Any panel discussion billed as a “deathmatch” should be a lively one, even if it’s about a highly technical subject like blockchain. And the session titled “Blockchain Deathmatch: Permissed-ed vs. -less” at this year’s SXSW conference and festival in Austin, Texas, did not disappoint in that regard.
The session on 14 March featured IBM engineer Christopher Ferris, who sits on the Hyperledger governing board; St. Mary’s University School of Law’s Angela Walch, who’s also a research fellow at University College London’s Centre for Blockchain Technologies; and Blockchain Capital venture partner/Bitcoin educator Jimmy Song. Cesare Fracassi, an associate professor of finance at the University of Texas at Austin, kicked off the fireworks with his opening question: are the critics right that a permissioned blockchain isn’t really a decentralised ledger?
Ferris’s response – that a permissioned blockchain is “an evolution of distributed databases but one that adds a layer of trust” – sent Song into full Bitcoin evangelist mode. Hyperledger, as a project with an orderer service that lets users choose their own consensus method rather than relying on Bitcoin’s proof-of-work method, Song said, “it’s not decentralised…, it’s not Byzantine fault-tolerant”.
“The ordering service is a central point of failure,” Song continued, noting that if the service were hacked the wrong data would go out to others on the network – something Bitcoin’s proof-of-work mechanism prevents, he added. “It’s centralised if there’s a single point of failure.”
Walch noted that it could be argued that Bitcoin also has similar centralisation, as resolving a serious inflation bug that was discovered last year relied on a small group of Bitcoin core developers who made the decision to delay publication of the bug’s details until users could patch the problem.
“I am struggling with how Bitcoin doesn’t have similar centralisation of power in the core developers,” Walch said. “That to me is power exercised by a very small number of people over all the people relying on the system.”
Song pointed out that Bitcoin Core software is not the only software running on the network, and that other software did not have the inflation bug. Rather than showing centralisation, he added, the developers’ response simply illustrated responsible disclosure and the benefits of running multiple nodes.
The discussion eventually reached the point where Ferris said Song, by advocating for systems with no governance beyond proof-of-work, was “arguing for anarchy” and Song said a permissioned blockchain is an oxymoron that makes no sense when an Oracle database that is “going to eat your lunch” is a far cheaper and less complicated solution.
Fracassi eventually stepped in and noted: “So now you understand why I called the panel ‘Deathmatch’”. He suggested that Song views systems as being either decentralised or centralised, with no grey areas in between, while Ferris and Walch consider distributed ledger technologies as existing on a spectrum from more to less centralised. The other panellists agreed.
Walch, who has argued that Bitcoin’s core developers essentially function as fiduciaries for the cryptocurrency, added that some regulation will eventually be needed if such systems are used for critical purposes like finance.
“We wish in these permissionless systems that we’ve escaped the need to trust in others and that there’s no power exercised and – hooray, hooray,” she said. “But we haven’t. It’s just shifted. It’s just shifted to another small group of people and I think we’re going to want that small group of people to owe us something.”
Song responded by pointing out that, while regulators might pass laws to control Bitcoin, those will be hard to enforce and users will find ways around them. (He had earlier noted that China’s effort to ban Bitcoin exchanges has resulted in people conducting transactions on WeChat and Telegram instead.)
In the end, Ferris acknowledged there will be room for both permissioned and permissionless blockchains, while Walch said it’s eventually likely they’ll be called different things because they have such different capabilities. Song agreed there are “good ways to do both”, but added: “If you’re going in that direction [permissioned], you don’t need blockchain.”
Song concluded: “Outside of Bitcoin, everything else has a central point of failure.”