Global Gas Price

Each miner node's PoW solution comes with a proposed minimum transaction processing gas price that the node is willing to accept. During DS block consensus, the DS committee runs an algorithm to compute the globally acceptable minimum gas price that the entire network will operate on. The DS committee then informs the shards, lookups, and seeds (through m_gasPrice in the DS block) about the agreed upon global minimum gas price. The network will accept any transaction that has a gas price larger than or equal to the minimum gas price and reject transactions that don't.

The algorithm to compute the global minimum gas price takes into account:

the average m_gasPrice used over the last n DS epochs
the average of the gas prices proposed by each individual miner for the next DS epoch (in case a price increase is necessary)
the trend in network congestion (i.e., the actual consumed gas in the Tx blocks) in the last DS epoch

Essentially, the algorithm decides on the gas price depending on the level of network congestion: if network congestion is high, then the miners get to have a say on the gas price; otherwise, the minimum gas price should not depend too much on their proposed gas prices.

Algorithm Inputs

Global Parameter	Description
`microblock_gas_limit`	Gas limit for each micro block
`num_shards`	Number of shards in the network, including the DS committee
`txblock_gas_limit`	Gas limit for the Tx block (`num_shards x microblock_gas_limit` for simplicity's sake, although the DS committee has a different gas limit value from the shards)
`default_min_gas_price`	The lowest value that the gas price can take

Data from Current DS Epoch	Description
`num_nodes`	Number of nodes in the network (including the DS committee)
`proposed_min_price_node[i]`	Minimum gas price proposed by the i th miner node

Data from Last n DS Epochs	Description
`min_price_epoch[j]`	Global minimum gas price value used in the j th DS epoch
`consumed_gas_tx_block[j][k]`	Total consumed gas in the k th Tx block in the j th DS epoch

Algorithm Steps

In the following, we describe an algorithm to compute global_gas_price for the next DS epoch.

We first compute percentage_full_tx_blocks, i.e., the number of Tx blocks mined in the last DS epoch for which the total gas consumed is at least 80% of txblock_gas_limit. This computation will require checking each consumed_gas_tx_block[j][k].
Then, we make the decision on how to set global_gas_price according to the table below.

Percentage of Full Blocks	Interpretation	Impact on Gas Price
< 10%	Blocks are mostly empty; demand is low and hence miners should accept a lower price	Increase in the gas price compared to the previous epoch
Between 10% and 70%	Blocks are mostly filled; demand and supply have met the sweet spot	No change in the gas price compared to the previous epoch
> 70%	Blocks are heavily filled and hence there is a sign of high demand	Decrease in the gas price compared to the previous epoch

Decreasing the Gas Price

Discard all proposed_min_price_node[i] values
Compute the average gas price over the last n DS epochs:

average_gas_price_val = mean(min_price_epoch[j-1], ..., min_price_epoch[j-n])

Compute the decreased gas price value:

decreased_gas_price_val = 99% of average_gas_price_val

Finally, compute the new global_gas_price:

global_gas_price = min_price_epoch[j] =
   max(default_min_gas_price, decreased_gas_price_val)

Increasing the Gas Price

Get the median proposed_min_price_node[i] value over all values from the N miners:

median_proposed_min_price = median(proposed_min_price_node[1], ..., proposed_min_price_node[N])

Compute the average gas price over the last n DS epochs:

average_gas_price_val = mean(min_price_epoch[j-1], ..., min_price_epoch[j-n])

Compute the lower and upper bounds for the increased gas price value:

increased_gas_price_val_lower_bound = 100.5% of average_gas_price_val
increased_gas_price_val_upper_bound = 101.5% of average_gas_price_val

Finally, compute the new global_gas_price:

global_gas_price = min_price_epoch[j] =
   max{
      max[
           min(median_proposed_min_price, increased_gas_price_val_upper_bound),
           increased_gas_price_val_lower_bound
         ],
      default_min_gas_price
   }