Nowadays, two types of forks are known and widespread, a soft fork and a hard fork. The former are considered less destructive and have backward compatibility, that is, they are compatible with the old version of the blockchain. When carrying out a soft fork, updated nodes can interact with the nodes of the "parent" network. The example of a soft fork is SegWit in the Bitcoin network. In the case of a hard fork, a new chain of the network is formed according to the rules that are incompatible with the old protocol of the blockchain. In this case, the nodes of the new and the old networks cannot interact since hard forks involve changes in the mechanism of reaching a consensus itself. In this case, the hard fork can split the blockchain, if not all users agree to make updates, as it happened with Ethereum Classic. The Ethereum hard fork was originally intended to recover assets stolen during an attack, rather than create two versions of the same cryptocurrency.
Velvet forks were described for the first time in the research work of Non-Interactive-Proofs-of-Proof-of-Work (NIPoPoWs) on how to reach a consensus and published on December 4, 2017, by the staff of leading universities in Edinburgh, Illinois, and Athens, Aggelos Kiayias, Andrew Miller, and Dionysis Zindros. They represent velvet forks as a mechanism that does not require "a change in the rules at the consensus level," so it does not need approval by all network members. In addition, the document indicates that such forks allow for "consistent deployment" of the software and do not harm miners who did not support the update, and in this sense, they have similar features with soft forks.
"We think the most interesting is that you can introduce new principles into an inclusive blockchain system, which does not require the support of the majority of network members," said Alexei Zamyatin, a research assistant at the Centre for Cryptocurrency Research and Engineering at the Imperial College London, who, at the beginning of March, at a conference on financial cryptography in Curaçao, presented a study entirely focused on velvet forks called "A Wild Velvet Fork Appears! Inclusive Blockchain Protocol Changes in Practice."
According to Zamyatin, "This is not rocket science, but rather a simple concept," and the system itself existed before, but was not described. Zamyatin gives the decentralized mining pool P2Pool as an example. Since there is no tool that controls the distribution of rewards to the miners, the P2Pool created the second blockchain with a lower level of complexity, which only the miners of this pool can support. This blockchain is used to estimate how much processing power each miner provides and pay a proportional reward. At the same time, all miners can accept both blocks created in the Bitcoin blockchain, as well as blocks of their own blockchain on P2Pool.
Zamyatin also describes how such forks can breathe new life into previously proposed but unimplemented innovations, particularly the Bitcoin-NG "next-generation Bitcoin" protocol developed by Cornell University professor Emin Gün Sirer to increase Bitcoin blockchain's throughput with his restructuring. "Although the document is not very detailed, the very idea of adding new functionality in the absence of risks and difficulties, which are usually associated with soft forks and hard forks, is very interesting," commented Sirer.
In his study, Zamyatin mentions the potential vulnerabilities of velvet forks but notes that the degree of these threats requires further study. So, in one of the scenarios, "velvet miners," as Zamyatin calls them, update the software, while the others ignore the new rules. If the blocks generated by velvet miners become more profitable than normal ones, other miners can "prefer the updated blocks to blocks created by the old system." Since today's attack patterns in most cases do not assume different block profitability, this can have "an unpredictable effect on the security assumptions of such systems," Zamyatin writes. Another possible way of attack is "selfish mining." This is a process in which miners hide the fact that they have generated a block, and the remaining miners continue to search for it, while the fraudsters are already looking for the next one. This gives them a head start for the next block.
Currently, Zamyatin examines proposals for changing the network, for which a velvet fork will be potentially useful. In particular, he expects that it could be used to add the GHOST protocol, which was originally developed for the Ethereum blockchain, to the Bitcoin network. To speed up the blockchain, it completely restructures the system and, in Zamyatin's opinion, may not receive sufficient support for implementation as a soft fork or a hard fork.
Thus, one of the main advantages of a velvet fork is that it helps circumvent the stage of disputes that usually flare up around major code changes: "Even several versions [of the blockchain—DeCenter] can exist in parallel; perhaps, they will even be compatible with each other, all without the need for often conflicting soft forks and hard forks," says Zamyatin.