EVM: Reboot scheduling

To make sure refutation is always possible, the kernel should have a reboot strategy to ensure it never go over the limit of tick/reboot.

Pb:

• max amount of tick per reboot is set
  • so the kernel need to know when to reboot
• max nb of reboot is set
  • so the kernel needs to know when to stop at a given level, and delay transaction to next level
• if transaction are delayed too much, then we have a liveness pb
  • a way to prevent that is to have a moving gas price, akin to elastic blocks on ethereum: if the blocks are full, gas price increase, until nobody wants to send new transactions, and the backlog is resorbed.

Remark: not all problems have the same priority

• The first point is necessary for refutability
• The second point is important for censorship ?
• The third is important for liveness