日常の記録 (2024 年 4月)

• I have completed introducing the Apache's function to log all access information, such as global IP address and FQDN, to prepare for future DDoS attacks. I want to learn more about how the real Internet works via hands-on training.

The part of yesterday's access logs. .

• Here, this is one of the simplest codes for verifying the results of Fig. 5 in the paper written by G. Ramezan et al.

gained from the below source codes(Python3).

simVal.py

import simpy
import numpy as np
import sys

def process(env, lamb, mu, D, maxNum):
    np.random.seed(0)
    meanWT, curProcNum = 0, 0
    queue = []
    for i in range(maxNum):
        # サービス実行開始(以降、waitQueueに貯まる)                                                                                                                                                         
        srvTime = np.random.exponential(1/mu)
        exeTime = srvTime + env.now
        while True:
            arvTime = np.random.exponential(1/lamb)
            if env.now + arvTime < exeTime:
                # キューにTxを入れる                                                                                                                                                                        
                queue.append(env.now + arvTime)
                yield env.timeout(arvTime)
            elif env.now + arvTime >= exeTime and len(queue) < D:
                # Queueを無理やりDまで増やしてサービス実行。PoWは考慮していない)                                                                                                                            
                yield env.timeout(exeTime - env.now)
                # DまでQを増やす、時間も経過                                                                                                                                                                
                while len(queue) < D:
                    arvTime = np.random.exponential(1/lamb)
                    yield env.timeout(arvTime) # 平均1/1000秒待機                                                                                                                                           
                    queue.append(env.now)
                procTime = env.now

                # 平均処理時間を計算                                                                                                                                                                        
                meanWT = (np.mean(procTime - np.array(queue)) * len(queue) + meanWT * curProcNum) / (len(queue) + curProcNum)
                curProcNum += len(queue)
                # Queue情報の更新                                                                                                                                                                           
                queue = []
                break
            else:
                # queueがD個以上で、サービスを即実行                                                                                                                                                        
                yield env.timeout(exeTime - env.now)
                # 平均処理時間を計算                                                                                                                                                                       
                meanWT = (np.mean(exeTime - np.array(queue)) * len(queue) + meanWT * curProcNum) / (len(queue) + curProcNum)
                curProcNum += len(queue)
                # Queue情報の更新                                                                                                                                                                           
                queue = []
                break
    print(meanWT)

plotVals.py

import matplotlib.pyplot as plt
import numpy as np

def calcN0(lamb, mu, D):
    den = D * (lamb + mu) / lamb + (lamb + mu) / mu * np.power(lamb / (lamb + mu), D)
    return 1 / den
def calcAveL(lamb, mu, D):
    N0 = calcN0(lamb, mu, D)
    term = (lamb + mu) / lamb * (D * (D - 1)) / 2 + lamb / mu * (lamb / mu + D) * np.power(lamb / (lamb + mu), D - 1)
    return N0 * term
def main():
    # データの準備                                                                                                                                                                                         
    lamb, mu = 1000, 1
    x = np.linspace(1, 2000, 100)
    y = calcAveL(lamb, mu, x) * mu / lamb
    x1 = np.array([1, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 1900, 2000])
    y1 = np.array([0.9975690907442448, 0.9905734288416077, 0.9816150305985084, 0.9850361091339387, 0.9478050715571071, 0.9540101248803657, 0.9084462285183247,
                    0.9073192404072165, 0.8976292031050388, 0.9098063513863652, 0.9036150641522788, 0.9068923298319815, 0.9320381920990181, 0.9351002615154222,
                    0.9630543412957223, 0.9758201499136501, 1.006384431635652, 1.0329673620743782, 1.0545567690515925, 1.0911256312581148, 1.1211433826940767])
    # グラフの描画                                                                                                                                                                                          
    plt.figure(figsize=(8, 6))
    plt.plot(x, y, '-', color="orange")
    plt.plot(x1, y1, 'o', color="blue")
    plt.xlim(0, 2000)
    plt.ylim(0.8, 1.2)
    plt.xlabel('$D$')
    plt.ylabel('$\mu \\bar{W}$')
    plt.title('Normalized average waiting time.')
    plt.grid(True)
    plt.show()
main()

• Today, I have changed the URL on these diary pages.