Bitcoin Fork Fails: Miners Refuse BIP-110 Upgrade, Halting Protocol Progress

A recent attempt to activate Bitcoin Improvement Proposal (BIP)-110 culminated in a minor, yet significant, chain split over the weekend, demonstrating Bitcoin's resilience against proposals lacking broad consensus. The split occurred on Saturday at block 961,632, precisely when BIP-110 entered its mandatory signaling window. Nodes running the BIP-110 software were programmed to reject any block that did not signal support via version bit 4. Consequently, when the first block at that height failed to include the required signal, these enforcing nodes branched off onto a separate chain.

This newly formed chain, however, quickly stalled. A pseudonymous mining collective known as Roughnecks was responsible for producing the only two blocks on the BIP-110 fork — blocks 961,632 and 961,633 — utilizing Ocean’s DATUM protocol, before ceasing their activity. By Sunday afternoon, the enforcing branch remained stuck at block 961,633, while Bitcoin's dominant chain had advanced significantly to block 961,744. This created a substantial gap of 111 blocks, representing approximately 17 hours during which no new blocks were added to the BIP-110 fork.

The core issue behind the failure was a stark lack of support for the proposal. In the difficulty period immediately preceding the split, only 51 out of 2,016 blocks signaled for BIP-110, amounting to a mere 2.53% of the network's hashpower. This figure fell far short of the 55% threshold required for voluntary lock-in. Furthermore, once the mandatory signaling window opened, none of the first 113 blocks on the dominant chain signaled their support for the proposal, cementing its lack of acceptance.

Compounding the problems for the stalled fork was Bitcoin’s inherent difficulty adjustment mechanism. The BIP-110 fork inherited the main chain’s mining difficulty, yet it commanded a negligible share of the overall hashpower. Under these conditions, the fork could not adjust its mining difficulty downwards until it completed a full 2,016-block period. Estimates for how long this would take ranged widely from under a year to several decades, depending on the assumed hashrate, effectively rendering the fork impractical and unsustainable.

BIP-110, formally known as the Reduced Data Temporary Softfork, aimed to cap arbitrary data within transactions for about a year. Its primary targets were Ordinals inscriptions and oversized OP_RETURN payloads, which have been a subject of debate within the Bitcoin community regarding their impact on network efficiency and decentralization. Prominent figures such as Michael Saylor and Blockstream’s Adam Back publicly voiced their opposition, not necessarily against the goal of the proposal, but rather against its chosen activation method, which they argued was prone to precisely this outcome.

Michael Saylor, a vocal proponent of Bitcoin, commented on Sunday that the network had functioned exactly as designed, with 99.85% of hashpower remaining steadfastly on the main chain. In light of the confusion, mining company Ocean, whose team had supported BIP-110, took steps to address its clients. They informed miners using Ocean’s Stratum templates that some might have inadvertently directed their hashrate to the new BIP-110 chain, believing they were mining on the main Bitcoin network. Ocean pledged to reimburse these affected miners for the rewards they would have earned on the non-BIP-110 chain during the period of misdirection, acknowledging the operational disruption caused by the split attempt.