Burst-Sized Linked List Elements for Queue Buffering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern routers and switches face challenges in efficiently buffering large amounts of data during silent periods in communication protocols, requiring effective data storage solutions to maintain quality of service and support features like pause and record in video applications.
Innovation Solution
The implementation of a buffering system that uses burst-sized elements stored in a linked list, where descriptors are stored in off-chip memory, allowing for dynamic queue sizing and fast burst access, enabling efficient data retrieval and reconstruction of data streams without the need for pre-allocation of memory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If descriptors are stored individually in a queue, then each descriptor can be accessed separately, but memory access efficiency deteriorates due to inability to perform burst transfers
Solution Approach 1:
The queue is segmented into multiple descriptor groups, where each group contains a burst-sized number of descriptors (e.g., 64 descriptors per group). This segmentation allows the memory interface to transfer entire groups in single burst operations while maintaining the ability to access individual descriptors within each group through group base addresses and offsets.
2Ease of manufacture
If queue size is fixed and pre-allocated, then memory allocation is simple, but adaptability deteriorates when handling variable data volumes during silent periods
Solution Approach 1:
The queue structure uses dynamic memory allocation where descriptor groups can be allocated and deallocated based on actual data volume requirements. The queue head and tail pointers dynamically track the current state, allowing the queue to grow and shrink during silent periods without requiring fixed pre-allocation, thus adapting to variable data volumes while maintaining efficient burst access patterns.
3Quantity of substance
If descriptors are stored in off-chip memory, then memory requirements are reduced, but access speed deteriorates compared to on-chip storage
Solution Approach 1:
Multiple descriptors are merged into single memory storage locations organized as descriptor groups. Each group contains multiple descriptors that can be transferred together in a single burst operation, effectively combining multiple individual access operations into one unified memory transaction, thereby maintaining high throughput despite off-chip storage.
4Productivity
If burst-sized elements are used, then data throughput is improved through efficient memory transfers, but device complexity increases due to additional buffer management requirements
Solution Approach 1:
The descriptor group structure serves multiple functions simultaneously: it enables efficient burst transfers, provides dynamic queue resizing capability, maintains individual descriptor accessibility, and simplifies memory management through unified group-based allocation. This multi-functionality reduces overall system complexity despite the enhanced capabilities.
Data Source
AI summary
Methods, apparatus, circuitry, and systems to construct a queue including a plurality of elements are provided. A method includes receiving metadata describing a first buffer; generating a descriptor based on the metadata; and storing the descriptor in an element. The element is configured to store a predetermined number of descriptors and the element includes an amount of memory corresponding to a burst size of a component configured to read the metadata to control access to the first buffer.