Multi-timescale Packet Marker Using Throughput-Value Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current per-packet marking based bandwidth sharing control methods are inefficient, particularly when using token buckets, as they are only suitable for a few drop precedence levels and require an unrealistic number of token buckets as the number of levels increases, leading to memory and computational demands.
Innovation Solution
Configuring multiple timescales with Throughput-Value Functions (TVFs) and determining distances between them to select a packet value for marking, reusing existing per-packet value core stateless schedulers, and optimizing bitrate measurement on longer timescales to manage shared resources efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If token buckets are used for per-packet marking based bandwidth sharing control, then resource sharing control can be implemented, but the number of token buckets required increases unrealistically as the number of drop precedence levels increases, leading to high memory and computational demands
Solution Approach 1:
The patent segments the single TVF into multiple TVFs, each corresponding to a different timescale (e.g., short-term, medium-term, long-term). Each TVF is associated with a token bucket, but now the number of token buckets equals the number of timescales (e.g., 3) rather than the number of drop precedence levels (e.g., 7). This segmentation allows the system to handle multiple precedence levels by evaluating packet values against multiple timescale perspectives simultaneously, reducing the unrealistic growth of token buckets while maintaining adaptability to different drop precedence requirements
Solution Approach 2:
The patent introduces a new dimension (timescale) to the resource sharing control mechanism. Instead of using a single TVF that considers all drop precedence levels in one dimension, the system creates multiple TVFs across different timescale dimensions. Each TVF evaluates packet importance from its specific timescale perspective, and the packet's final marked value is determined by combining evaluations across all timescale dimensions. This dimensional transformation allows the system to handle complex multi-level precedence control with a manageable number of token buckets
2Ease of operation
If per-packet marking is implemented with multiple drop precedence levels, then fine-grained bandwidth sharing control is achieved, but memory and computational requirements increase significantly
Solution Approach 1:
The patent segments the control mechanism into multiple independent TVFs, each handling a specific timescale. Each TVF maintains its own token bucket and evaluation logic, allowing independent computation for each timescale. This segmentation enables the system to achieve fine-grained control by combining results from multiple segmented evaluations rather than using a single complex evaluation that would require excessive resources
Solution Approach 2:
The patent applies partial action by having each TVF evaluate only the packet value from its specific timescale perspective, rather than all TVFs evaluating all aspects. The system performs partial evaluations across multiple timescales and combines them to achieve the complete control decision. This approach avoids the excessive computational burden of having a single comprehensive evaluation mechanism while maintaining fine-grained control capability
Data Source
AI summary
A network node (120), such as a packet marking node, efficiently measures the bitrates of incoming packets on a plurality of timescales (TSs). A throughput-value function (TVF) is then graphed to indicate the throughput-packet value relationship for that TVF. Then, starting from the longest TS and moving towards the shortest TS, the packet marking node determines (88) a distance between the TVFs of different TSs at the measured bitrates. To determine the packet marking, the packet marking node selects a random throughput value between 0 and the bitrate measured on the shortest TS. Depending on how the random value relates to the measured bitrates, a TVF, and the distances to add to the random value, is then selected to determine (92) a packet value (PV) with which to mark the packet. The packet marking node then marks (94) the packet according to the determined PV.