Hierarchical Queue Selection Circuit for High-Speed Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional queue selection methods in communication networks require increased circuit scale when dealing with a large number of queues, making them impractical for high-speed and high-capacity operations.
Innovation Solution
The method involves grouping queues into a tree structure and selecting groups in parallel steps, reducing the need for registers to manage packet presence and allowing for high-speed queue selection without increasing the circuit scale.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the number of queues is increased from several hundred to several tens of thousands, then the queue selection capability and traffic handling capacity are improved, but the circuit scale becomes enormous and impractical
Solution Approach 1:
The patent divides the queue selection process into multiple stages. Instead of selecting from all queues directly, the system first divides queues into groups, then selects a group, and finally selects a queue within that group. This multi-stage segmentation reduces the circuit scale at each stage while maintaining the ability to handle tens of thousands of queues.
Solution Approach 2:
The patent introduces a hierarchical dimension to the queue selection process by organizing queues into groups and then into supergroups. This adds a grouping dimension that reduces the direct selection complexity from O(N) to O(log N), where N is the number of queues.
2Productivity
If the number of queues is increased to handle larger traffic volume, then the traffic handling capacity is improved, but the register scale for managing packet presence becomes enormous
Solution Approach 1:
The patent segments the register management by creating group presence registers instead of individual queue presence registers. Each group has a single presence register indicating whether any queue in that group has packets, reducing the total number of registers from N (number of queues) to sqrt(N) or log(N) depending on the hierarchy level.
Solution Approach 2:
The patent uses a hierarchical copying structure where group presence information is maintained at multiple levels. The presence information is copied from individual queues to groups, and from groups to supergroups, allowing efficient querying without maintaining full individual queue state at all levels.
3Reliability
If conventional queue selection methods are used with a large number of queues, then complete queue coverage is achieved, but selection errors increase and high-speed operation becomes difficult
Solution Approach 1:
The patent segments the selection process into parallel stages of group selection and queue selection. This segmentation allows each stage to operate independently and in parallel, reducing the overall selection time while maintaining accuracy through the hierarchical structure that ensures complete queue coverage.
Data Source
AI summary
A queue selection method for controlling selection of many queues without increasing the circuit scale is provided. Queues are organized into groups, and each group is created as a tree structure with a plurality of steps, and a queue is selected by selecting a group of each step. By this, even if the number of queues is enormous, it is sufficient to provide registers for managing the presence of packets only for the number of groups selected in each step, and it becomes unnecessary to provide registers for all of the queues, so an increase of registers can be suppressed even if the number of queues increases. It is preferable that group selection in each step is performed in parallel independently from pipeline processing so as to maintain high-speed operation.


