for all packets. This partial application of the computationally intensive adaptive routing methodology provides sufficient routing accuracy for the packets that need it most while avoiding the excessive computational overhead that would result from applying adaptive routing to every packet regardless of size.
Data Source
AI summary
Embodiments of the present disclosure are directed to network devices utilizing routing engines of different types. One routing engine or set of routing engines can use, for example, adaptive routing which, while computationally heavy, more accurately routes packets and therefore provides a more efficient use of available bandwidth. The other routing engine or set of routing engines can utilize, for example, random routing which is much less computationally heavy. Packets can be routed by selectively utilizing one of these engines based on the size of the packet or the type of packet. For example, large packets such as Maximum Transfer Unit (MTU) packets can be routed using the adaptive routing engine and small packets such as an ACKnowledge (ACK) packet can be routed using the random routing engine.