Add documentation for our calculation of a _reasonable_ gas fee for storage

deployment/scripts/chainspec/beta_chainspec.json

@@ -64,7 +64,7 @@
     },
     "fee_params": {
       "gas_price_factor": 92,
-      "gas_per_byte": 4
+      "gas_per_byte": 63
     },
     "chain_id": 0,
     "gas_costs": {

deployment/scripts/chainspec/dev_chainspec.json

@@ -34,7 +34,7 @@
     },
     "fee_params": {
       "gas_price_factor": 1000000000,
-      "gas_per_byte": 4
+      "gas_per_byte": 63
     },
     "chain_id": 0,
     "gas_costs": {

docs/fee_calculations.md

@@ -0,0 +1,65 @@
+## Fee calculations
+
+### Storage fees
+
+Fuel doesn't currently support separate fees for execution costs vs storage costs.
+This is a planned future feature, but in the mean time we need a system for
+compensating the network for storing contract data. The temporary solution is
+to include an additional cost to op codes that write new data to storage or to
+transactions that add new contracts to the chain.
+
+There are a number of ways we might calculate this value; we have decided to go 
+with a simple calculatoin based on the our target storage growth and working
+backward from there.
+
+#### Pessimistic Estimate
+
+If we arbitrarily set the target growth to 500 GB/year, we can ask, "what would 
+be the gas price if users maxed out each block gas limit with storage?" 
+This gives us this graph:
+
+| bytes per year  | block gas limit | blocks per year | bytes per block | gas per bytes |
+| --------------- | --------------- | --------------- | --------------- | ------------- |
+| 500,000,000,000 |      10,000,000 |        31536000 |      **15,855** |    **630.72** |
+
+This is a harsh estimate that isn't taking into the additional base cost of tx
+execution and the cost of any additional op codes. It is also assuming that 
+all blocks would be maxing out the storage.
+
+#### Generous Estimate
+
+Additionally, this will only apply to our early networks, which won't be long-live.
+This allows us to take a bigger risk on the storage price and increas it over 
+time to compensate for users adding a lot of data.
+
+All this included, if we re-estimate the yearly storage limit as 5 TB we get:
+
+| bytes per year    | block gas limit | blocks per year | bytes per block | gas per bytes |
+| ----------------- | --------------- | --------------- | --------------- | ------------- |
+| 5,000,000,000,000 |      10,000,000 |        31536000 |     **158,549** |     **63.07** |
+
+Giving us a cost per byte of **63 gas** as our initial gas price for storage.
+
+#### Additional justifications
+
+If we compare our numbers to what Ethereum charges for gas, we get similar numbers.
+
+If we had a pessimistic estimate for Ethereum and every block on Ethereum was full of new contract creations:
+
+max_contract_size = 24,000 bytes
+gas_per_max_contract = 32,000 + max_contract_size * 200 = 4,832,000
+
+gas_per_block = 30,000,000
+contracts_per_block = gas_per_block / gas_per_max_contract = ~6
+
+blocks_per_year = 365 * 24 * 60 * ~5 = ~2628000
+
+yearly_new_contract_bytes = blocks_per_year * contracts_per_block * max_contract_size = **378,432,000,000**
+
+Which rougly lines up with our pessimistic estimate.
+
+#### Conclusion
+
+We'd recommend starting at something like **63 gas/byte**. If this is underpriced,
+this will be tested in a testnet before being launched, hopefully giving us 
+ample time to gather more data and update the price