An overview of the cluster mempool proposal

An overview of the cluster mempool proposal

Original Postby harding

Posted on: January 12, 2024 01:26 UTC

In the proposed approach to Replace-By-Fee (RBF) validation, a transaction is accepted into the mempool only if it results in an overall improvement, assessed through a comparison of feerate diagrams.

The process involves determining whether the cumulative fees of incoming transactions—up to a specific byte size—are greater than the accumulated fees at the corresponding point of the existing transactions in the mempool. A critical inquiry arises when transactions have different ending byte sizes, particularly when a new transaction with fewer bytes but higher fees per byte replaces an older, larger transaction with lower fees per byte. This raises concerns about the potential for RBF pinning to persist as an issue, especially when large transactions are replaced by much smaller ones, leading to a disproportionate increase in fees.

The conversation also touches on the Child Pays For Parent (CPFP) carve-out rule and its discontinuation. The sender questions the lack of straightforward alternatives to this rule, suggesting a possible solution where transactions that meet specific criteria could be flagged to bypass restrictions related to cluster size limits. These flags would prevent such transactions from being negatively impacted by cluster size constraints or from being evicted if the cluster exceeds its maximum size. Despite the sender's preference for removing special case rules like the carve-out, they express surprise over the absence of a simple workaround.