Two-Layer Alpha Buffer Management with Dynamic RED
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network systems face challenges in dynamically managing shared buffers across multiple entities, such as hosts and sub-queues, to ensure fair resource allocation and efficient congestion management, particularly when using protocols like DCTCP and RED, which may lead to inefficient packet dropping and ECN marking.
Innovation Solution
A two-layer Alpha-based buffer management scheme with Dynamic RED (DRED) is implemented, using alpha parameters to dynamically allocate buffer space and adjust RED thresholds, ensuring fairness across hosts and sub-queues, and providing guaranteed and maximum buffer limits, while dynamically deciding on packet dropping or ECN marking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional RED or DCTCP protocols are used for congestion management, then packet dropping and ECN marking are performed, but resource allocation fairness and buffer management efficiency deteriorate
Solution Approach 1:
The buffer management system is segmented into two hierarchical layers: host-level buffers and sub-queue-level buffers. Each layer has its own alpha parameter for independent control. This segmentation allows fine-grained buffer allocation to multiple hosts and their respective sub-queues, improving fairness and efficiency while maintaining manageable complexity through modular parameter control.
2Reliability
If static buffer allocation is used for multiple hosts, then buffer guarantee is provided, but buffer utilization efficiency and adaptability worsen
Solution Approach 1:
The system employs dynamic buffer allocation where each host and sub-queue has an alpha parameter that determines its buffer share. The available buffer space is dynamically calculated based on current occupancy and the alpha parameters, allowing the buffer allocation to adapt to changing traffic conditions while maintaining guaranteed minimum buffers for each entity.
3Productivity
If dynamic buffer sharing is implemented across hosts and sub-queues, then buffer utilization improves, but allocation fairness and control complexity worsen
Solution Approach 1:
The system applies local quality by assigning specific alpha parameters to each host and sub-queue based on their requirements. Each entity receives differentiated buffer allocation according to its local needs, with the calculation logic applied locally at each level of the hierarchy, simplifying overall control while achieving fair and efficient resource distribution.
Data Source
AI summary
Methods and apparatus for two-layer Alpha-based buffer management with dynamic RED. A two-layer hierarchical sharing scheme using alpha parameters is provided. A buffer is dynamically shared across upper-level entities, such as hosts, using one set of alpha parameters, then a dynamically-adjusted buffer portion allocated for an upper level entity is shared among its lower level entities (e.g., sub queues) using a separate set of low-level alpha parameters. The memory spaces for the upper- and lower-level entities may be dynamically redistributed. Determinations to drop and/or mark and ECN field of received packets are performed using Dynamic RED, which employs dynamic thresholds and associated dynamic probabilities.