Dust Expiry: Clean the UTXO set from spam

The discussion revolves around the implementation of a simple absolute limit to manage the Unspent Transaction Output (UTXO) set size in blockchain networks. This approach aims at excluding coins above a certain value from being subjected to an accumulator, thereby establishing a fixed upper limit for the non-accumulated UTXO set. This method is likened to introducing a new form of dust limit, which might lead wallets to treat values below this threshold as unspendable, unintentionally confiscating small balances.

The proposal also touches on the potential for abuse, illustrating how setting a low threshold could still result in significant storage requirements under attack scenarios. For example, even with a threshold as high as 20,000 satoshis, an attacker could inflate the UTXO set to consume up to 7 terabytes of disk space. The inclusion of a budget constraint is suggested as a mitigation technique, where discarding transactions under a minimal value could significantly increase the cost of carrying out such an attack, thus serving as a deterrent.

Furthermore, the implications of this system on Layer 2 (L2) protocols are considered, especially concerning low-value UTXOs that might be used briefly within these systems. It's suggested that moving such UTXOs to the accumulator only if they fall below a specific value and age could mitigate complications, supporting the functionality of L2 designs without imposing undue complexity.

Lastly, the idea of hybridizing the approach to UTXO management is explored. This could involve combining different types of proofs for coin identity and spentness over time to streamline the process and potentially disincentivize UTXO spam attacks. However, there's an acknowledgment that easing the restrictions on UTXO growth might inadvertently encourage the creation of more dust outputs, as they would no longer pose a burden on nodes. This could lead to their increased use in L2 constructions for contingency purposes, ultimately resulting in a larger number of UTXOs.