Burning as a Transaction Fee

PlasmaPower mentioned the idea of using a burn to replace PoW. I think it's reasonable to consider a dynamic burn amount that increases with network traffic similar to Dynamic PoW. This burn would be required for sends only.

Pros:

  • Burns would act as a fee denominated in Nano rather than a fee denominated in electricity. This removes PoW generation time & the gaming of PoW processing speed and efficiency.
  • Burns would subsidize spamming's effect on the network. Burns increase scarcity and relative value of all other accounts. Assuming node operators hold Nano, the burns by spammers would help offset the increased hardware costs required to handle ledger bloat.
  • I believe this will get around the prioritization problem, as a burn amount should be easy to prioritize fairly.

Cons:

  • As a burn is a fee, it is a significant departure in terms of user experience.

Additional thoughts:

  • Instead of distributed PoW donating GPU processing, a service for donating "Burns" could step in to help provide fill in the UX gap for trusted wallets. I'm not sure how this could be implemented.
  • Subtracting the burn from the sender's remaining balance rather than from the transaction itself could be a slight UX improvement.
3 Likes

Having fees in the protocol is such a fundamental bad idea. Imagine having fees on every TCP or UDP packet. The Internet wouldn't be what the Internet is. I am not against making the network more resilient. Even though there are fundamental design attributes which have to be like they are. Evey node owner can also have local protective measures, like every ISP does. Cryptocurrencies should learn from the design of the Internet. Yes, it has it flaws, but overall it has a decent resilient design. To unlock the full potential of cryptocurrencies, it has to be feeless.

8 Likes

This idea has been around since the beginning but it also misses the more fundamental point of how the network determines the upper TPS limit. If the TPS limit isn't being hit, congestion throttling won't have an effect. A spammer will just issue as many free transactions as they can under the TPS limit.

Lots of chains pick this in a centralized way i.e. the block size, nano the TPS limit is implicitly set in a decentralized way by the bandwidth limits on PRs.

4 Likes

I think this argument is flawed. PoW is a fee denominated in electricity (and whatever currency is used to pay for that electricity) and time, which also requires access to the hardware to generate it. A burn serves as a fee denominated in Nano which has many pros for the network, which I have outlined.

The true con as I see it, and as I pointed out, is the user experience.

1 Like

There could be a minimum burn required for all send transactions.

How does this scale with value fluctuations?

I imagine it would scale similar to DPoW.
I worked out this algorithm to determine what the burn should be if a simple dynamic burn wasn't considered appropriate:

dpob / (1 + total_burnt) / log(prior_balance)^2 * max(prior_balance / send_amount, 1 / dpob)^(1/(1+log(prior_balance)))

It's meant to penalize transactions that are small relative to the account balance. It also penalizes very small accounts so they can't send dust either.
dpob is "dynamic proof of burn" that rises with congestion. total_burnt is the sum of all burnt funds (so the burn rate decreases over time, and honestly i'd probably give this value some exponent instead of it being linear). prior_balance is the account balance before initiating the send.

DPoW doesn't change the minimum pow threshold.

Ah, then raising dpow/dpob would increase user's priority level? Still, a minimum PoB similar to minimum PoW could be required, right?

Technically yes.

But then how does that scale with value fluctuations?

To clarify, what value fluctuation? If you're referring to the fiat valuation of Nano at any given time, I don't think it should have any relation. The fee should be entirely denominated in Nano. I'll now contradict myself: if the concern is pricing out small players as the price of Nano increases, then I agree that is a concern and I am not sure how to tackle that problem other than to make amendments to the burn amount in the future.

This is why the equihash and memory-hard PoW discussion is so important. You're statement is correct but it seems to imply that electricity cost is the more significant "fee". I'd argue that with the right algorithm, "time" is the bigger spam deterrent. Fortunately most honest users have significant "time" they can spare between transactions.

Any sort of proof of burn or $$$ fee approach takes this natural advantage away from honest users (particularly users in poor countries).

2 Likes