[Protocol] 3 seconds dbft consensus

[Jim8y]

This pr tries to formalize our so called 3-seconds consensus.

It requires hardfork and:

• reduce the block time from 15 seconds to 3 seconds
• reduce the gasperblock from 5 GAS to 1 GAS

Affected Configurations:

• MillisecondsPerBlock: 15000 => 3000
• MaxTransactionsPerBlock: 512 => 256
• GasPerBlock:
  • Current: 5 GAS
  • At hardfork: Automatically reduced to 1 GAS
  • Post-hardfork: Configurable by council
• MaxValidUntilBlockIncrement: 100 => 500?

[roman-khimov]

Isn't it a policy change, rather than a protocol change? Do we track policy changes in NEPs?

[roman-khimov]

Maybe

An Informational NEP describes a NEO design issue, or provides general guidelines or information to the NEO community, but does not propose a new feature. Informational NEPs do not necessarily represent a NEO community consensus or recommendation, so users and implementors are free to ignore Informational NEPs or follow their advice.

But then this means that any mainnet policy changes (like fees) should be described in NEPs. Which would be nice, but are we ready for it?

[AnnaShaleva]

Which would be nice, but are we ready for it?

It seems that we're not yet ready. We'd better start with NEPs for protocol changes only, because otherwise we'll end up in a very large amount of pending NEPs describing all Neo policies that need to be created/reviewed/approved.

So vote up for not including this change into NEPs.

[cschuchardt88]

PolicyContract doesn't hold this information. It's a configuration in a file. How else you going to change either setting for privatenet?

[Wi1l-B0t]

Is it necessary to add a method to get block time?

[superboyiii]

Yes, we have set then we should have get.

[superboyiii]

Suggested change

	===Events===
	====SetBlockGenTime====
	<pre>
	{
	"name": "SetBlockGenTime",
	"parameters": [
	{
	"name": "milliseconds",
	"type": "Integer"
	}
	]
	}
	</pre>

We'd better add an eventlog in the end of the method.