-
Notifications
You must be signed in to change notification settings - Fork 0
/
init.py
253 lines (218 loc) · 7.31 KB
/
init.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
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
import random
from disk import Disk
from data import Data
from replica import Replica
from nodeheapbig import *
from LP1 import *
import time
# import global_vars as gl
import matplotlib.pyplot as plt
def init_disks(DISK_num, Latency_range):
disk_id = 0
disks = list()
_min = Latency_range.get("min")
_max = Latency_range.get("max")
print(" min = "+str(_min)+", max = "+ str(_max))
aver = (_min + _max) / 2
sigma = (aver - _min) / 2 # 95% 的面积
xli = list()
yli = list()
while disk_id < DISK_num:
latency = int(round(random.gauss(aver, sigma)))
if latency > 0 and latency < DISK_num:
xli.append(disk_id)
yli.append(latency)
disk = Disk(disk_id, latency)
disks.append(disk)
disk_id = disk_id + 1
# plt.figure()
# plt.xticks(fontsize=14)
# plt.yticks(fontsize=14)
# plt.xlabel("disk", fontsize=14)
# plt.ylabel("latency", fontsize=14)
# plt.step(xli, yli, label="DISK-LATANCY", where="post")
# plt.show()
# print(len(disks))
return disks
def init_datas(DATA_NUM, REP_FAC):
print("--------------- Init_datas ---------------")
data_list = list()
replica_list = list()
k = 0
for i in range(0, DATA_NUM):
data = Data(i)
for j in range(0, REP_FAC):
replica = Replica(k, i)
data.rep_set.append(k)
replica_list.append(replica)
k = k + 1
# print(len(data.rep_set))
data_list.append(data)
# print(len(data_list))
return data_list, replica_list
def deploy_replica(disks, replica_list):
print("--------------- deploy_replica ---------------")
for replica in replica_list:
while True:
disk_id = random.randint(0, int(len(disks))-1)
flag = False
for rep in disks[disk_id].rep_list:
if replica_list[rep].get_data_id() == replica.data_id:
flag = True
break
if flag == False:
break
# print(disk_id)
replica.set_disk_id(disk_id)
disks[disk_id].rep_list.append(replica.id)
# for disk in disks:
# print(len(disk.rep_list))
def hadoop_naive(disks, data_list, replica_list):
for data in data_list:
disks[replica_list[data.rep_set[0]].disk_id].add_active(data.rep_set[0])
heapbig = HeapPriorityQueueBig()#大顶堆
for disk in disks:
heapbig.add(len(disk.active_list)*disk.read_latency, disk.id)
print("hadoop_naive max (load , id) = ", heapbig.max())
xli = list()
yli = list()
sum = 0
for disk in disks:
xli.append(disk.id)
# sum = sum + len(disk.active_list)
yli.append(len(disk.active_list)*disk.read_latency)
# print("sum:", sum)
# print(xli)
# plt.figure()
# plt.xticks(fontsize=14)
# plt.yticks(fontsize=14)
# plt.xlabel("disk", fontsize=14)
# plt.ylabel("latency", fontsize=14)
# plt.step(xli, yli, label="DISK-load", where="post")
# plt.show()
return xli, yli, heapbig.max()
def random_naive(disks, data_list, replica_list):
for data in data_list:
replica_id = random.randint(0, int(len(data.rep_set))-1)
disks[replica_list[data.rep_set[replica_id]].disk_id].add_active(data.rep_set[0])
xli = list()
yli = list()
heapbig = HeapPriorityQueueBig() # 大顶堆
# sum = 0
for disk in disks:
# sum = sum + len(disk.active_list)
xli.append(disk.id)
yli.append(len(disk.active_list) * disk.read_latency)
heapbig.add(len(disk.active_list) * disk.read_latency, disk.id)
# print("sum:", sum)
# plt.figure()
# plt.xticks(fontsize=14)
# plt.yticks(fontsize=14)
# plt.xlabel("disk", fontsize=14)
# plt.ylabel("latency", fontsize=14)
# plt.step(xli, yli, label="DISK-load", where="post")
# plt.show()
print("random max (load , id) = ", heapbig.max())
return xli, yli, heapbig.max()
def greedy(disks, data_list, replica_list):
for data in data_list:
pre = 100000000
index = 0
for i in range(0, len(data.rep_set)):
if pre > len(disks[replica_list[data.rep_set[i]].disk_id].active_list)*disks[replica_list[data.rep_set[i]].disk_id].read_latency:
pre = len(disks[replica_list[data.rep_set[i]].disk_id].active_list)*disks[replica_list[data.rep_set[i]].disk_id].read_latency
replica_id = i
disks[replica_list[data.rep_set[replica_id]].disk_id].add_active(data.rep_set[0])
xli = list()
yli = list()
sum = 0
heapbig = HeapPriorityQueueBig() # 大顶堆
for disk in disks:
# sum = sum + len(disk.active_list)
xli.append(disk.id)
yli.append(len(disk.active_list) * disk.read_latency)
heapbig.add(len(disk.active_list) * disk.read_latency, disk.id)
# plt.figure()
# plt.xticks(fontsize=14)
# plt.yticks(fontsize=14)
# plt.xlabel("disk", fontsize=14)
# plt.ylabel("latency", fontsize=14)
# plt.step(xli, yli, label="DISK-load", where="post")
# plt.show()
# print("sum:", sum)
print("greedy max (load , id) = ", heapbig.max())
return xli, yli, heapbig.max()
def DLP_random(disks, data_list, replica_list, REP_FAC):
# x=1
# y=1
# print(len(disks))
T = list()
for di in disks:
T.append(di.read_latency)
#data 也就是task数量
data = list()
for dat in data_list:
data_disk = list()
for i in range(0, len(dat.rep_set)):
data_disk.append(replica_list[dat.rep_set[i]].disk_id)
data_disk.sort()
data.append(data_disk)
# print(T, data)
# ans, = LP_solver(T, data)
ans, tasks = LP_solver(T, data, REP_FAC)
end = time.clock()
# print(end - init_end, "seconds LP time ")
p = ans["x"]
q = list()
for i in range(0, len(data)):
q.append(random.random())
# print(tasks)
# print(q)
# print(p)
results = list()
start = 0
end = start + REP_FAC
j = 0 # task
while end < len(p):
pis = p[start:end]
sum = 0
i = 0
for pi in pis:
sum = sum + pi
if q[j] <= sum:
if start + i != 0:
d_i = tasks[j].index(start + i)
results.append([j, d_i])
break
else:
d_i = tasks[j].index(-1)
results.append([j, d_i])
break
i = i + 1
j = j + 1
start = start + REP_FAC
end = end + REP_FAC
# print(tasks)
# print(q)
# print(results)
for result in results:
disks[result[1]].add_active(result[0])
xli = list()
yli = list()
sum = 0
heapbig = HeapPriorityQueueBig() # 大顶堆
for disk in disks:
# sum = sum + len(disk.active_list)
xli.append(disk.id)
yli.append(len(disk.active_list) * disk.read_latency)
heapbig.add(len(disk.active_list) * disk.read_latency, disk.id)
# plt.figure()
# plt.xticks(fontsize=14)
# plt.yticks(fontsize=14)
# plt.xlabel("disk", fontsize=14)
# plt.ylabel("latency", fontsize=14)
# plt.step(xli, yli, label="DISK-load", where="post")
# plt.show()
# print("sum:", sum)
print("HTS-rdm max (load , id) = ", heapbig.max())
return xli, yli, heapbig.max()