Multidimensional EIP 1559

I think this would be a reasonable suggestion although I imagine it would be a significant engineering undertaking.

Would it be compatible with a non-linear basefee escalation model? I can’t find the exact piece I remember seeing (I don’t think I’m thinking of your post along the same lines), that proposed that fixed 12.5% steps up and 12.5% steps down, although effective, were not optimal, and that an algorithm that could take dynamic steps up and down could handle a burst and return to its normal load faster. My napkin math suggests the basefee can currently rise at most 80% in a minute (1.125^5 \approx 1.8...) and takes as long to come back down.

My question is whether dynamism in the escalator algorithm(s) would be incompatible with having many distinct escalation algorithms.

If I had to guess, I would assume many distinct escalator algorithms, both static and dynamic are feasible in theory, but I’m not confident it wouldn’t make the equilibria too unpredictable, and whether we could cause unintended consequences to the stability of the system from different dynamic systems all conflicting with one another in a manner that might end up hurting the smooth inclusion instead of facilitating it. e.g. a spike in one resource triggering a larger spike in another resource through follow on effects.

3 Likes