Network Apparatus Traffic Microburst Detection via Histogram Thresholding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In network systems, detecting microbursts—sudden increases in traffic within a short period—is challenging due to the difficulty in setting appropriate thresholds, as traffic tendencies vary among customers using different types of data communication, making uniform threshold settings inadequate.
Innovation Solution
A network apparatus with a traffic amount measuring unit and frequency counters that measure and visualize traffic trends by determining the range each traffic amount falls into, allowing for optimized threshold setting tailored to individual customers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a uniform threshold is set for all customers based on physical band and number of customers, then the threshold setting process is simple, but the threshold is not always appropriate for actual operation because traffic tendencies differ between customers
Solution Approach 1:
The patent segments the single threshold value into multiple thresholds corresponding to different customers. Each customer is assigned their own threshold based on their historical traffic tendencies, allowing customized microburst detection for each customer rather than using a uniform threshold for all customers.
Solution Approach 2:
The patent applies local quality by making each customer's threshold specific to their traffic characteristics. Instead of a global uniform threshold, each customer receives a locally optimized threshold based on their own traffic history, improving detection accuracy for diverse traffic patterns.
2Measurement precision
If an optimized threshold for each customer is used, then microburst detection accuracy is improved, but it is necessary to find a technique capable of visualizing traffic tendency which increases system complexity
Solution Approach 1:
The patent performs preliminary action by accumulating and analyzing traffic amounts over time before setting thresholds. The system collects historical traffic data for each customer, visualizes traffic tendencies through histograms, and establishes customized thresholds in advance, enabling accurate microburst detection without complex real-time analysis.
Solution Approach 2:
The patent implements self-service by enabling the system to automatically analyze its own traffic data and generate customized thresholds for each customer. The switch apparatus autonomously accumulates traffic statistics, visualizes patterns, and determines appropriate thresholds without requiring external manual configuration, reducing operational complexity.
3Reliability
If traffic amount is monitored and compared with a threshold to detect microburst, then microburst detection is achieved, but it is not easy to determine the appropriate threshold
Solution Approach 1:
The patent applies feedback by using historical traffic data to inform threshold setting. The system accumulates past traffic amounts, analyzes patterns through histograms, and uses this feedback to establish appropriate thresholds for each customer, making threshold determination systematic rather than arbitrary.
Solution Approach 2:
The patent changes parameters by transforming raw traffic amount data into histogram distributions across multiple ranges. This parameter transformation reveals traffic tendencies and enables data-driven threshold selection, converting the difficult qualitative judgment of threshold appropriateness into a quantitative analysis process.
Data Source
AI summary
A traffic amount measuring unit measures the traffic amount of frames received in each of cycles, and outputs the measured traffic amount for each cycle. A plurality of frequency counters are provided so as to correspond to a plurality of ranges, and retain frequencies of each of the plurality of ranges, respectively. A frequency counter updating unit determines in which of the plurality of ranges the traffic amount output for each cycle is included, and updates the frequency of the frequency counter corresponding to the range obtained by the determination result.


