Dynamic Flow Threshold Control in Electro-Optical Hybrid Switch Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current electro-optical hybrid switch networks face inefficiencies in data transfer due to the inability to effectively distinguish and manage Mice and Elephant flows, leading to suboptimal use of optical lines and increased power consumption in data centers.
Innovation Solution
A communication control method that dynamically adjusts a flow threshold based on monitoring data such as blocking rates, buffer utilization, and packet transfer latency to determine whether flows should be routed through optical or electric networks, optimizing the use of optical lines and reducing latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a fixed threshold is used to distinguish Mice flows from Elephant flows, then the classification process is simple, but the data transfer efficiency is suboptimal due to inability to adapt to varying network conditions
Solution Approach 1:
The patent applies dynamics by transitioning from a fixed threshold to a dynamic threshold that adapts to network conditions. The threshold is adjusted based on real-time monitoring of network load, blocking rates, and optical line availability, allowing the system to optimize flow classification continuously rather than using a static value.
Solution Approach 2:
The patent implements feedback mechanisms by monitoring network performance metrics (blocking rates, buffer utilization, packet transfer latency) and using this information to adjust the flow threshold. This closed-loop control enables the system to learn from past performance and adapt the threshold to achieve optimal data transfer efficiency under varying conditions.
2Productivity
If neural network learning is used to dynamically determine the threshold, then the data transfer efficiency improves, but the calculation amount for determining the threshold increases
Solution Approach 1:
The patent changes parameters by using multiple monitor data parameters (blocking rate, buffer utilization, packet transfer latency) rather than relying solely on flow size. This multi-parameter approach allows for more accurate threshold determination without requiring complex neural network learning, reducing computational energy consumption while maintaining high data transfer efficiency.
Solution Approach 2:
The patent applies partial action by selectively using monitoring data and threshold adjustment mechanisms only when necessary to optimize performance. Rather than continuously performing complex calculations, the system adjusts the threshold based on significant changes in network conditions, reducing unnecessary computational energy consumption while still achieving efficient data transfer.
3Use of energy by stationary object
If optical lines are set up for Elephant flows, then power consumption is reduced compared to electric switches, but the optical lines may not be used efficiently without proper flow rate monitoring
Solution Approach 1:
The patent implements feedback by continuously monitoring flow rates on optical lines and using this information to adjust routing decisions. When flow rates drop below certain thresholds, the system can disconnect optical lines or redirect flows to electric networks, ensuring that optical resources are used efficiently and power consumption is minimized when optical lines are not fully utilized.
Solution Approach 2:
The patent applies dynamics by making the optical line configuration adaptive rather than static. The system dynamically sets up, maintains, or disconnects optical lines based on real-time flow rate monitoring and network conditions, allowing optimal power consumption and resource utilization to be achieved under varying traffic patterns.
4Use of energy by stationary object
If the threshold is lowered to transfer more flows through optical lines, then power consumption decreases, but the blocking rate at optical line setup increases
Solution Approach 1:
The patent changes parameters by using multiple monitor data parameters (blocking rate, buffer utilization, packet transfer latency) rather than relying solely on flow size. This multi-parameter approach allows for more accurate threshold determination without requiring complex neural network learning, reducing computational energy consumption while maintaining high data transfer efficiency.
Solution Approach 2:
The patent applies partial action by selectively using monitoring data and threshold adjustment mechanisms only when necessary to optimize performance. Rather than continuously performing complex calculations, the system adjusts the threshold based on significant changes in network conditions, reducing unnecessary computational energy consumption while still achieving efficient data transfer.
Data Source
AI summary
In order to perform data transfer efficiently in an electro-optical hybrid switch network, monitor data that is a blocking rate in an optical line switching network, an amount concerning flows transferred by a connection apparatus to an electric packet network, a buffer utilization state in telecommunication devices within the electric packet network, or packet transfer latency or the like in telecommunication devices is obtained, and based on the monitor data, a threshold for distinguishing a first flow to be transferred by the connection apparatus through the optical line switching network from a second flow to be transferred by the connection apparatus through the electric packet network is changed, wherein the threshold is to identify, as the first flow, a flow that has a size exceeding the threshold and identify, as the second flow, a flow that has a size that is equal to or less than the threshold.