Limit the maximum block size to 256 KB

[erikzhang]

In order to prevent the block data from growing too fast, it is necessary to limit the maximum size of each block.

[kevaundray]

How did you decide on 256KB?

[wy]

Is there currently a limit? How do you pick which tx's to drop (based on the priority sorting order) or random?

[erikzhang]

There is currently no limit.

[wy]

To help with the discussion @erikzhang, @decentralisedkev, I looked at 287 thousand blocks in recent history (June-Aug 2018). Of them, there is only 15 above 256KB (262144 bytes). That's less than 0.0052% of blocks.

[f27d]

And it's 257 above 128KB (131,072 bytes), that's 0.089% for the same dataset.
Hope that helps..

[vncoelho]

Good data and precise studies.

[jsolman]

• Would this be necessary if there were a workable type of throttling for bad actor addresses?
• As normal usage increases and hardware of nodes and networking improve, shouldn't this value need to increase also?

I added a comment on the end of issue #357 concerning a potential solution to throttle bad actors for which I've started drafting a NEP.

[vncoelho]

@erikzhang, do you still believe that this is necessary?

Maybe with the correct Economic model and the refined verification+application layers we are not going to need to have this limit.

However, since blocks can be faster this can also be a good practice for the P2P decentralization.

[erikzhang]

I think we can set the maximum block size in PolicyContract.

[vncoelho]

Ok, @erikzhang, so, still proceed with this idea, right?
Remembers me BTC, BCH. aheuaheuahea In our case, from BCH to BTC...

@shargon, @igormcoelho, @jsolman, limit or not limit, this is the question...aheuahuea

[shargon]

We must implement the max, is totally needed.