According to Threefold the Commercial value of a ThreeFold Token (TFT)
Commercial value of 1 TFT = commercial_value_of_grid (over 5 years) / nr_of_liquid_tokens (available today)
At time of writing this is about USD $0.35
How do we measure the commercial value?
each node is being registered in the TF Directory
the total resources available are measured by Zero-OS
the resources are measured by means of [resource_units]
The following formula's are used to calculate from resource units to [cloud_units].
1 CU = min(MRU/4*(1-5%),CRU*2)
1 SU = HRU / 1093 + SRU / 137
1 NU = CU x 10 + SU x 2% x 1000
each cloudunit has a commercial value at time of writing (jan 2019) we use
USD 15$ for a compute unit
USD 12$ for a storage unit
reasoning
We use as basis the definitions in [cloud_units].
compute unit = CU
1 std CU is 4GB of mem and we take 5% buffer but we can never oversubscribe more than 4 times. 1 std CU has 2 virtual CPUs thanks to min CRUx2 we make sure that oversubcription is max of 4.
We need to take min because the most conservative measurement needs to be used.
storage unit = SU
From experience we know the required resources that allow us to deliver the specs as defined in the definitions. It's the combination of SSD & HD capacity.
The detailed calculations are described below.
HARD DISK CAPACITY
= (archivecapacity x 70% + nascapacity x 30%) / redundancy_factor = (1000 x 70% + 400 x 30%) / 0.75 = 1093
0.75 is the redundancy factor, which means we take 25% overhead for redundancy
we take 70% of archive capacity in the field
we take 30% of nas capacity in the field
SSD capacity
= (( SSD_disk x 75% + DB_disk x 5%) / redundancy_factor + temp_disk x 20%) x free_space_facttor = (( 50 x 75% + 5 x 5%) / 0.5 + 80 x 20%) x 1.5 = 137
For SSD we take 75% for std purposes, 5% is for database & 20% for temp space. The temp space is not redundant so there is no redundancy factor. It's on SSD so the redundancy factor is 2x, we copy each block 2x. We take a free space factor of 1.5 which means 33% of capacity is free on the SSD and will not be used by customers.
Network Unit
This is the most tricky one because we don't know the relation between network bandwidth requirements and storage or compute workloads. The NU's are a result of the available CUs and SUs.
We did best effort estimates in this phase, this will improve as we get more data.
compute_units x 20 + storage_units x 2% x 1000
estimation
We estimate that averaged out each compute unit will require 20 GB of transfered data per month (which is conservative).
We estimate that averaged out each storage unit will require 2% of its capacity transfered per month
