The Hard Fork: Will Bitcoin XT Take?

By David Floyd

“So this is it. Here we are.” With that uninspired opening, Mike Hearn announced on August 15 a change to Bitcoin’s software that has the potential to generate a hard fork in the blockchain.

It would be his second. In March 2013, a tweak Hearn introduced to version 0.8 caused a discrepancy between miners running that version and the previous one. The fork was resolved quickly, so no harm done. Ironically, though it had to do with the size of blocks.

The difference between that fork and this one is that this time, it’s intentional. The first shot has been fired in a years-long standoff over what may seem like a finicky technical issue, but in fact has enormous implications for Bitcoin’s future: the maximum size of a block in the blockchain.

There’s also a less bellicose metaphor we can use, which supporters seem to prefer: the issue has been put to a vote, a sort of miners’ referendum, so that the bickering and trench-digging can finally stop.

The Block Size Controversy

If you’re unfamiliar with the fight over the maximum block size, this post explains the issue as it stood in mid-July. For a truly exhaustive review, BitcoinWiki’s “Scalability FAQ” leaves no stone unturned.

Briefly, however, the issue revolves around Bitcoin’s potential for scalability. At the moment, the crypto-currency remains relatively niche; it has an active and growing user-base, but by no means the kind of capital or usage a reserve currency enjoys. At $3.2 billion, its market cap is less than that of Dillard's (DDS) or Pandora (P).

If Bitcoin were ever to be adopted by even a couple billion people, the network would grind to a halt, since the current block size limit of 1MB allows for a theoretical maximum of seven transactions per second (in practice, the limit is even lower). People could wait years or decades for their transactions to process. Or, more likely, intermediaries would sidle in, demanding trust and becoming the very financial institutions Bitcoin exists to circumvent.

How to solve this issue has splintered the community. There are those who oppose any change to the block size: leave it as it is, nobody should be using Bitcoin to pay for every cup of coffee anyway. There are many who want to raise the limit, but the specifics of individual proposals vary: BIP 100, in which miners vote on a limit not exceeding 32MB; the Lightning Network, which would track bilateral transactions and only submit the net balance to the blockchain; etc.

But until now, a premium has been placed on near-complete consensus. While this has in effect meant inaction, the alternative could be worse: larger-than-1MB blocks will generate a hard fork, that is, a change that older versions of the Bitcoin software will not accept. A hard fork could be disastrous for the network’s security, primarily because it hamstrings safeguards against double-spending. Bitcoin held before the fork can be spent twice, once on each chain, and Bitcoin received after the fork will only be viable on one chain.

These chains could both continue to build separately, making one or the other in effect an “altcoin.” The longer this situation persists, the harder it will be to reconcile the two blockchains. Even in the best-case scenario, those on the wrong side would lose money, so attitudes would harden, and the division would calcify.

In the worst-case scenario, both pretenders to the throne would be perceived as altcoins, and the fragile germ of legitimacy Bitcoin has cultivated would fade. Things fall apart; the centre cannot hold; Mere government is loosed upon the world.

Mike and Gavin to the Rescue

That is what Mike Hearn, a former Google (GOOG) developer, and Gavin Adresen, one of five Bitcoin Core developers, are risking with Bitcoin XT. They rationalize their technical-political brinksmanship by insisting that there is no time: by the middle of next year, or maybe some time in 2017, the average block will be pushing the 1MB limit.

What happens then? According to some, market forces take over, driving transaction fees up and discouraging those who would spam the system with miniscule transactions. Equilibrium is restored, and fees, while higher, remain low compared to any other payment system.

According to both Adresen and Hearn, this scenario is delusional. Hearn’s take: because of the way the blockchain communicates with wallet apps—or rather, doesn’t—users will see that their transactions are stalled, but not why. Some users will resubmit them with a higher fee, but that may not work, because they will be bidding blind. Imagine an online brokerage that displays no price data, so you have to guess what the stock you want to purchase is worth, increasing your bid incrementally and exacerbating the system’s overload of transaction requests.

Nor, according to Hearn, can Bitcoin Core really handle a backlog. Nodes might simply crash. It wouldn’t be long before users moved on to some other asset, which is exactly what Hearn fears. “Quite simply,” he writes in another post, “we are out of time. There are no credible technical proposals that could gain widespread adoption within the next twelve months, beyond simply raising the capacity on the existing system.” So that’s what he and Andresen have decided to do.

Bitcoin XT allows the block size limit to rise to 8MB in January 2016 and then to double every two years until it reaches a cap of 8,192 MB. This process will not begin, however, unless 750 of 1,000 consecutive blocks are mined using Bitcoin XT. Once that threshold is reached, two weeks will be given for dissenting miners to upgrade.

Criticisms of Bitcoin XT

For many, 75% is much too small a majority to implement such a sweeping change. A group of miners accounting for a quarter of new blocks may have the critical mass to challenge the majority-Bitcoin’s legitimacy. Or they may attack it. In any case, trust—which ironically does matter when it comes to Bitcoin—has been breached.

Other worries, perhaps baseless, abound. There is concern that Tor users will be added to a hard-coded “blacklist” which may reveal their IP addresses; others have dismissed this claim as a fabrication or misunderstanding of the code. Credible or not, the privacy concern responds to a widespread suspicion that this update is rushed, sloppy and potentially detrimental to the Bitcoin project.

Nor are these divisions limited to the various forums where the fractious Bitcoin populace airs its grievances. Andresen is one of five core developers who manage Bitcoin Core’s upkeep. A second, Jeff Garzik, has urged a measured debate on the issue. On August 17 he tweeted: “This ‘contentious hard fork’ stuff is being blown way out of proportion. There is a lot more theoretical risk than practical risk.”

Others are less generous in their assessments. Gregory Maxwell has compared Hearn and Andresen to “a guy standing on the sidelines with a beer cup hat.” Nick Szabo likened the fork to the Challenger space shuttle disaster.

For all the emphasis on Bitcoin’s decentralization and its merits, a very small group of people in effect controls it. A schism within that small group is dangerous to the project’s future. Whichever side comes out on top, the divisions might not heal quickly, completely or at all, and the crypto-currency could lie in a yet-smaller number of hands.

Will It Take?

In any case, the cat’s out of the bag. The only question that matters now is, will Bitcoin XT get its 75% of 1,000 consecutive blocks or not? Will the hard fork in the blockchain come to pass?

The fourth Bitcoin XT block was recently mined—unless, that is, it wasn’t. Saboteurs have created a special version of Bitcoin’s software, Not Bitcoin XT, which seems in every way to be Hearn’s and Andresen’s brainchild, except that it doesn’t actually support larger blocks. So it may be that a portion of the 14.7% of miners who have seemingly embraced XT are in fact a sort of fifth column. Perhaps even one of the mined XT blocks (representing 0.04% of a needed 75%, to put it in perspective) is in fact Not.

At least two large payment processors, BitPay and Coinbase, have publicly supported XT. Someone claiming to be Satoshi Nakamoto has opposed it, although it seems unlikely that the founder has in fact emerged from his (?) Greater Occultation, and his email address is thought to have been hacked.

In the end, all that matters is the miners, and at the moment, no conclusions can be drawn. Adoption could clear 50% of miners easily enough, given a few more weeks, and maybe even break into a majority. The blocks are the only variable that matters, and that metric is lagging behind miners’ adoption by a mile.

And If It Does?

Bitcoin’s price, as measured by the CoinDesk BPI, is below $220 for the first time since April; between August 15 and 18, it lost 16% of its dollar value. The market is understandably nervous. After all, if the block chain forks, users could double-spend their Bitcoins in an attempt to sink the chain they don’t approve of. Someone would lose their money.

Even if users weather that storm, the long-term security of the network would be uncertain. Larger blocks are more bandwidth-intensive to mine, and the overhead could drive smaller operations out of business. The result could be fewer, larger miners.

Ghash.io, the largest pool, has already briefly touched 50% of mining power; with a majority, a malicious actor could double-spend coins, prevent transactions from being confirmed and prevent other miners from producing new blocks. Retroactive edits to the blockchain are a feared prospect, but are probably too difficult to accomplish in practical terms.

Conclusion

The release of Bitcoin XT has thrown the currency’s vulnerabilities into sharp relief. Far from being “decentralized,” it is dependent on a group of core committers you can count on one hand. That there is a rift in this group only raises the specter of an even smaller group consolidating control. The hard fork, the ultimate violation of trust, only demonstrates how important trust remains in the “trustless currency.”

And the precedent is dangerous: who’s to say intentional hard forks won’t be tried again and again, over ever-smaller issues. We’ve resorted to the nuclear option. How many forks can Bitcoin withstand before there is no main chain—just a plethora of altcoins calling themselves the real thing? Will large mining pools become ever more powerful and break the 51% threshold? Will vindictive purists double-spend Bitcoin into the ground? Will anyone care by the time this is over?

Anybody remember b-money?