Proof of Storage Cost Analysis

	A	B	C	D
1	Input Parameters	Value	Unit	Notes
2	GPU Rate for mask() function	512	MB/s	Nvidia 3080 outputs 12GB/s for 128 bytes Ethash/hashimoto output in single round. We can add more rounds to lower the throughput (and thus physical storage will have more cost advantage)
3	GPU Power for mask() function	240	W	See above
4	GPU Lifetime	4	years	Typical GPU lifetime for Ethash mining
5	GPU Price	1000	$	3080 current price
6	CPU Rate	0.917562724	MB/s	AMD 5900 single core
7	Disk Rate	6144	MB/s	2x3GB/s for NVME SSD 980 Pro
8	Disk Power	14.4	W	2x7.2W for NVME SSD 980 Pro
9	Disk Price	400	$	2x2TB NVME SSD 980 Pro price
10	Disk Lifetime	5	years	5 years for manufature warrenty, but average is 10 years (since most IOs are reads)
11	Epoch Time	384	s	Beacon chain's RANDAO value
12	Total Random Sampling Data Size	262144	MB	In each epoch, bigger sample data will be more friendly to physical storage, but more power cost
13	kW Electric Hour Cost	0.068	$/kWh
14	Sample Size (On-Demand)	128	KB	16KB per BLOB * 16
15	Sample Size (Storage)	4	KB	4KB (IO) * 16
16	4TB GPU Sync Time	2.275555556	hours
17
18
19
20	Cost Estimation Per Random Sampling Solution
21	Storage Cost Per Solution
22	Disk Read Duration	42.6666666666667000	s
23	Solutions Per Epoch	67,108,864.00
24	Disk Power Per Solution	2.54E-12	kWh/solution
25	Disk Power Cost Per Solution	1.73E-13	$/solution
26	Disk Lifetime Cost Per Solution	1.45E-11	$/solution
27	Disk Total Cost Per Solution	1.47E-11	$/solution
28
29	On-Demand Computation Cost Per Solution			Raw data storage cost is not included
30	Solutions Per Sec	4,096.00
31	GPU Power	1.63E-08	kWh/solution
32	GPU Power Cost	1.11E-09	$/solution
33	GPU Lifetime Cost	1.94E-09	$/solution
34	GPU Total Cost	3.04E-09	$/solution
35
36	Cost Ratio of On-Demand Computation / Storage	207.1129129338050000
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