-
Notifications
You must be signed in to change notification settings - Fork 3
/
mec.py
405 lines (326 loc) · 17.3 KB
/
mec.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
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
#~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+
# DEFINE WORKFLOW & INFRA CONSTANTS
# wfD range of Data units in a workflow
# wfV range of Task units in a workflow
# wfL range of initial task locations in a workflow
#
# E no of edge servers
# C no of cloud servers
# A no of actions possible = 1+E+C
#
# DR Data Rate : Time Delay in moving a unit of data b/w two locations
# DE Data Energy : Energy consumed in handling per unit of Data
# VR Task Rate : Time Delay in executing a unit of task
# VE Task Energy : Energy consumed in executing a unit of task
# randomize() a function that can uniformly randomize infra constants DR VR DE VE
#~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+~~+
import numpy as np
import matplotlib.pyplot as plt
from math import floor
from scipy.sparse.csgraph import floyd_warshall
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
# Basic Shared functions
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
def effective_bandwidth(M):
""" function for finding shortest path and effective Data Rate (bandwidth) b/w edge or cloud servers
return dist_matrix """
lm = len(M)
for i in range(lm):
for j in range(i+1, lm):
if M[i,j]!=0:
M[i,j] = 1/M[i,j]
M[j,i] = M[i,j]
dist_matrix, predecessors = floyd_warshall(csgraph=M,
directed=False,return_predecessors=True)
return dist_matrix #<--- this turns the diagonals to inf #print(dist_matrix)
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
def int2base(num, base, digs):
""" convert base-10 integer to base-n array of fixed no. of digits
return array (of length = digs)"""
res = np.zeros(digs, dtype=np.int32)
q = num
for i in range(digs):
res[i]=q%base
q = floor(q/base)
return res
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
def base2int(arr, base):
""" convert array from given base to base-10 --> return integer"""
res = 0
for i in range(len(arr)):
res+=(base**i)*arr[i]
return res
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
def strA(arr, start="[", sep=",", end="]"):
""" returns a string representation of an array/list for printing """
res=start
for i in range(len(arr)):
res+=str(arr[i])+sep
return res + end
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
def strD(arr, sep="\n", caption=""):
""" returns a string representation of a dict object for printing """
res="=-=-=-=-==-=-=-=-="+sep+"DICT: "+caption+sep+"=-=-=-=-==-=-=-=-="+sep
for i in arr:
res+=str(i) + "\t\t" + str(arr[i]) + sep
return res + "=-=-=-=-==-=-=-=-="+sep
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
# core classes & functions
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
def COST(infra, workflow, solution):
iLoc = int(workflow.V[0])
L = [x for x in solution]
L.insert(0, iLoc)
L.append(iLoc)
c = 0
for i in range(1, workflow.T+1):
a = L[i]
c += (workflow.D[i-1] * infra.DR[L[i-1], a] + \
workflow.D[i-1] * infra.DE[a] + \
workflow.V[i] * infra.VR[a] + \
workflow.V[i] * infra.VE[a] + \
workflow.D[i] * infra.DE[a] + \
workflow.D[i] * infra.DR[a, L[i+1]] * int(workflow.T==i))
return c
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
class BASELINE:
def __init__(self):
pass
def geteff(self, r): # efficiency
return 1- ((r-self.min_cost) / self.cost_delta)
def total_pies(self, infra, workflow):
return infra.A**workflow.T
def allpie(self, infra, workflow, verbose=0):
""" Compare Cost for all policies on current workflow and find the min & max costs
return TP, costs, min_count, min_cost, min_pies, max_count, max_cost, max_pies """
self.A = infra.A
self.T = workflow.T
self.TP = self.A**self.T
costs = []
for i in range(self.TP):
costs.append( COST (infra, workflow, int2base(i, self.A, self.T)) )
costs = np.array(costs)
min_cost, max_cost = np.min(costs), np.max(costs)
min_pies, max_pies = np.where(costs==min_cost)[0], np.where(costs==max_cost)[0]
min_count, max_count = len(min_pies), len(max_pies)
self.costs = costs
self.min_cost = min_cost
self.min_count = min_count
self.min_pies = min_pies
self.max_cost = max_cost
self.max_count = max_count
self.max_pies = max_pies
self.cost_delta = (self.max_cost-self.min_cost)
if verbose>0:
self.render()
return # TP, costs, min_count, min_cost, min_pies, max_count, max_cost, max_pies
def render(self):
print('TOTAL-POLICIES:', self.TP)
print('\nMIN-Cost:', self.min_cost)
print('MIN-Cost COUNT:', self.min_count)
for i in range(self.min_count):
print('[',self.min_pies[i],']', int2base(self.min_pies[i], self.A, self.T))
print('\nMAX-Cost:', self.max_cost)
print('MAX-Cost COUNT:', self.max_count)
for i in range(self.max_count):
print('[',self.max_pies[i],']', int2base(self.max_pies[i], self.A, self.T))
fig, ax = plt.subplots(1,2, figsize=(16,4))
ax[0].plot(self.costs)
ax[0].scatter(self.min_pies, np.zeros(self.min_count)+self.min_cost, color='green')
ax[0].scatter(self.max_pies, np.zeros(self.max_count)+self.max_cost, color='red')
_=ax[1].hist(self.costs, bins=50)
plt.show()
return
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
class WORKFLOW:
def __init__(self, arg_T, wfD=[1,100], wfV=[1,60], wfL=[0,0], seed=None):
self.T = arg_T
self.wfD = np.array(wfD) # all data size
self.wfV = np.array(wfV) # all task size
self.wfL = np.array(wfL) # initial location
self.prng = np.random.default_rng(seed)
self.new_flow()
def new_flow(self):
self.V = self.prng.uniform(self.wfV[0], self.wfV[1], size=self.T+1)
self.V[0] = self.prng.integers(self.wfL[0],self.wfL[1]+1) # initiak work destination location
self.D = self.prng.uniform(self.wfD[0], self.wfD[1], size=self.T+1)
self.D[0] = self.prng.uniform(self.wfD[0] + (self.wfD[1]-self.wfD[0])/2, self.wfD[1]) # initial data is a bit higher
return
def render(self):
return ('V:{}\nD:{}'.format(self.V, self.D))
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
class INFRA:
def __init__(self, seed=None):
self.prng = np.random.default_rng(seed)
pass
def render(self):
return ('E,C,A: {},{},{}\nVR:{}\nVE:{}\nDE:{}\nDR:\n{}'.format(self.E, self.C, self.A, self.VR, self.VE, self.DE, self.DR))
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
# pre-defined infra objects
#=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
def infra_1():
x=INFRA()
x.E = 1 # no of Edge servers
x.C = 1 # no of Cloud servers
x.A = x.E + x.C +1 # action space
# 1 [DR] Data Rate(delay) # DR contains delay per unit of data
gw8_DR = 1
x.DR = effective_bandwidth(np.array([
# #i0 e1 c2 #
[ 0, 5, 0, ], # i0
[ 0, 0, 300, ], # e1
[ 0, 0, 0, ], # c2
], dtype='float') ) * gw8_DR
x.DRi = np.copy(x.DR)
x.DRrL, x.DRrH = 0.5, 1.5
# 2 [DE] Data Eng # DE contains energy cost per unit of data
gw8_DE = 1
ar_DE = np.array ([0.56, 0.52, 0.55, ], dtype='float')
w8_DE = np.array ([ 1, 1, 1, ], dtype='float') * gw8_DE
x.DE = np.array(np.multiply(ar_DE, w8_DE))
x.DEi = np.copy(x.DE)
x.DErL, x.DErH = 0.5, 1.5
# 3 [VR] Task Rate(delay) # VR contains delay per unit of computation
gw8_VR = 1
x.VR = np.array([ 1/4.5, 1/5, 1/10, ], dtype='float') * gw8_VR
x.VRi = np.copy(x.VR)
x.VRrL, x.VRrH = 0.5, 1.5 # randomize ratio
# 4 [VE] Task Eng # # VE contains energy cost per unit of computation
gw8_VE = 1
ar_VE = np.array ([0.56, 0.52, 0.55, ], dtype='float')
w8_VE = np.array ([1, 1, 1, ], dtype='float') * gw8_VE
x.VE = np.array(np.multiply(ar_VE, w8_VE))
x.VEi = np.copy(x.VE)
x.VErL, x.VErH = 0.5, 1.5 # randomize ratio
#-=============================================
return x
#-=============================================
def infra_2():
x=INFRA()
x.E = 3 # no of Edge servers
x.C = 2 # no of Cloud servers
x.A = x.E + x.C +1 # action space
# 1 [DR] Data Rate(delay) # DR contains delay per unit of data
gw8_DR = 1
x.DR = effective_bandwidth(np.array([
# #i0 e1 e2 e3 c4 c5 #
[ 0, 5, 5, 5, 0, 0, ], # i0
[ 0, 0, 300, 0, 300, 0 ], # e1
[ 0, 0, 0, 300, 300, 0, ], # e2
[ 0, 0, 0, 0, 0, 300,], # e3
[ 0, 0, 0, 0, 0, 500,], # c4
[ 0, 0, 0, 0, 0, 0, ], # c5
], dtype='float') ) * gw8_DR
x.DRi = np.copy(x.DR)
x.DRrL, x.DRrH =0.5, 1.5
# 2 [DE] Data Eng # DE contains energy cost per unit of data
gw8_DE = 1
ar_DE = np.array ([0.56, 0.522, 0.521, 0.522, 0.55, 0.55, ], dtype='float')
w8_DE = np.array ([ 1, 1, 1, 1, 1, 1, ], dtype='float') * gw8_DE
x.DE = np.array(np.multiply(ar_DE, w8_DE))
x.DEi = np.copy(x.DE)
x.DErL, x.DErH = 0.5, 1.5
# 3 [VR] Task Rate(delay) # VR contains delay per unit of computation
gw8_VR = 1
x.VR = np.array([ 1/4.5, 1/5, 1/5.05, 1/5, 1/10, 1/10, ], dtype='float') * gw8_VR
x.VRi = np.copy(x.VR)
x.VRrL, x.VRrH = 0.5, 1.5 # randomize ratio
# 4 [VE] Task Eng # # VE contains energy cost per unit of computation
gw8_VE = 1
ar_VE = np.array ([0.56, 0.52, 0.52, 0.52, 0.55, 0.55, ], dtype='float')
w8_VE = np.array ([1, 1, 1, 1, 1, 1, ], dtype='float') * gw8_VE
x.VE = np.array(np.multiply(ar_VE, w8_VE))
x.VEi = np.copy(x.VE)
x.VErL, x.VErH = 0.5, 1.5 # randomize ratio
#-=============================================
return x
#-=============================================
def infra_3():
x=INFRA()
x.E = 5 # no of Edge servers
x.C = 2 # no of Cloud servers
x.A = x.E + x.C +1 # action space
# 1 [DR] Data Rate(delay) # DR contains delay per unit of data
gw8_DR = 1
x.DR = effective_bandwidth(np.array([
# #i0 e1 e2 e3 e4 e5 c6 c7 #
[ 0, 5, 5, 5, 5, 5, 0, 0, ], # i0
[ 0, 0, 300, 0, 0, 0, 300, 0, ], # e1
[ 0, 0, 0, 300, 0, 0, 300, 0, ], # e2
[ 0, 0, 0, 0, 300, 0, 300, 0, ], # e3
[ 0, 0, 0, 0, 0, 300, 300, 0, ], # e4
[ 0, 0, 0, 0, 0, 0, 0, 300, ], # e5
[ 0, 0, 0, 0, 0, 0, 0, 500, ], # c6
[ 0, 0, 0, 0, 0, 0, 0, 0, ], # c7
], dtype='float') ) * gw8_DR
x.DRi = np.copy(x.DR)
x.DRrL, x.DRrH = 0.5, 1.5
# 2 [DE] Data Eng # DE contains energy cost per unit of data
gw8_DE = 1
ar_DE = np.array ([0.56, 0.522, 0.521, 0.522, 0.521, 0.522, 0.55, 0.55, ], dtype='float')
w8_DE = np.array ([ 1, 1, 1, 1, 1, 1, 1, 1, ], dtype='float') * gw8_DE
x.DE = np.array(np.multiply(ar_DE, w8_DE))
x.DEi = np.copy(x.DE)
x.DErL, x.DErH = 0.5, 1.5
# 3 [VR] Task Rate(delay) # VR contains delay per unit of computation
gw8_VR = 1
x.VR = np.array([ 1/4.5, 1/5, 1/5.05, 1/5, 1/5.05, 1/5, 1/10, 1/10, ], dtype='float') * gw8_VR
x.VRi = np.copy(x.VR)
x.VRrL, x.VRrH =0.5, 1.5 # randomize ratio
# 4 [VE] Task Eng # # VE contains energy cost per unit of computation
gw8_VE = 1
ar_VE = np.array ([0.56, 0.52, 0.52, 0.52, 0.52, 0.52, 0.55, 0.55, ], dtype='float')
w8_VE = np.array ([1, 1, 1, 1, 1, 1, 1, 1, ], dtype='float') * gw8_VE
x.VE = np.array(np.multiply(ar_VE, w8_VE))
x.VEi = np.copy(x.VE)
x.VErL, x.VErH = 0.5, 1.5 # randomize ratio
#-=============================================
return x
#-=============================================
def infra_4():
x=INFRA()
x.E = 8 # no of Edge servers
x.C = 3 # no of Cloud servers
x.A = x.E + x.C +1 # action space
# 1 [DR] Data Rate(delay) # DR contains delay per unit of data
gw8_DR = 1
x.DR = effective_bandwidth(np.array([
# #i0 e1 e2 e3 e4 e5 e6 e7 e8 c9 c10 c11 #
[ 0, 5, 5, 5, 5, 5, 5, 5, 5, 0, 0, 0, ], # i0
[ 0, 0, 300, 0, 0, 0, 0, 0, 0, 300, 0, 0, ], # e1
[ 0, 0, 0, 300, 0, 0, 0, 0, 0, 300, 0, 0, ], # e2
[ 0, 0, 0, 0, 300, 0, 0, 0, 0, 300, 0, 0, ], # e3
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 300, 0, ], # e4
[ 0, 0, 0, 0, 0, 0, 300, 0, 0, 0, 300, 300, ], # e5
[ 0, 0, 0, 0, 0, 0, 0, 300, 0, 0, 0, 300, ], # e6
[ 0, 0, 0, 0, 0, 0, 0, 0, 300, 0, 0, 300, ], # e7
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 300, ], # e8
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 500, 0, ], # c9
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 500, ], # c10
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ], # c11
], dtype='float') ) * gw8_DR
x.DRi = np.copy(x.DR)
x.DRrL, x.DRrH =0.5, 1.5
# 2 [DE] Data Eng # DE contains energy cost per unit of data
gw8_DE = 1
ar_DE = np.array ([0.56, 0.522, 0.521, 0.522, 0.521, 0.522, 0.522, 0.521, 0.522, 0.55, 0.55, 0.55, ], dtype='float')
w8_DE = np.array ([ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ], dtype='float') * gw8_DE
x.DE = np.array(np.multiply(ar_DE, w8_DE))
x.DEi = np.copy(x.DE)
x.DErL, x.DErH = 0.5, 1.5
# 3 [VR] Task Rate(delay) # VR contains delay per unit of computation
gw8_VR = 1
x.VR = np.array([ 1/4.5, 1/5, 1/5.05, 1/5, 1/5.05, 1/5.05, 1/5, 1/5.05, 1/5, 1/10, 1/10, 1/10, ], dtype='float') * gw8_VR
x.VRi = np.copy(x.VR)
x.VRrL, x.VRrH = 0.5, 1.5 # randomize ratio
# 4 [VE] Task Eng # # VE contains energy cost per unit of computation
gw8_VE = 1
ar_VE = np.array ([0.56, 0.52, 0.52, 0.52, 0.52, 0.52, 0.52, 0.52, 0.52, 0.55, 0.55, 0.55, ], dtype='float')
w8_VE = np.array ([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ], dtype='float') * gw8_VE
x.VE = np.array(np.multiply(ar_VE, w8_VE))
x.VEi = np.copy(x.VE)
x.VErL, x.VErH = 0.5, 1.5 # randomize ratio
#-=============================================
return x
#-=============================================