Capacity Planning via Distribution Model Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current capacity planning methods for mobile edge computing networks are inadequate in accurately predicting network traffic and user experience, as they rely on traditional average performance counters that mask resource insufficiency and poor user experience during busy hours, especially in bursty traffic scenarios.
Innovation Solution
A capacity planning method that uses distribution models, such as Zeta and Pareto distributions, to match the quantity and length of service packets, allowing for more accurate bandwidth control based on user experienced rate distribution models, ensuring precise capacity planning and improved user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional average performance counters are used for capacity planning, then the measurement system is simple and easy to implement, but the measurement precision is insufficient and masks resource insufficiency during busy hours
Solution Approach 1:
The patent segments the measurement approach by introducing multiple distribution models (Poisson, Zeta, Pareto) to represent different traffic patterns at granular levels. Instead of using a single average counter, the system divides traffic measurement into packet quantity distribution and packet length distribution, each modeled separately to capture bursty traffic characteristics accurately.
Solution Approach 2:
The patent changes the measurement parameters from simple average counters to distribution model parameters (e.g., Poisson parameter λ, Zeta parameter s, Pareto parameters m and α). These parameters capture the statistical characteristics of bursty traffic, enabling accurate representation of traffic volume variations during different network conditions including busy hours.
2Measurement precision
If distribution models with fine granularity are used to accurately describe user experienced rate, then the capacity planning accuracy is improved, but the device complexity increases due to multiple distribution models and parameters
Solution Approach 1:
The patent implements dynamic model selection where the system adapts to different traffic conditions by choosing appropriate distribution models. The measurement system dynamically adjusts which distribution model (Poisson, Zeta, or Pareto) to use based on the observed traffic characteristics, allowing accurate representation of user experienced rate under varying network conditions without requiring all models to operate simultaneously.
Solution Approach 2:
The patent adds a new dimension to capacity planning by introducing distribution model parameters as additional measurement dimensions. Instead of only measuring average traffic volume, the system measures traffic volume distribution characteristics (packet quantity distribution and packet length distribution), providing a multi-dimensional view that captures bursty traffic patterns and enables accurate user experienced rate calculation.
3Measurement precision
If bandwidth control is performed based on matched distribution models and parameters, then the capacity planning accuracy is improved, but the ease of operation decreases due to the complexity of model matching and parameter analysis
Solution Approach 1:
The patent implements self-service through automated model matching and parameter extraction. The system automatically fits observed traffic data to appropriate distribution models (Poisson, Zeta, or Pareto) and extracts the corresponding parameters without requiring manual intervention. This automation handles the complexity of model matching and parameter analysis, making bandwidth control operationally simple while maintaining high accuracy.
Solution Approach 2:
The patent incorporates feedback mechanisms where the system continuously monitors traffic patterns, compares observed distributions with theoretical distribution models, and adjusts bandwidth allocation based on the fitted parameters. This feedback loop enables automatic adaptation to changing traffic conditions, maintaining accurate capacity planning without requiring manual reconfiguration or complex operational interventions.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
This application provides a capacity planning method and apparatus. The method includes: matching a distribution model based on a quantity of service packets in each transmission time interval within specified duration, to obtain a matched first distribution model, matching a distribution model based on a length of the service packet, to obtain a second distribution model, and performing bandwidth control based on the first distribution model, the second distribution model, a distribution parameter of the first distribution model, and a distribution parameter of the second distribution model. Because bandwidth control is performed based on a matched distribution model and a distribution parameter of a distribution model, it helps provide more accurate capacity planning for a user.