Antpool, bitcoin’s biggest pool with around 14% of the network’s hardware share, began mining yesterday with the new grassroots client, Bitcoin Unlimited.
Only the Beijing servers appear to have upgraded, finding the first block around 7PM London time, with six total blocks found at the time of publishing. That gives BU’s share of Antpool hardware around 7% of the total bitcoin network.
It is unclear whether the other pool servers will follow as Jihan Wu, Antpool’s co-founder, has not responded in time for publishing, but it appears unlikely at this point in time.
The main reason is probably because a switch to Bitcoin Unlimited is a network wide upgrade, requiring caution and a gradual roll out to give businesses and node operators time to upgrade as it would be the first ever upgrade without any leadership guidance, undertaken solely by direct mining choice.
The public reaction appears positive overall, even in highly censored public bitcoin forums. Somewhat surprisingly, price did not react either way, moving sideways for more than eight hours. It then fell by $100 in what appears to be an uncorrelated move, but quickly recovered, now standing around $1,230.
The lack of price reaction to what is fairly big news regarding scalability appears interesting because it may suggest the market has priced in an upgrade. Like the shut-down of the US government over debt ceiling arguments, the market may consider a resolution as highly probable. Thus, seemingly, not pricing in transaction delays or increased fees, nor segwit or BU moves - in most cases in any event - so probably considering an increase in capacity as unavoidable and just a question of time.
Proof of Nodes?
Regarding timing, the Antpool move was made after suggestions that segwit could be implemented by a "flag day activation." That is, at a certain block upgraded nodes enforce the new rules.
This is a useful method for hardforks when they are uncontroversial or when it appears likely that around 70-80% of the overall community supports the new rules, but this method has never been tried as a soft-fork because in a soft-fork non-upgraded nodes do not actually have a choice.
In a hard-fork, non-upgraded nodes reject the new rules and cannot operate on the upgraded node chain. Thus, if there are miners on the non-upgrade node chain, they create a clean split.
In a soft-fork, non-upgraded nodes ignore the new rules while continuing to operate on the chain of upgraded nodes. They, therefore, have no say on the new rules that are nonetheless applied, in effect, by the entire network, even if they strenuously disagree with the new rules - which can include an increase of the 21-million-coin limit - with the only option being to create a new client.
As far as miners are concerned, there is no such thing as hard or soft fork. For miners, all forks are hard forks. We have seen this in practice many times when, during a soft-fork upgrade, accidental chain splits are created, requiring miners to quickly choose one fork.
However, in a flag day soft-fork, the miner’s role becomes far more complicated. Assuming the entire network is mining with one client, non-upgraded miners are accepted by non-upgraded nodes, upgraded miners are accepted by upgraded nodes and non-upgraded nodes, upgraded nodes reject non-upgraded miners.
This is a mess in itself, but add other clients, such as Bitcoin Unlimited, and we have utter chaos. Alternatively, if non-upgraded miners are forced by easily sybilable upgraded nodes to upgrade, then we have a dangerous centralized point of failure as developers – who are just volunteers with no real stake – are given all the power.
While miners have a complex incentive system, developers have no robust incentive system that prevents malicious actions. Moreover, technically, development is centralized as it is just one person, the maintainer, who ultimately decides what goes into the client.
It is, therefore, unnecessary to elaborate on the dangers of giving unaccountable and volunteer developers sole say, but the mere fact that it has been suggested and somewhat seriously considered in some corners has been sufficient to finally incentivize Antpool to mine with Bitcoin Unlimited.
That means Bitcoin Unlimited now stands at around 30% of the network’s hardware-share. The highest ever achieved by any alternative client. The decision now moves to F2Pool. If they too upgrade, then businesses will probably follow and this endless debate may finally end.
Image from Shutterstock.