When blocks get full under the current protocol parameters, my block transmission delays already bump over the 1 second cliff.
In order of block size:
Size Transactions Propagation delay
71030 28 1.10s
55518 30 0.97s
24511 6 0.79s
7929 5 0.62s
Any increase in block size will presumably make this situation worse. Without protocol changes like I suggested at Subpar handling of blocks from the future · Issue #4251 · IntersectMBO/ouroboros-network · GitHub the only solution is to centralise my block producer by moving it to an Amazon data centre in EU or USA. If I don’t do that when my propagation delays are over 1 second then my pool will suffer 3 times the number of “fork battles”.
The world is not just the EU and USA. To be truly decentralised we also need to be geographically decentralised. Possibly even space decentralised, with block producers in satellites. We can’t propagate larger blocks to all these areas within 1 second. We need to make some minor protocol changes so that those that can’t stay under the 1 second barrier are not unfairly disadvantaged. We need to increase this barrier to 3-4 seconds.
The changes I suggested should be discussed with IOG technicians prior to any increase in block size.