Network Path Measurement in Clos Switching Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network path measurement methods in Clos switching networks, such as the stage-by-stage backpressure method, are not accurate enough and inefficient, as they require significant link bandwidth and fail to determine shared available bandwidth among multiple paths, leading to suboptimal load balancing.
Innovation Solution
A method that calculates an aggregate available bandwidth by segmenting paths into series-parallel segments, determining the smallest available bandwidth among these segments, which allows for dynamic traffic scheduling between egress ports, thereby improving the efficiency and accuracy of network path measurement and load balancing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the stage-by-stage backpressure method is used to measure network path congestion, then congestion information can be obtained through switching nodes, but the measurement accuracy is insufficient and the method cannot determine shared available bandwidth among multiple paths
Solution Approach 1:
The patent segments the network measurement task by introducing probe packets that traverse specific paths through the Clos switching network. Each probe packet carries identification information to track its path, and the measurement is segmented into probe generation, path traversal, and response collection phases. This segmentation enables accurate measurement of shared available bandwidth across multiple paths simultaneously.
Solution Approach 2:
The patent implements a feedback mechanism where probe packets are sent from ingress switching nodes through the network, and egress switching nodes send back responses containing bandwidth measurement data. This feedback loop provides real-time information about the shared available bandwidth of multiple paths, enabling accurate congestion status determination without requiring complex coordination between intermediate switching nodes.
2Productivity
If individual path bandwidths are measured separately, then each path's bandwidth can be determined, but the link bandwidth occupancy is high and the measurement efficiency is low
Solution Approach 1:
The patent merges multiple path measurements into a single coordinated operation by using probe packets that can traverse multiple paths simultaneously. The ingress switching node generates probe packets that are distributed across multiple paths, and the egress switching node collects responses from all paths in one measurement cycle. This merging approach reduces link bandwidth occupancy compared to measuring each path separately while maintaining measurement accuracy.
Solution Approach 2:
The probe packet mechanism serves multiple functions: it identifies the path traversed, measures the available bandwidth, and provides congestion status information. This multi-functional approach eliminates the need for separate measurement protocols for each path, thereby improving measurement efficiency and reducing the overall bandwidth occupancy required for network monitoring.
3Ease of operation
If the stage-by-stage backpressure method is used, then congestion information can be propagated through the network, but the overhead is significant and the dynamic traffic scheduling is suboptimal
Solution Approach 1:
The patent performs preliminary actions by pre-configuring probe packets with path identification information and bandwidth measurement parameters before they enter the network. The ingress switching node prepares the probe packets in advance, specifying the paths to be measured and the bandwidth thresholds for congestion detection. This preliminary preparation reduces the processing overhead during actual measurement and enables faster dynamic traffic scheduling decisions.
Solution Approach 2:
The patent changes the measurement parameter from individual link bandwidth (in backpressure method) to shared path bandwidth. By measuring the aggregate available bandwidth across multiple paths rather than individual link bandwidths, the system reduces measurement overhead while providing more useful information for dynamic traffic scheduling. The probe packets carry parameters that aggregate bandwidth information from multiple paths into a single shared available bandwidth value.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
A network path measurement method, apparatus, and system are provided, so as to improve efficiency of network path measurement. The method includes: obtaining, by a first switching node in a Clos switching network, a first aggregate available bandwidth of a path from the first switching node to a second switching node; obtaining, by the first switching node, a first available bandwidth of a path from a first target port of a third switching node to the first switching node, where the third switching node is a next-stage switching node connected to the first switching node; obtaining, by the first switching node, a second available bandwidth of a path from the second switching node to a fourth switching node, where the fourth switching node is a next-stage switching node connected to the second switching node; and determining, by the first switching node, a second aggregate available bandwidth of a path from the first target port of the third switching node to the fourth switching node, where the second aggregate available bandwidth is a smallest available bandwidth among the first aggregate available bandwidth, the first available bandwidth, and the second available bandwidth.