Distributions Dataset for Fuzzy Based Fog Load Balancer

Published: 28 September 2019| Version 1 | DOI: 10.17632/fz8rsr7c2k.1
Simar Preet Singh


This dataset consists of 5 columns, namely X1, X2, X3, X4 and Y and has 3000 entries corresponding to each random variable. These random variables (X1, X2, X3 and X4) correspond to the following variables in paper: "Design and Exploration of Load Balancers for Fog Computing Using Fuzzy Logic": X1: Traffic Load X2: Delay Sensitivity X3: Energy Consumption X4: Link Saturation Y: Link Health Here, X1, X2, X3 and X4 are independent variables and Y is a dependent variable. Value of Y depends on all independent variables. These variables are mapped to Fuzzy Terms on the basis of Network Traffic Load Balancing. The fuzzy terms mapped for different scenarios in a day, are as follows: Sensor_TrafficLoad MinimumLoad Trapezoid (0.000, 0.080, 0.240, 0.330) AverageLoad Trapezoid (0.240, 0.330, 0.570, 0.660) PeakLoad Gaussian (0.500, 0.200) Sensor_DelaySenstivity NoDelay GaussianProduct (0.500, 0.200, 0.500, 0.200) AverageDelay Sigmoid (0.500, 20.000) TimeOutDelay SShape (0.000, 1.000) Sensor_EnergyConsumption LowConsumption Sigmoid (0.510, 0.204) AverageConsumption Sigmoid (0.510, 0.204) PeakConsumption SShape (0.020, 1.000) Sensor_LinkSaturation LowSaturation Ramp (0.000, 1.000) AverageSaturation Ramp (0.000, 1.000) FullSaturation SShape (0.000, 1.000) OUTPUT: LinkHealth LowQuality Ramp (0.000, 1.000) MediumQuality Ramp (0.000, 1.000) HighQuality Ramp (0.000, 1.000) This dataset is used to devise 3-level fuzzy load balancer design in the mentioned paper. This dataset consists of 3000 records, where each sub-parameter/sub-variable corresponds to 1000 records for 3-level fuzzy load balancer design. Eg: For the parameter 'Traffic Load', sub-variables are: MinimumLoad, AverageLoad and PeakLoad. So this dataset consists of 1000 records computed using the random distributions of MinimumLoad, 1000 records computed using the distributions of AverageLoad and 1000 records computed using the distributions of PeakLoad, thus resulting in a total of 3000 records for 'Traffic Load'. This division is computed to cover maximum different possible scenarios to which Network Traffic Load belongs to.


Steps to reproduce

This is a dataset for simulating fuzzy based load balancer. This fuzzy based fog load balancer requires distributions, that are based on parameters: Traffic Load, Delay Sensitivity, Energy Consumption and Link Saturation. Based on all these parameters, Link Health parameter is computed. The values of these all parameters were generated based on random variable distributions. The steps to reproduce the dataset are as follows: The dataset can be reproduced by modifying random.py python file (https://svn.python.org/projects/python/trunk/Lib/random.py) in python environment. The same random.py file can also be found on github repository (https://github.com/python/cpython/blob/master/Lib/random.py). Modified random.py file is also attached in this dataset. The fuzzy terms selected for this dataset are mapped to the following functions in random.py file: GaussianProduct --> Gaussian() Sigmoid --> lognormvariate() SShape --> lognormvariate() Ramp --> trapezoid() Triangle --> triangular() For Trapezoid distributions, methods and python code that were used are placed at URL: https://docs.scipy.org/doc/scipy-1.2.0/reference/generated/scipy.stats.trapz.html.


Thapar University


Artificial Intelligence, Cloud Computing, Fuzzy Computing, Fuzzy Logic, Fuzzy Set, Machine Learning, Machine Learning Algorithm, Machine Learning Theory, Bio-Inspired Computing, Fuzzy Programming, Fuzzy Control, Fog Computing