forked from mkheirkhah/cno-rl
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathrl_env.py
200 lines (161 loc) · 8.71 KB
/
rl_env.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
###########################################################################
# Title: Implementation of RL algortihm for UC2, 5G-MEDIA project
# Author: Morteza Kheirkhah
# Institution: University College London (UCL), UK
# Email: m.kheirkhah@ucl.ac.uk
# Homepage: http://www.uclmail.net/users/m.kheirkhah/
# Note: Original code has been written for the Pensieve paper (SIGCOMM'17)
###########################################################################
import numpy as np
MBPS = 1000000.0
BITS_IN_BYTE = 8.0
RANDOM_SEED = 42
BITRATE_LEVELS = 10
TOTAL_VIDEO_CHUNCK = 48 # 48 #100 #20
FRAME_INTERVAL = 1 #50 # frame
#VIDEO_SIZE_FILE = './video_sizes/video_size_'
#VIDEO_BIT_RATE = [3000, 5000, 8000, 12000, 15000, 20000, 25000, 30000, 40000, 50000]
#VIDEO_BIT_RATE_AVG = [3129, 5216, 8349, 12505, 15630, 20841, 26055, 31294, 41727, 52156]
#VIDEO_BIT_RATE = [5000, 6000, 7000, 8000, 9000, 10000, 11000, 12000, 15000, 20000]
#VIDEO_BIT_RATE_AVG = [5216, 6500, 7400, 8349, 9400, 10440, 11420, 12630, 15630, 20841] # dry-run
#VIDEO_BIT_RATE = [4000, 8000, 12000, 20000, 40000, 45000] #Kbps
#VIDEO_BIT_RATE_AVG = [3855, 7551, 11244, 18740, 37480, 42000] #Kbps
BACKGROUND_TRAFFIC_1 = [0, 30] #Mbps
BACKGROUND_TRAFFIC_2 = [0, 5, 10, 15, 20, 25, 30, 35, 30, 25, 20, 15, 10, 5, 0]
BACKGROUND_TRAFFIC_3 = [0, 10, 20, 30, 20, 10, 0]
BACKGROUND_TRAFFIC_4 = [0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22,
24, 26, 28, 30, 32, 34, 36, 38, 40, 38, 36, 34,
32, 28, 26, 24, 22, 20, 18, 16, 14, 12, 10, 8, 6, 4, 2]
BACKGROUND_TRAFFIC_5 = [0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 40, 35, 30, 25, 20, 15, 10, 5, 0]
BACKGROUND_TRAFFIC_5_1 = [0, 2, 4, 6, 8, 10, 12, 14, 15, 14, 12, 10, 8, 6, 4, 2, 0] # dry-run
#BACKGROUND_TRAFFIC_5_1 = [0, 2, 4, 6, 8, 10, 12, 13, 10, 8, 6, 4, 2, 0]
BACKGROUND_TRAFFIC_6 = [0, 0, 0, 0, 0, 0, 0, 0, 10, 10, 10, 10, 10, 10, 10, 10]
BACKGROUND_TRAFFIC_7 = [0, 0, 0, 0, 0, 0, 0, 0, 20, 20, 20, 20, 20, 20, 20, 20]
BACKGROUND_TRAFFIC_8 = [0, 0, 0, 0, 0, 0, 0, 0, 30, 30, 30, 30, 30, 30, 30, 30]
BACKGROUND_TRAFFIC_9 = [0, 0, 0, 0, 0, 0, 0, 0, 35, 35, 35, 35, 35, 35, 35, 35]
BACKGROUND_TRAFFIC_10 = [0, 0, 0, 0, 0, 0, 0, 0, 40, 40, 40, 40, 40, 40, 40, 40]
BACKGROUND_TRAFFIC_0 = [BACKGROUND_TRAFFIC_1, BACKGROUND_TRAFFIC_2,
BACKGROUND_TRAFFIC_3, BACKGROUND_TRAFFIC_4,
BACKGROUND_TRAFFIC_5, BACKGROUND_TRAFFIC_6,
BACKGROUND_TRAFFIC_7, BACKGROUND_TRAFFIC_8,
BACKGROUND_TRAFFIC_9, BACKGROUND_TRAFFIC_10]
# CNO_RAND_MAX = 50000000 / 8.0 #Mbytes/s
# CNO_RAND_MIN = 1000000 / 8.0 #Mbytes/s
# ALGO = {0: "REAL", 1: "UNIFORM", 2: "NORMAL", 3: "SAWTHOOTH", 4: "NORM_BIT"}
# TRAFFIC_MODEL = ALGO[0]
class Environment:
def __init__(self, all_cooked_time, all_cooked_bw, random_seed=RANDOM_SEED):
np.random.seed(random_seed)
self.video_chunk_counter = 0
# self.video_size = {} # in bytes
# for bitrate in range(BITRATE_LEVELS):
# self.video_size[bitrate] = []
# with open(VIDEO_SIZE_FILE + str(bitrate)) as f:
# for line in f:
# self.video_size[bitrate].append(
# (int(line.split()[0]), float(line.split()[1])))
# def get_video_size(self, quality):
# video_chunk_size = self.video_size[quality][self.video_chunk_counter][0] # equal to frame's size
# video_chunk_br = self.video_size[quality][self.video_chunk_counter][1] # equal to frame's bit_rate
# self.video_chunk_counter += 1 # to keep track of chunks/frames globally
# return video_chunk_size, \
# video_chunk_br
def get_video(self, quality, video_bit_rates):
video = video_bit_rates[quality] * 1000 #bps
video = np.random.normal(video, video / 20.0)
return video
def get_background_0(self, video_count):
outer_index = video_count % len(BACKGROUND_TRAFFIC_0)
inner_index = self.video_chunk_counter % len(BACKGROUND_TRAFFIC_0[outer_index])
background = BACKGROUND_TRAFFIC_0[outer_index][inner_index] * MBPS #bps
background = np.random.normal(background, background / 10.0)
return background
def get_background(self, rand1, rand2, bg_traffic_pattern, bg_traffic_dist):
if (bg_traffic_dist == 'normal'):
index = self.video_chunk_counter % len(bg_traffic_pattern)
background = bg_traffic_pattern[index] * MBPS #bps
background = np.random.normal(background, background / 10.0)
return background
if (bg_traffic_dist == 'uniform'):
index = self.video_chunk_counter % len(bg_traffic_pattern)
background = bg_traffic_pattern[index] * MBPS #bps
delta = background/10.0
background = np.random.uniform(background-delta, background+delta)
return background
# index = self.video_chunk_counter % len(BACKGROUND_TRAFFIC_5_1)
# background = BACKGROUND_TRAFFIC_5_1[index] * MBPS #bps
# background = np.random.normal(background, background / 10.0)
# return background
# background = 0.0
# if (TRAFFIC_MODEL) == "UNIFORM":
# background = np.random.randint(CNO_RAND_MIN, CNO_RAND_MAX)
# elif (TRAFFIC_MODEL == "NORMAL"):
# background = np.random.normal(CNO_RAND_MAX, CNO_RAND_MAX / 10.0)
# elif (TRAFFIC_MODEL == "REAL"):
# background *= rand1
# elif (TRAFFIC_MODEL == "NORM_BIT"):
# bitrate = (VIDEO_BIT_RATE_AVG[rand2] * 1000) / 8.0
# background = np.random.normal(bitrate, bitrate / 10.0)
# return background
def get_video_chunk(self, quality, video_count, bg_traffic_pattern, bg_traffic_dist ,lc, video_bit_rates):
# print("\n********** get_video_chunk **********")
# print(bg_traffic_pattern, bg_traffic_dist, video_bit_rates,lc)
self.video_chunk_counter += 1 # to keep track of chunks/frames globally
#video_chunk_size, video_chunk_br = self.get_video_size(quality)
loss_rate_list = []
ava_ca_list = []
loss_rate_frac_list = []
ava_ca_frac_list = []
frame_counter = 1
rand1 = np.random.randint(1, 15)
rand2 = np.random.randint(1, len(video_bit_rates))
capacity = lc
#background = 0.0
loss_rate = 0.0
video = 0.0
while True: # download video frames
background = self.get_background(rand1, rand2, bg_traffic_pattern, bg_traffic_dist) #bps
#background = self.get_background_0(video_count) #bps
# if (video_count < 5000):
# background = 0.0
# else:
# background = 30 * MBPS
video = self.get_video(quality, video_bit_rates)
ava_ca = capacity - video - background #bps
ava_ca = 0.0 if ava_ca < 0 else ava_ca
ava_ca_frac = ava_ca / capacity
loss_rate = (video + background) - capacity
loss_rate = 0 if loss_rate < 0 else loss_rate
loss_rate_frac = loss_rate / float(video + background)
########
# print("bg [{0}] video[{1}] ava_ca [{2}] ava_ca_frac [{3}] loss_rate [{4}] loss_rate_frac [{5}]"
# .format(background/MBPS, video/MBPS, ava_ca/MBPS, ava_ca_frac, loss_rate, loss_rate_frac))
# print (capacity,"\t", video,"\t", background)
ava_ca_frac_list.append(ava_ca_frac) # bytes/s
loss_rate_frac_list.append(loss_rate_frac)
loss_rate_list.append(loss_rate)
ava_ca_list.append(ava_ca)
if (frame_counter < FRAME_INTERVAL):
frame_counter += 1 # a frame ends so start another one
# if (self.video_chunk_counter >= TOTAL_VIDEO_CHUNCK): # a chunk with several frames ends
# break
# video_chunk_size, video_chunk_br = self.get_video_size(quality)
# continue
else:
break
# exit while loop
mean_loss_rate_frac = np.mean(loss_rate_frac_list)
mean_ava_ca_frac = np.mean(ava_ca_frac_list)
mean_loss_rate = np.mean(loss_rate_list) #bps
mean_ava_ca = np.mean(ava_ca_list) #bps
assert(len(ava_ca_frac_list) == 1)
end_of_video = False
if self.video_chunk_counter >= TOTAL_VIDEO_CHUNCK:
end_of_video = True
self.video_chunk_counter = 0
#print("\n***** end of video ******")
return end_of_video, \
mean_loss_rate_frac, \
mean_ava_ca_frac, \
mean_loss_rate, \
mean_ava_ca