This is the first part of an extensive interview with Hyperledger Director Brian Behlendorf in advance of the Hyperledger Global Forum which takes place in Phoenix USA on March 3-6. Register by February 18th to get the discounted rates.
Highlights
- Hyperledger to serve as a bridge across the spectrum of permissioned to permissionless blockchains
- Hyperledger has a royal flush of different approaches to enterprise blockchain
- Unsure if there’s room for a seventh, but possible consolidation
- Enterprise Ethereum Alliance compared to Hyperledger: standards versus software
- Permissioned networks need to become more decentralized
- Vast majority of transactions will be on permissioned blockchains indefinitely
Thejourney started with a 2016 visit to Shanghai
Behlendorf: I guess the kind of biggest thing that came in last year was Hyperledger Besu, and one of my first trips as the leader of this project was to China in 2016 to the second Ethereum Devcon in Shanghai. Mainly I wanted to go because I wanted to learn more about Ethereum. And I actually had met Vitalik and Bo Shen when they did their original ICO fundraise for this a year earlier.
Butreally, I wanted to see what the developer community was like aroundit and be super sharp about where we were going to positionHyperledger when it came to permissioned versus permissionlessblockchains.
Becausecertainly at the time, and even to a large extent today, thetechnology worlds behind public and permissioned are very differentand very different consensus mechanisms; very different algorithms;communities of developers with very different ideas about use casesand that sort of thing.
Aspectrum from permissioned to public
Butwhat became clear to me while there, was that it was eventually goingto be more of a spectrum. That as permissioned blockchains gotlarger, they would probably need to inherit and learn from theexperiences and maybe even adopt some of the algorithms that thepublic LEDGER communities were starting to pioneer.
Notso much around proof of work. I think there’s still a lot ofskepticism in the enterprise space around that, if only for theenergy load, that sort of thing. And not so much for the DAOdistributed autonomous organization kind of automated robotization ofmanagement kind of ideas.
But partly for how do you do these things really at scale? And so I felt it was important to have an olive branch out to that community. And make sure that as the permissioned side of the blockchain world grew in acceptance and deployment, that we could evolve Hyperledger to a point where it could serve as a bridge across the spectrum.
Iwas still pretty adamant I did not want Hyperledger to be running amain net or a token. I told people that you’ll never see a hypercoin. I still believe that. And thereby avoid the minting money outof thin air kind of thing.
ConsenSyswas part of original 2015 Hyperledger cohort
LikeI said, it was important to be close to the technology, and thatmeant being close to ConsenSys. ConsenSys was one of the companieswhich was in the initial cohort ofcompaniesinvolved with Hyperledger at the announcement in 2015, and they werearound for a year. They kind of went off and focused on publicblockchains and the ICO market for a while.
Butstarting about two years ago, they cameback around and realized the enterprise space was going to beimportant to them. And prior to last year, one project came in calledBurrow, which was a tiny piece of Ethereum technology, as well as arelationship with the EnterpriseEthereum Alliance.
Sowhen ConsenSys said: “Hey, we’ve been building this alternativestack for Ethereum enterprise technologies that could run both publicmain net Ethereum, as well as permissioned blockchains, and we’rebuilding it to be Apache licensed,” it felt like ready-made forHyperledger.
Soa lot of diplomacy, (we had) a lot of conversations with engineersabout how open source works, but also how Hyperledger’s communityworks, through a lot of conversations with the existing Hyperledgercommunity leaders. We brought in Besuand had a very frank and public conversation about how all of thisshould work. And that led to the project being accepted. And now it’sin.
Theroyal flush of enterprise blockchain approaches
I’mnot going to say that we’re done adding new frameworks. But the sixthat we have now represent a pretty royal flush of all the differentkinds of approaches I think you could take to building enterpriseblockchains.
Fabric,which is very much like the granddaddy of the project, it just hit2.0. You saw the announcementof that. It’s still the most widely deployed enterprise blockchainplatform out there. Very flexible, very much an operating system,very generalizable.
Andthen, we’ve got the one focused on identity (Indy).You’ve got the one focused on digital assets, Iroha.The one focused on being a bridge to public and private being Besu.And then Sawtoothwhich is still a more experimental platform. Those six (includingBurrow),I don’t know if there’s room for a seventh, to be honest.
Thecommunity will decide
Thegood news is it’s not up to me. It’s really up to the community.
The community had to be convinced there is room for Besu, as a number six. I would say we might even consider seeing consolidation before we see expansion of that set, but anything’s possible.
It’shard for me to say that there’s a major missing approach to buildingdistributed ledger now inside Hyperledger.
AnotherEthereum link – Hyperledger Avalon
We’vehad a few other projects come in recently, such as HyperledgerAvalon, which is the main other one that I’ll highlight. Avalonis an implementation, actually again it’s Ethereum related because itis implementing a specification that came out of the EnterpriseEthereum Alliance around what they call the TrustedCompute Framework.
It’sa generalizable way of trying to describe privacy on blockchains,whether that’s implemented through secure enclaves, like Intel’schips or through zero-knowledge systems. And in doing that it mightbe a way for us to bring that better balance between confidentialityand auditability, which is the whole point of using blockchainsanyways. If you want confidential, don’t put it on a blockchain,right?
Butwhat we also want is the ability to, you know, track spending, theability to track a diamond as it goes through the supply chainwithout revealing every intermediary’s complete business flows.
Soproject Avalon is really about moving us further along those lines asa whole community. It’s more of a library. It’s more of a set ofconcepts and tools right now. But I hope that we’ll start to see thefirst deployment of that into at least pilot environments this year.
Q:Can you clarify your relationship with the Enterprise EthereumAlliance (EEA)? Because as an outsider you look like you have moreand more overlap.
SoI see some pretty sharp distinctions. One is, and this is true in alot of other technology domains, it’s really, really good to have astandards body in a domain separate from the leading open sourceproject in a domain or from the open source projects in a domain.
Thekinds of stakeholders you want to pull together around a standard.The kinds of IP processes you want to manage in the development of astandard. The fact that (for) a standard, once you eventually set it,it (should) not really be changed all that often. So there’s a lot ofpressure when you publish it to make sure you’ve gotten it exactlyright.
Whereaswith software these days, you know, being agile and publishingupdates frequently and continuing to refine and add features, thatsort of thing is important.
Allof these lead to very different collaboration cultures and differentorganizational structures, even different agreements between theparticipants. And so we’ve always said that Hyperledger is not astandards body. And it’s important for somebody else out there to bedoing that kind of work.
Ifnot the EEA then another standards body
Soif the EEA hadn’t come along, I would expect that you would have seensome other type of enterprise blockchain standards alliance comealong that perhaps wasn’t directly focused on Ethereum. Andobviously, there’s standards efforts at ISO and the identity relatedstandards work and a couple of other works, and all that iscompletely compatible with Hyperledger.
It’snice for the development teams at Hyperledger to have the choice ofwhich standards to implement, how quickly, and potentially even comeup with new de facto standards that could eventually get proposedupstream to somebody else’s standards body.
Sothat’s going to be a pretty sharp distinction between us and the EEA.And that’s borne fruit for us in, for example, project Avalon.Hyperledger being able to now take this standard defined elsewhereand build implementations of it. So that’s something that I think isimportant to keep in mind. And it’s always good to know where theboundaries are in any relationship like this. That’s just kept itvery productive.
Q:Do you have any views on the path of some applications moving ontopublic?
Ithink it’s inevitable that there will be some applications running onthe public ledger networks. DeFi seems to be the kind of thing takingoff there. But I think the vast majority of transactions for ageneration at least, and I don’t see any reason why this changesafter the generation frankly, will take place on permissionedblockchains.
Thereasons for that include a blockchain use case will probably define acertain jurisdictional kind of coverage. This blockchain is governedby the laws of country X or GDPR or something like that. And often,those regulations will have some sort of data residency requirements,and privacy requirements that will be really hard to enforce if youdon’t have the ability to bind all the different participants with acopy of the dataset to a set of agreements.
Hopefully,you can use smart contracts and others to provide a lot ofconfidentiality. But you know, if you and I have some sort ofbusiness arrangement and you end up with a copy of data, there’s nosmart contract in the world that can delete that data out of yourhands if I wished it.
Therehas to be, in many cases, a contractual relationship between partiesthat describes the use of that data no matter how thoroughly we’veencrypted it on whatever blockchain we’re using. And so for mostparticipants, most people, they’ll want that kind of agreement boundinto the network.
Permissionedbut more decentralized
Nowthe thing that permissioned networks need to do is themselves be moredecentralized than many of the ones that you see today. I think, manyof these networks that have launched, they’re still somewhat in theirearly stages, where it makes sense to have one technology partner tohelp bootstrap to get everyone on board and push it forward.
Butmy take is, as soon as they’re in production, you should be open toadding nodes to that network, not only from other end users. If it’sa banking network from other banks, that sort of thing if theyqualify and are able to sign whatever participation agreement isrequired. But also from other technology providers, from other cloudproviders or from nodes that are hosted by the end users themselves.
Ithink if you do that, and then I think if you also make it easy forsmall and midsize businesses to either participate as kind of fullyvested citizens on a blockchain, able to submit transactions, reviewtransactions, confirm the validity of transactions. Or do thatthrough an intermediary of some sort, with the choice of who to trustin doing that. If you make it easy for small, midsize businesses tojoin these blockchains, then that basically erases the advantage ofdoing some of these things as a public blockchain, which isaccessibility.
Accessibilityand blurry lines between permissioned and public
Arguablythat’s the main reason why advocates of public blockchains forenterprise use cases are advocates. They say it’s because then youdon’t have to ask anybody for permission. You just jump on and startengaging. I have no doubt that even enterprise stuff done on publicchains will still implement access control, or KYC (know yourcustomer) or some other type of criteria threshold in order toparticipate.
Andso I think this is why I’m saying that the line between permissionedand public will get awfully blurry. I do think that the vast majorityof transactions will be taking place on permissioned blockchainsindefinitely. Just because that’s an architectural model that mosttechnologists and most companies are going to find more familiar.
