Bitcoin Core merged a significant architectural change on November 25, 2025, that fundamentally restructures how nodes organize pending transactions. Cluster Mempool, developed by Suhas Daftuar and Pieter Wuille, addresses long-standing inefficiencies in transaction management that have created challenges for Layer 2 protocol developers and mining operations.
Technical Architecture Changes
The mempool serves multiple critical functions: fee estimation, transaction replacement validation, and block construction. Prior to version 30.0, Bitcoin Core maintained two separate transaction orderings—descendant feerate (looking forward to child transactions) and ancestor feerate (looking backward to parent transactions). This dual-ranking system created misalignments between what nodes accepted and what miners would actually include in blocks.
Cluster Mempool introduces a unified approach by grouping related unconfirmed transactions into "clusters" and subdividing them into "chunks" sorted by feerate. This structure enables consistent behavior across all mempool operations:
- Block construction now simply selects the highest feerate chunks across all clusters
- Eviction logic removes lowest feerate chunks when the mempool reaches capacity
- Transaction replacement compares chunk feerates directly
The new system implements a global limit of 64 transactions and 101 kvB per cluster, a computational constraint that makes constant re-sorting practical while maintaining optimal performance.
Implications for Protocol Developers
For professionals building on Bitcoin, this change addresses critical reliability concerns. The previous architecture created uncertainty around whether transactions would propagate to miners, particularly when replacing unconfirmed transactions—a fundamental operation for Lightning Network penalty transactions and other Layer 2 enforcement mechanisms.
Malicious actors could previously exploit the inconsistent ranking logic to evict necessary ancestor transactions or block valid child transactions from entering the mempool. The unified cluster approach eliminates these attack vectors by aligning eviction, inclusion, and replacement logic with actual mining incentives.
Protocol developers can now build with greater confidence in predictable mempool behavior. Fee estimation becomes more reliable, and transaction confirmation times more consistent. This reduces pressure for private transaction relay channels to miners, supporting decentralized mining operations.
For blockchain engineers and protocol architects working on second-layer solutions, understanding cluster mempool mechanics will become essential as the upgrade propagates across the network. The change represents years of development work now available in production Bitcoin Core releases.
