IBFT Parameters

Command line options

Block period

--istanbul.blockperiod 1

Setting the block period is used for how long blocks should be minted by the validators. It is also used for validation of block times by all nodes, so should not be changed after deciding a value for the network. The setting is a positive integer, and measures the minimum numbers of seconds before the next block is considered valid.

The default value is 1.

Request timeout

--istanbul.requesttimeout 10000

The request timeout is the timeout at which IBFT will seek to trigger a new round if the previous one did not complete. This period increases are the timeout is hit more often. This parameter sets the minimum timeout in the case of normal operation and is measured in milliseconds.

The default value is 10000.

Genesis file options

Within the genesis.json file, there is an area for IBFT specific configuration, much like a Clique network configuration.

The options are as follows:

{
    "config": {
        "istanbul": {
            "epoch": 30000,
            "policy": 0,
            "ceil2Nby3Block": 0
        },
        ...
    },
    ...
}

Epoch

The epoch specifies the number of blocks that should pass before pending validator votes are reset. When the blocknumber%EPOCH == 0, the votes are reset in order to prevent a single vote from becoming stale. If the existing vote was still due to take place, then it must be resubmitted, along with all its votes.

Policy

The policy refers to the proposer selection policy, which is either ROUND_ROBIN or STICKY.

A value of 0 denotes a ROUND_ROBIN policy, where the next expected proposer is the next in queue. Once a proposer has submitted a valid block, they join the back of the queue and must wait their turn again.

A value of 1 denotes a STICKY proposer policy, where a single proposer is selected to mint blocks and does so until such a time as they go offline or are otherwise unreachable.

ceil2Nby3Block

The ceil2Nby3Block sets the block number from which to use an updated formula for calculating the number of faulty nodes. This was introduced to enable existing network the ability to upgrade at a point in the future of the network, as it is incompatible with the existing formula. For new networks, it is recommended to set this value to 0 to use the updated formula immediately.

To update this value, the same process can be followed as other hard-forks.