Fees trilemma

People like trilemmas :slight_smile: Here is another one.

With increasing complexity of the blockchain the following are getting increasingly difficult to satisfy simultaneously:

  • Fees are not severely overestimated โ€“ the more complex the blockchain operations are the larger is the gap between the worst and the average scenario, and since we need to charge based on the worst case to avoid the adversarial abuse, the average case is getting overcharged;
  • Fees are not abusable โ€“ the more complex the operations are the more their fees become functions of multiple context-dependent variables, e.g. the depth of the MPT;
  • Fees have good DevX โ€“ we can make fees non-overestimated and non-abusable, but then they will likely have a complex behavior from the contract development perspective, because they would depend on a lot of context that developer wouldnโ€™t want to care about.

Right now NEAR is leaning towards allowing fees to be severely overestimated, since we donโ€™t want to sacrifice safety or DevX.

6 Likes

Do you have an example to understand the scale of โ€œseverelyโ€?

1 Like

For some fees it is 3x, but we are doing it intentionally right now, to stay on the safe side. With the time, as we are spending more resources on tuning our fee estimation (and incidentally the blockchain performance) this will become less severe.

2 Likes