Most observers of the ecosystem will probably agree that 2015-2017 were the years of the enterprise blockchain. It was during that time that many startups catering to enterprise were founded and funded, including Monax and Stratumn (where Brian and Sebastien previously worked). For the permissioned blockchain camp, adoption would come in the form of enterprise use cases, arguing that public networks carried too much risk, and lacked needed features like privacy. While much of the hype has subsided, large companies in every sector are forming consortia and leveraging blockchain to provide better process traceability and transparency, and reduce their reliance on third parties.
We’re joined by Brian Behlendorf, Executive Director of Hyperledger. When Brian was last on the show, he had recently started his role at the Linux Foundation. Now, two years later, Hyperledger has gone from a nascent project to a mature ecosystem of technologies with hundreds of members. With dozens of networks in production, and backed by companies like IBM, Hyperleger is the most widely used blockchain technology for permissioned networks.
Topics we discussed in this episode
- The most important developments in Hyperledger in the last two years
- Hyperledger’s family of technologies
- Production networks on Hyperledger
- How Hyperledger Fabric differs from Tendermint and Ethereum
- The evolution of the enterprise blockchain ecosystem
- The separation between the permissioned and public blockchain ecosystems
- Industry use cases for Hyperledger Fabric
- Brian’s skepticism about ICOs and tokens
- The growing Hyperledger community and upcoming Global Forum in Basel this December
- Toptal: Simplify your hiring process & access the best blockchain talent . Get a $1,000 credit on your first hire at toptal.com/epicenter.
Sebastien Couture: Hi, welcome to Epicenter. My name is Sebastien Couture.
Brian Fabian Crain: And my name is Brian Fabian Crain. Today our guest is Brian Behlendorf. He’s the executive director of the Hyperledger projects and we’ve already had him on just over two years ago. It was a very interesting conversation.
Sebastien Couture: Yeah, I mean, I listened to the episode before having him back on again, and it’s just really fascinating to see where our mindset was two years ago with regards to enterprise blockchain. And I mean, not just his, but like what he was saying really resonated with me I think like where I was at back then. And I’m wondering if you’ve felt the same way.
Brian Fabian Crain: Yeah. I mean, many listeners will be aware that I used to work for this company Monax which was the first enterprise Ethereum company there that started 2014 to like fourth Ethereum and make it an interest version of it. And so I was working there until just about when we did that interview I think with Brian just about two years ago. And then of course Monax, their project became part of Hyperledger. It’s not one of the Hyperledger projects. But then in the last two years I haven’t really spent much time in the enterprise blockchain space at all, and haven’t followed it too much, so it’s very interesting to hear, okay, what has happened since then?
What progress has been made? And even though the hype has died down and it’s not really in the news so much anymore, all of these enterprise blockchain projects, it seems that people have delivered stuff. And we’ll hear much more about that from Brian in a second.
Sebastien Couture: Yeah. There definitely was a lot of hype, and I mean, for me I spent most of those years working in that space, of course, working in Stratumn, and I think I’ve come out of that with a lot of skepticism, but I must say that, I guess, after speaking with Brian today that some of that skepticism has fallen to the side. And I can see how a lot of those promises that were made early on are starting now to come to fruition. And apparently some of these networks are now in production and there’s actually use cases being built on Hyperledger. Speaking about that, there’s a conference going on from December 12 to 15 in Basel is the Hyperledger Global Forum and I’m going to be there. I’m going to be doing some content.
They were nice enough of you to pay my flight over there, and invite me to the conference so that I could produce some interviews with some of the key members of that community. If you’re interested in attending that, again, it’s December 12th to 15th. It’s in Basel, in your hometown. And I’ve never been, so I’ll get to see what your childhood and teenage years were like living there.
Brian Fabian Crain: Pretty neat.
Sebastien Couture: Yeah. How it set me like the places you used to hang out at, and tell me what you’re hanging out was used to be. We’ll check them out.
Brian Fabian Crain: I used to live in Basel, of course, until I was around 18 years, it’s my hometown.
Sebastien Couture: Yeah. It’s going to be a pretty big event. Actually, there’s going to be about a thousand people there apparently, so if you’re interested in attending, I’m not sure what the price is, but definitely check out the website. And I think they told me there were some coupon code, they haven’t sent it to me yet but if you are interested in going, and tweet at us and we’ll send you the coupon code for the event. I think there’s a coupon code, they just haven’t sent to me yet. If you’re going there, it’ll be great to see you. Definitely come say hi. Without further ado, here’s Brian Behlendorf.
Brian Fabian Crain: We’re here today with Brian Behlendorf. He was a guest on the podcast today that was a very long time ago. I think just around two years ago. And we spoke, of course, about Hyperledger, and he’s the executive director of Hyperledger. And that was a fascinating episode, we just listened to it again in the last days and revisit it. We were excited to have Brian on again and talk a bit about Hyperledger and about how it has developed and about, generally, the use of blockchain in enterprise use cases or other use cases that Hyperledger’s just been working on. Thanks so much for joining us, Brian.
Brian Behlendorf: Thank you. It’s great to be here. I think that was two years ago almost exactly on our 14th blockchain years, I think.
Brian Fabian Crain: Yeah, exactly. And we do recommend people to check out, it was interested. Brian has a very interesting background. Among other things, she was the main initiator around the Apache software license and the Apache Foundation. Also key part of The Linux Foundation of the Apache web server, so he’s played a pivotal role in many of the key at web technology. We spoke quite a bit about that in the last episode too. And including interesting things like the pros and cons of different software licenses, and the case for it, Apache license. I think that was very interesting and I’m sure a lot of people would enjoy that. And I think it’s just as relevant as the episode years ago.
Sebastien Couture: So episode 160. Episode 160 for those who are curious and want to listen to it.
Brian Fabian Crain: Maybe give us just two minutes of a very brief background, Brian, about how you originally ended up taking on the Hyperledger project, and what drives you in this project.
Brian Behlendorf: Sure. Briefly for those who didn’t hear the last podcast or might otherwise not know, I’ve been an executive director for Hyperledger since May of 2016. The project actually started about six months before I joined. Previously to that, I mean, I’ve done a bunch of different roles. And you mentioned a few. I’ve also worked as a C.T.O. for the World Economic Forum. I worked in the White House during at the beginning of the Obama administration. Most of the time, though, I’ve been a startup entrepreneur and then right before this I was working at a venture capital fund. And in addition to nuclear fusion and robotic surgery and all these other sci-fi fun things, obviously, we were getting pummeled with requests from the Bitcoin, a blockchain community to look at their companies, their proposals.
And even met Vitalik and Bo Shen when they came around to do the original I.C.O. for Ethereum, really started to understand the technology and the rationale in the business, but joined Hyperledger because I felt like it had really the most realistic sense of how this technology was going to play out. That it was something that wasn’t being paid attention to as much as some of the others. And was closest to a lot of my background with both enterprise software and obviously had a great organization behind it in the form of The Linux Foundation. And so I jumped right in.
Sebastien Couture: When we last had you on in December of 2016, it had been about six months since you’ve been working with Hyperledger. Can you give us a sense of like in the last two years, how has this particular organization changed? Just what steam look like, but then also from your perspective, how has it grown now and like, what does it look like?
Brian Behlendorf: The most important thing is we’ve actually been shipping now production code for just over a year. Fabric, I think the September 2017 it was, went 1.0 with general availability. And in the year and a quarter or so since it’s been launched, it’s now incredibly widely deployed. We’re tracking 50 different production networks that are running on top of Fabric in one form or another, and there’s probably quite a few we don’t know about. There’s also now production networks on top of Sawtooth. Sawtooth in 1.0 in January of this year, they’re about to cut their 1.1 release. And now, both of these are different from each other as I think I talked about in the last call.
They’re different in the same way that MySQL and Cassandra might be different as technology platforms, but both of them have now found their footing and are being used in to track real digital assets, right? We’ve also now added quite a few additional projects. I forgot which ones were a part of the group when we talked in December, but I think we had launched Indy by that point, I’m not sure, but maybe not. But Hyperledger Indy is becoming this really amazing project. It’s the platform for building distributed digital identity networks, formed around this concept of self-sovereign or user centric ID. And it’s now the basis for this large network convened by an organization called The Sovrin Foundation.
And there are pilots now involving nation-states projects on the horizon to do national ID systems for the government of Sierra Leone. In Canada in British Columbia, there’s a big project using it for self-sovereign ID for businesses. There lots I can go into there, but Indy has really taken off. It hasn’t in 1.0 release yet, but that feels pretty eminent. And then also pretty noteworthy is we picked up in Ethereum, virtual machine project called Hyperledger Burrow, which is an implementation of the E.V.M. designed for enterprise use, non-mainnet use. Pretty noteworthy, doesn’t have any aspiration of being a mainnet client, but it is now running on quite a few production networks.
And it’s also been reported to run on top of Sawtooth and on top of Fabric. That’s pretty exciting as well, and it’s the basis for Monax’s agreements network that they launched, which is like a legal technology platform. Lots of other projects. We’ve actually just added an 11th project called Hyperledger Ursa, which is a library of cryptography routines used by the other projects. We’re refactoring all that out into a common library so that projects can more easily pick up advanced hashing algorithms or zero-knowledge proof types of techniques. And that’s hopefully something that’ll allow us to easily add zero-knowledge stuff to Fabric, Sawtooth, Burrow, Indy, that sort of thing.
And all these technology platforms are continuously improving, even after they had production release that’s not the end. If anything, that’s the beginning. Fabric 1.4 just came out, for example, it has support for zero-knowledge asset transfer inspired by the Zcash approach to obscuring participants and amounts in transactions even though you should be able to make strong guarantees with the overall integrity of the ledger, right? Lots of activity that way, and correspondingly more developers coming in as core contributors. We have over 800 different developers now make a contribution of some form to the software behind the 11 different projects.
And that might sound small, there are millions of developers out there working on things like GitHub that sort of thing, but those are 800 devs who’ve actually gone outside of just using the code to actually feeding back upstream. And I’m pretty happy with that number. I think it’ll grow to larger, no doubt, just as a function of time. But now especially on Fabric and Sawtooth we have real evidence of the multi-stakeholder, multi-vendor community growing around that. And on that front, I talked about the membership model for Hyperledger. We grew pretty quickly in 2016, 2017, and this year continued to grow. We have 288 members at last count, and most of these are companies, but a few of them include nonprofits, and central banks, and government agencies, and that sort of thing, and really happy at the diversity of that crowd.
And the vendor directory now lists over 70 different companies building products and services on top of Fabric, or Sawtooth, or one of the other projects. That’s by the numbers, I’d say generally speaking, this year has been a year where a lot of P.o.C.s did go into production, when pilots did go to production, where we started to see banking networks, trade finance networks get set up. In China there’s one that is about extending letters of credit to trading partners. This is set up between the People’s Bank of China, actually CITIC Bank, China Minsheng Bank, and a bunch of others. And they’re currently issuing about a billion renminbi a day, which is about $160 million worth of letters of credit.
Likewise, the Hong Kong Monetary Authority just launched a system. There’s another startup in Singapore called dltledgers that has done a project with a couple of big banks and agricultural firms for 4,500 farmers in Australia. And in the two months that they’ve been in production, they’ve issued $750 million worth of trade finance extensions, that sort of thing. Seeing lots of those, we’re seeing supply chain stuff all over the place. I think I mentioned Diamonds when I was last on. That the Everledger network has gone into even deeper production now. And now we’re claiming they’ve been saving millions of dollars in attempts of fraud on that network, which is pretty cool.
But we also see now production supply chain networks in coffee, in the Walmart Network, of course, which is initially just tracking leafy green goods, green vegetables, right? As well as big network around rice in China. And then in healthcare, that’s been a little slower to take off but I feel like the impact there will be pretty big. And in healthcare, there’s a network run by Change Healthcare that is processing millions. It’s an even bigger number than that sounds. I’m only allowed to say millions of transactions a day in management of these kind of claims between parties, insurance claims between parties. A lot of good stuff going on. And the technology is doing well when it stood up against its competitors.
We see people running fairly open and sometimes not open trials between two or three different technologies and choosing us a lot more often than they seem to be choosing the alternatives. And it’s real buoyancy to be waking up every morning to a new post somewhere on CoinDesk or somewhere else saying a new project being launched by three companies we’ve never directly engaged with, using Hyperledger Fabric, or Hyperledger Sawtooth, or whatever, to go build something that we hadn’t had to push on anybody. We hadn’t had to sell. And that feels like an inflection point of some form. That’s what 2018 has been.
Brian Fabian Crain: Wow. That’s super impressive. Yeah, that sounds like a tremendous amount of achievements and accomplishments. I mean, I’m curious that you mentioned that it’s standing up well against competition and people tend to go with that. Are you talking here about competition in the sense of other blockchain platforms or are you talking about alternative approaches to maybe accomplishing some of the same aims that people generally want to use blockchains for?
Brian Behlendorf: Certainly the number one competition is doing it with a centralized database, right? And I think I even might have said on the call two years ago, there’s no blockchain use case that couldn’t also be done by a centralized database. It’s something you don’t want to do when you have a really either politically sensitive type of thing or there is no central party that everyone trusts to be able to run that central server. But trust can always be bought for a price, right? Like there are some organizations that are neutral organizations that sit at the center of marketplaces that serve as public utilities and some of those are dabbling in blockchain technology, to understand them, that’s like SWIFT, and D.T.C.C., and that sort of thing.
But there’s others who are still happy to trust central providers if they can be as neutral as possible. And so that’s always the number-one competition. And I’d say any time I’ve heard of pilots not continuing, it’s been for that reason. And that trust, though, is also a function of cost, right? The cheaper that we get these technologies to deploy and manage and easier it is to scale up and the more developers there are out there who understand how these technologies work, then the more and more use cases, the more and more situations where the cost versus trust issue will resolve in favor of watching technology of one form or another. And then yeah, to be fair, there are other blockchain systems out there.
There are some who will sometimes tell you they’re not a blockchain, sometimes they are like R3’s Corda. And R3 has a sales team out there that is aggressively selling into many of these operations. And on our site you do have I.B.M. with their sales team and Oracle, and Accenture, and a whole lot of startups selling, but sometimes a determined sales force will get a product into a space. There’s some interesting R3 projects happening out there and you could say they’re competitive with some of the other trade finance networks, but at the end of the day I think these converge in some way. And then there’s Quorum, and Quorum, had a lot of interest in it. And some production usage, I’m aware of a gold trading platform that’s using or aiming to use Quorum when it launches.
I haven’t seen as many out there. And by the way, one interesting site people might want to visit is something called the Unbounded Network, unbounded.network. It’s the beginnings of what hopefully could become a coin market cap-like system for tracking where these consortium blockchains are. It’s what’s built by Jonathan Levi and then Sarah. And it’s tracking who’s out there, and what are they running on, and what is their intent, that sort of thing. And actually, I believe some sort of directory like this in the long-term becomes a way that people discover and join these consortia networks, but the ones that are listed there, the vast majority of them are on Fabric.
Quite a few on Sawtooth. And then a trickle of them on R3, and Quorum, and other Ethereum private ledger equivalents. That’s competition, but by far, the biggest competition is simply people who say, “I’m happy with the centralized network.” And that’s who I spend more time thinking about is how do we make the technology easier to use, more people adopt it, and more appropriate for more situations.
Sebastien Couture: Looking back in the last two years, I think a lot of our assumptions about how these networks would work have either turned out to be correct or perhaps we look back on them and realized that are something we perhaps made some false assumptions about how these networks would sprout up, these consortiums. What are some of the biggest mistakes you can point to or the things you can look back on and say, “Okay, maybe my thinking about this has evolved and I was wrong about this specific thing and the way that things have evolved to today.”?
Brian Behlendorf: That’s a good question. I don’t know that we’ve made any observations that we’re fundamentally wrong. I think the timing is always a question. I don’t mean timing in the way that . . . what’s that phrase? Even a stopped clock tells the right time twice a day. But I mean timing in a more . . . just how long it takes for certain things to flush out. I think we have gotten to production use cases about as quickly as I would have thought. We know that this is not technology like a new programming language or a new kind of web server where somebody with a motivated need can get something up and running and demonstrated overnight, right? That to really get the stuff in production it’s always about building a coalition of businesses that have a common need.
And demonstrating value means getting it into production across a number of different companies to all of their own integration needs and that sort of thing. Necessarily this stuff is going to be slow burn. I think one thing that we spent a lot of time on and it hasn’t been as easy to achieve as I would have liked, would have been getting more developers in. I mean the 800 number, I’m pretty happy with, I think that’s a testament to the breadth of the technology portfolio we have, but this has been early enough on and people don’t really like dealing with the plumbing as much as perhaps I did and my generation did 20 years ago, I don’t know. People like getting to the apps like using this stuff but then using it to accomplish something.
I can’t begrudge them that, but getting more developers into the core of some of these technologies has been a challenge. And many of those 800 developers themselves are new to open source and so getting them to understand how to work publicly and transparently and when they’re bringing a technology from that might have started in-house and bringing it out, how do you do that in a way that really gets other people to feel like stakeholders in the process and directly engage. We’ve had to invest a lot in that, to get that there. And it’s working, but it’s taking perhaps more than we thought. But no major strategic mistakes. Another big strategic that I think I even said it two years ago was thinking about how we would relate to the public ledgers and to the currency community, cryptocurrency community.
And while I’ve famously been skeptical about I.C.O.s and not so much about tokens, just the cryptocurrency-driven token model. It’s always been about figuring out how do we make that part of a hybrid rather than treating that as the other or treating that as the enemy worst of all, because it’s really not the enemy in any meaningful way. But what could we do constructively from Hyperledger to go and build bridges with that community where there’s good technology coming out of that side of the spectrum, right? And that’s where putting effort in to reach out to the Ethereum community, bringing them to bridge in and their M.V.M., but also joining the Enterprise Ethereum Alliance, and hopefully the cryptography library work that we’re building will be useful.
In fact, one of the co-sponsors on that project are a couple of developers from DFINITY. That’s pretty encouraging, right? Ultimately want people to think of who Hyperledger is, less as the enterprise kids trying to be cool again or something like that, or trying to be anti-public ledger in any sense. And really to be more about that full spectrum of ways to use distributed ledgers and smart contracts.
Brian Behlendorf: There’s a project in British Columbia being run by an innovative group within the government of B.C. And they’re called what they build the Org Book which is like Facebook but for businesses that are registered in British Columbia. And they wanted to try to tackle the problem that small businesses have engaging with government when there’s different levels of government, different agencies that ask different things. If you run a restaurant, you need to file as a business and then go get a permit to serve food and then a permit from somebody else to serve alcohol. And I mean all these different engagements where lots of people in government have tried to provide a unified frontend to government services which is useful and great, but that means a whole lot of integration work behind the scenes on things like identity and that sort which self-sovereign identity gives us an opportunity to pivot that a bit.
And so their approach was that, “Can we use self-sovereign ID for businesses to register?” And so they picked Hyperledger Indy. I’m not sure the backstory of why they chose it. They’ve now loaded into, and are working with The Sovrin Foundation on that network. They’ve now loaded into the system 50,000 business registrations. They built a frontend. I mean, the idea ultimately is that these businesses get wallets to manage their digital identity documents, and permits, and data stations, and claims, and all these kinds of things. For the short term they built a web frontend that acts as a virtual wallet for them. And that’s in production now. And now that team is sharing what they’re doing, they’ll be at Hyperledger Global Forum.
For example, talking about it, but they’re sharing it with other government agencies and the chance there to use that same identity, that same registration, that same wallet for a business to engage in trade finance, and to pay their federal level taxes, to do all these different things is like built-in to the system. And that’s really encouraging and I wonder if in the 10-year horizon, like the biggest impact that we have especially when it comes to things that the average consumer will be able to see and experience might be in this reinventing how digital identity works. And that project is getting attention now.
In fact, Kiva, the nonprofit that pioneered the peer-to-peer lending space announced that they’re going to be building a national ID system for the government of Sierra Leone using Hyperledger Indy as the identity layer and using Fabric to implement a credit history system so that the cost of lending to businesses in Sierra Leone goes way down, which is their bread and butter and why they’re getting involved. And so that feels really awesome to me. It feels really cool. And that gets me excited. And I mean, the supply chain traceability of the healthcare projects, all those are also interesting and happy they talk about specifics over there too, but . . .
Sebastien Couture: Let’s maybe just spend just a bit of time on this particular use case, because it’s interesting because it incorporates a lot of the things that I feel enterprise blockchains promised to solve. And I just want to dig in. And so I agree, I think identity is one of the components that is needed for these systems of disintermediation to function, and especially if you’re talking about known entities in the systems such as businesses in this case. It’s encouraging to see that that is playing out and entering this production phase. But when you say that these systems are in production, what constitutes production other than the fact that other actual users are using it and it’s serving its actual use case?
But what are the volumes? And more importantly, are these use cases that are in production actually doing what they were meant to do in the first place which was like disintermediate the central actors that we’re all trying to separate ourselves from?
Brian Behlendorf: A lot of questions, let me see if I can get to the majority of them.
Sebastien Couture: Sure.
Brian Behlendorf: For me, the defining characteristic of a blockchain network in production is that it serves as the system of record for something. It’s not just an echo of some other business process that’s actually defining what’s going on, but that if there’s any question or dispute about what’s the reality on the ground it’s what’s written to the ledger. And to the best of my knowledge all 50 of those use cases that we know or sorry, of those production networks we know of are serving that purpose. They’re not just documentary in function. They’re actually transfers of assets of one form or another, and again, serving as the system of record. The transaction volumes, though, it could be all over the map.
I do not know what the transaction volume is on the British Columbia network today. When we’re lucky we get information about this from other parties being able to share with you. Change Healthcare, for example, being willing to say that there’s millions of transactions on that network, even that is hard to get a sense for scale from. The network I mentioned, the startup in Singapore that’s working with D.B.S. Bank, I’m sorry, that’s the Development Bank of Singapore, Agri Corp which is a large agricultural firm, 4,500 farmers, that’s where I have more concrete data. In two months of operation they’ve conducted 2,400 transactions as they call them, which actually represents 3.2 million entries to the ledger, because a given transaction is a lot of data points and a lot of steps in it, that sort of thing.
I mean that sounds like a small number of transactions compared to some blockchains that you could say, “Well, we could do 2,400 transactions a second,” right? But each transaction in that network is average size in terms of the amount of credit being extended of $300,000. Over those two months that means, sorry, $750 million worth of transactions where the credit being extended to farmers and other freeholders. But all these scales are going to be all over the map. There’ll be some networks that are extremely valuable that will be a small number of transactions per day. Other networks that are going to be trying about into the facilitating micro transactions in some way, that might be tens of thousands a second.
I think the fact that we’re talking about lots of different networks out there means that this tool to bear, when you’re asking about scalability, well, how do you guarantee we’ll be able to scale up to a certain amount? Well, by segmenting by interest, there’s a lot that we can do to make sure that scalability here is a question that can be managed locally rather than trying to answer the scalability question globally. But it does make it really hard to try to come to an objective measure of value. Again, there’s no coin market cap for consortium ledgers, right? I would love there to be. I’ve talked to people about building something like that. And because doing something like that will help build confidence that these technologies work at the scales people would like to use them at.
I think we’ll get there over time and it certainly are our goal to collect even if it has to be anecdotal at this point, there’s enough information about transaction volumes and numbers of nodes on those networks and that sort of thing. Something we’re also doing through the performance and scalability working group in Hyperledger.
Brian Fabian Crain: You’ve spoken about actually a lot of different examples of places where blockchains are being used or people have built applications on Hyperledger and these consortia. What do you think are the biggest value asks that have turned out? And are there some ways in which benefits have derived from these applications that may be like unexpected and you didn’t expect that turning out as a benefit?
Brian Behlendorf: Well, I think the biggest value, especially with trade finance, you’re often digitizing a process that heretofore has barely been standardized, let alone digitized. And the standardization has come in the form of a given geography, a given region, or a set of banks getting together and say, “Okay. We’ll use a certain standard process or a template for issuing these kinds of agreements,” but then it still defaults to lawyer-driven processes, such as legal contracts, signatures on paper, fax machines, that sort of thing. And that’s been an area that has been resistant to digitize, not for any lack of computers in that part of the world, but because there hasn’t been a central authority who, especially since many of these supply chains cross from China into United States or into the West where people have been happy to leave the accounting to somebody no matter how neutral you could arguably present them as, right?
And because in a lot of cases these trade finance decisions are essentially know your customer, any money laundering, credit history types of things, any sort of system of accounting for how well businesses are performing, it has been hindered by information sharing laws and processes that were very batch oriented rather than real-time. And were the knowledge of whose shared what with who was opaque to be honest. Some of the benefits of launching these networks are around digitizing a process that has been analog before, and arguably you could have seen many of those benefits by digitizing and centralizing, but by being decentralized you created the political will to create these networks that might not have existed before.
And so that’s a hard thing to measure. It’s a hard thing to go, “Well, that was worth a hundred billion dollars,” or something like that, or worth even the money that was spent to build that system. But that’s again, without deflecting too much, you could say that the internet 1998-1999 wasn’t quite worth the money that was being spent on websites. You couldn’t really point to more or less the traffic on amazon.com in 1998. Probably enough to justify raising some rounds of capital, but probably not enough to say homerun R.O.I. Not yet, right? I think a lot of slack has to be cut. I don’t think there’s any of the projects that right now could say they would stand up to external scrutiny, “We’ve generated this much of a return on the money that we’ve spent building a blockchain project.” But I think that’s eminent. I think that’s really close.
Sebastien Couture: I’d like to come back to my previous question and really understand at the base level where these networks are providing value. And so you mentioned digitization and standardizing processes, and I think that this is one of the things that for sure enterprise blockchains forces upon the actors engaging in the consortia because if you’re going to engage with competitors, partners, regulators or anybody involved in the process, you have to be speaking the same language and standardizing the way that these organizations communicates is definitely one of the benefits of implying these systems. But in terms of actually disintermediating the central authorities or the incumbent service providers that might be in previous times providing that facilitation or that exchange of data.
How have these use cases turned out? Just as an example, if you take the supply chain use case for example. And I say this is someone who’s previously worked at a company that was working on these types of use cases. Actually, both Brian and I were previously at companies in this space. You think the supply chain use case typically what I would see is a client would approach us and this client was a company that, for example, a chemical company. And this chemical company has supply chain where they were cultivating some sort of natural resource. And this natural resource was ending up in a product. And these projects were often being commissioned by these service providers, by these companies.
And what we would see is that the networks themselves would be operated by one, or two, or three actors, and all the smaller actors in the chain, so be it, the shipping company or even all the way down to the farmer were actually just uses a network but didn’t really have any that true validating power in the network. And that if you pull the veil the system was still somewhat fairly centralized. My question to you is that the systems that we’re seeing in production, are they still fairly centralized in this sense or is there really a true distribution of the actors that are participating in these consortia networks?
Brian Behlendorf: Good question. One form of disintermediation that’s going on is definitely in the settlement layer. One of these trade finance networks in Asia is also facilitating settlement through the system in a way that allows the parties to keep that record of who’s settling up with who and then to true-up basically at the end of some periodic process, day, week, month, that sort of thing. But in the course of that eliminating SWIFT as the platform for which that settlement was previously taking place, right? Now SWIFT themselves are conducting trials around reinventing how their network works using a blockchain partly because they realize there’s this competitive risk that emerges, they may wake up and discover their member banks don’t need them anymore because they figured out either a different [insurgent] [00:38:39] provider or they themselves have stirred up their own network.
But at the end of the day, in all these networks there’s still some governing entity. I would argue even the public ledger still has these governing entities out there who decides who’s in the network and who’s out, right? And that should ideally be a very objective decision, a very approachable kind of transparent thing, something to that, is open no matter what size you are as long as you can stand up a [node] [00:39:02] on the network and run, right? But the reality is right now we’re working with businesses who have existing notions about how to form these consortium, those are not always the most open processes. I think we are going to see a lot of experiments in governance take place out there.
I think we’ve seen some of those experiments not work out as well. There was a network called Batavia that was a trade finance network that didn’t really pick up any speed and now has been folded into one of the other Hyperledger trade finance platforms out there. There’s another one which is the joint venture between I.B.M. and Maersk which despite signing up about 80 different partners, all of those have been customers of Maersk. And the agreement there is very Maersk centric and that has meant that you don’t have the competitors with Maersk on that platform. And the competitors have gone off and formed their own blockchain networks as well around shipping, and tracking, and containers in particular.
And ideally you have these large networks because I think the most valuable networks are going to be . . . I mean, we know it intuitively, but are going to be demonstrated to be the networks that have the most participants on it even for those participants who don’t control those participants as customers. And so I think we will actually see competition between networks even in the same space. And I don’t think that’s a bad thing. I actually think it’s good to have competing governing organizations because there’s a trade-off, right? The larger the network is, the more valuable it is, the more people you can do business with directly. Although you can always do cross ledger transactions, that sort of thing, if parties are on two different ledgers.
But the larger the network, the better, but you don’t want that largesse to lead to the governing entity having a controlling position, charging fees, creating unfair rules, that sort of thing. And so sometimes a smaller network will be better if there’s a higher standard for participants on it. More money at stake for anybody who is a bad actor, for example, greater penalties for doing bad things, which will necessarily shut out some other smaller participants. We’re going to let water find its own level in this thing and encourage competition out there, I think. My hope is that if the technology is not only standardized. If people are running a small handful of different platforms then interoperability between them gets easier.
In addition to explicit interoperability projects like we have Hyperledger Quilt, which is an implementation of the Interledger standard, which should hopefully make having a company deal with multiple blockchains even easier in conducting transactions across ledgers. And it’s not the final word in this. That’s one of those projects we would love more contributors to and there’s probably other interoperability standards and technologies yet to come in this space. And we also need to make sure that as we’re building these frameworks that there’s never a technology reason why you can’t add another node or another participant to the network. We can’t make it the case where all these networks run well up to about 25 nodes and or 25 organizations, and then after that there’s no more room.
Because then, the technology becomes blamed for whatever systemic hegemony the industry might try to present there. That’s a challenge to us as programmers. That’s something we should take on, but the governance models, I think, there’s need for competition there, and need for people to map the landscape and figure out what works best on a use case by use case perspective. I don’t think it bubbles up to one gigantic global network, nor does it mean a hundred thousand different networks. But the more common the technologies are across all those networks, the easier they are to deal with.
Brian Fabian Crain: This is great. This is tied into something that I wanted to come back to anyway, because one of the ways that you could potentially see this turning out, let’s say now you have some industry and then you have a bunch of the players, they come together, they build this consortium. But now, don’t you just have at least a high risk that some of these consortiums become very dominant in the industry and they can then actually decrease competition and enforce rules in that industry, keep out new entrants, and so that, let’s say, startup comes up down the line, it will actually be very, very hard to compete, not maybe harder than it is today? Is that something that you’re concerned about?
Brian Behlendorf: Look, any trust rules will still need to play, right? And first up, there’s nothing the technology platforms can do to keep collusion from happening, right? Even if we were talking about gaps on Ethereum. You could have gaps that are built by big players who have a preference for even enforced interoperability with other big players in a way that serves an antitrust, anti-competitive kind of purpose, right? And we will need technologies that help us with that open. We’ll also need regulators who know how these technologies work and can recognize when anti-competitive behavior is happening. I think it’d be awfully hard to hide antitrust behavior on a blockchain network compared to the alternatives, right?
It’s easy right now for companies to collude off chain and come up with secret deals that exclude competitors, provide favorable pricing, that sort of thing. There’s very little the technology can actually do to prevent that kind of thing from happening if at all, right? But if it’s easy to form these networks, if it’s easy to form a rival network, to form a competing network and easy actually for a business to be on two networks at the same time, then I think we have a much more competitive environment to be able to hold that kind of antitrust behavior to account and limit its ability to truly lock things down.
Brian Fabian Crain: I mean, I think that’s a fair point, I guess we’ll see how it turns out. We wanted to speak a little bit about the project Fabric, which is the best known most widely used and most mature projects in Hyperledger. Can you tell us a little bit about what is Fabric and what’s the essence of Fabric?
Brian Behlendorf: Yeah. Fabric is easily the most mature of our technology stacks, the one that has had the most different number of contributors to it, both as individual developers and as different companies that have played a role in building it. It’s the one that forms the majority of the different production networks out there that we know about. And Fabric, first of all, first initially it came from I.B.M.’s internal R&D around exploring the use of distributed Ledger’s to reform a business that I.B.M. has been in since day one, which is not just having a mainframe in the basement of your big company, but how do those mainframes conduct business with other companies’ mainframes, right?
And so there’s a pretty deep history going well past before Satoshi, whoever she was, say, to put together these systems that operate cooperatively but in a way that can be audited, right? In a way that maps how business processes take place that involve steps between parties and things in a way that is more elegant than the web services approach that typically has been taken, one that’s more orchestrated, one that has this extra layer now of Merkle trees, and proofs, and that sort of thing, that the public blockchain showed us were possible. Fabric is based on the notion that you have a defined set of who the participants are in that network, that can grow dynamically, that can shrink dynamically.
Those participants run one or more different nodes on that network that are all capable of publishing to the network, as well as reading all of the transactions off of it. Largely speaking rather than a competition in terms of compute power. Say, that for proof of work you use consensus mechanisms that are more about making sure, hey, we can fiddle these transactions in an order, and if that order is out of whack, for example, somebody tries to spend the same resource twice or transfer the same asset twice, then the second of those gets rejected and thrown back to try again. And so it’s simpler, it’s less ambitious you might say, than what Bitcoin tries to accomplish and it’s a decentralization kind of trust, but it trades that off for really high performance that verify ability and integrity that you would expect from a system that is transferring digital assets.
The very first consensus mechanism was a Byzantine fault tolerant one, which actually proved complex, and unwieldy, and slow. They punted on that for 1.0 and went to use just Kafka as the consensus mechanism. Adding to that a couple of things around endorsement of transactions and such that helped make it truly more of a blockchain. But after that, they’ve now added support for Raft consensus mechanisms that’s coming, although that’s there I believe in 1.3 Byzantine fault tolerance will come back in at some point in the future maybe by 1.4, maybe later.
But when your networks are on the order of a couple dozen to maybe even a hundred participants, as long as you could detect bad actors, as long as you have built-in protections against certain categories of fraud like prevention of double spend, there’s a lot that you can do on that network without having to resort to competition between C.P.U. cycles as a way to decide what’s the next data structure in the chain. And how do you prevent bad actors from being able to get away with attempts to corrupt the system. And one of the benefits you also get from this approach is immediate finality rather than the eventual consistency that you get from less deterministic systems.
And in some cases, above and beyond. There’s a startup called Blockdaemon, which is aiming to be the Heroku of blockchain as they’re saying. And they’ve been doing a lot of work with Fabric as has Amazon Web Services. Particularly for some reason in the Hong Kong office, they’ve been doing a ton of Fabric work. There’s again, pre-built images for A.W.S., for Azure, which is a lot of Fabric use as well. And Baidu, Tencent, and Ali all offer blockchain as a service platforms as well on Fabric. That’s what’s exciting about Fabric.
Sebastien Couture: In the enterprise blockchain space there are other, I guess, protocols that are also used by projects that launch consortium networks. And I think Tendermint is probably one of them, also Ethereum, some assent with the Ethereum Enterprise Alliance. Briefly, can you give us may be like the sales pitch on how Hyperledger Fabric sets apart from those other protocols, and how it’s better?
Brian Behlendorf: Fabric is definitely more of a complete programming environment than Tendermint. And in fact, Hyperledger Burrow runs directly on Tendermint as default. And so I think if all you’re looking for is consensus between a number of nodes as a driver for some smart contract work and to build a raw ledger, then Tendermint plus Burrow which works completely fine there. Let’s see. And obviously there’s other technology stacks out there. I think Corda, the structure for Corda is not really a blockchain, it’s more about being able to correlate transactions between entries in your ledger and entries in my ledger and not sharing a universal sense of the truth, but more converging on a common system of record.
And so Fabric is much more like traditional enterprise blockchains in that way. And when it comes to the Ethereum stack, we just joined as you might have seen a few months ago, we announced a tie-up with the Enterprise Ethereum Alliance. This was after, I mean, when they got started we followed many of the conversations about who they are and what they aimed to do. And I think there was some notion early on that they might be both about building code and building standards. And they hired Ron Resnick who was previously head of a lot of standards work in the telecom space. And so I met him. Actually, I saw him in Davos last January when it was like his first week on the job.
And we started talking about what E.A. could be. And I saw that he really wanted to steer it in the standards direction. We’ve been supportive of that and as we’ve increased the amount of Ethereum support inside of Hyperledger being able to say Burrow standalone, or Burrow plus Sawtooth, or Burrow plus Fabric is a conformant enterprise Ethereum stack and that can guarantee portability with other Ethereum stacks, it’s pretty useful, pretty important. That certification process hasn’t yet launched, but we are tracking what’s going on in that community closely and trying to be supportive of them. And in general want to be good members of the Ethereum enterprise community.
That’s how I would relate to that. And then the competition can be about performance, it can be about additional features, and tie ups with other business. It doesn’t have to be about a fight over your protocol versus my protocol. And I think ultimately that’s how we all win is when we converge to a small number, not necessarily one, but a small number of standards and protocols and just try to build the best engines under those standards that we can.
Sebastien Couture: That’s a good point there. I think that falls into my next question which is with regards to, I guess, if you remember Serenity, and Polkadot, and the substrate S.D.K. or whatever you want to call it. Before I feel like, and maybe perhaps still now, there was really this separation between the public blockchain tied to a token or some sort of currency and the enterprise blockchain which doesn’t necessarily have an asset. And actually most cases is tracking some sort of process and ensuring that every participant in that process is doing what they’re supposed to be doing and that we have that lot of trail. And in a lot of cases I think the use cases that we were envisioning in sort of the enterprise blockchain space specifically . . . well, the public blockchain networks were not really suited for that, purely because they were using a token and we didn’t necessarily need one, but also just because on feature set.
We talked earlier, I think, about private transactions, the ability to have a predefined set of actors that are either validating transactions or just participating in a network. What we’re seeing now I think is that even public networks are starting to embody some of these functionalities and some of these features. For example, the ability to launch a private network that is using a public set of validators that is inheriting that security model, the ability to do private transactions, this sort of thing, is certainly coming in the future. How do you see Hyperledger’s role moving forward, I guess not necessarily competing, but how does Hyperledger and these public networks interact with each other in the future?
Do you think we’ll see private networks somehow connecting to public networks that use tokens? And what would be the interoperability there?
Brian Behlendorf: Sure. I think it was 2015, might have even been a little bit earlier, but definitely before I joined Hyperledger I met the guys at Factom, and they were showing, “Hey, we’ve got these private networks and they’re fence posting their transactions into the Bitcoin network as a way to help participants on those private ledgers have high confidence that when somebody presents a history of that ledger it actually is anchored to the public ledger in a way that guarantees the integrity. And no matter what two parties present different histories you can tell which one’s telling the truth and which one’s lying.” And to me that was like an obvious way in which these two worlds would work together, using the public ledger.
I know this sounds like a little bit of a joke and I think I might have even used this metaphor the last time we talked, tell me if I’m wrong, but almost, that’s like a public newspaper system of public record. In fact, you’ve probably seen some people taking out ads in like the Wall Street Journal, the classifieds posting hashes for contracts and things like that as a way to demonstrate, even if you don’t give the details, here’s the hash of a whole lot of stuff behind this. This is also Tierion’s business model, right? This is something that people have been playing with. It’s also been somewhat gratifying to see this idea that since a lot of the world’s transactions deal with personal information and since a lot of the world’s networks really need that sense of independence of operation, even the chains like with Ethereum, the conversation of plasma sub-chains, a lot of that is about branching things off into B.F.T. or other consensus mechanisms that then report back to the main network.
In fact, perhaps even the Lightning Network could be thought of as a consortium network in that way where a lot of transactions happen off chain and then come back and get settled on chain, right? This seems like a pretty natural thing. And combining that with the idea that actually there is a fair degree of tokenization that happens on these private networks. I’d actually say when you have a supply chain traceability network where title to that diamond or title to that bag of rice is being tracked, not just metadata about it but actual title to it, you’ve essentially tokenized that diamond or bag of rice. You’ve tokenized a house title if you’re tracking that. You’ve tokenized an insurance claim to some degree that probably not a very fungible token but you’re still moving it around.
I think this idea is these are both about tokens and moving around. One of them obviously more about public tokens and tokens that with a trust in them doesn’t have to be tied to something real-world. But even on the private ledger, you can have tokenization of insurance processes, other types of things. I think we’ll see lots of hybrid approaches. I think we’ll see most of the world’s transactions, especially those dealing with personally identifiable information even if encrypted taking place on consortium ledgers where you can implement things like the G.D.P.R., the right to be forgotten, or certain other legal agreements between parties around data sharing and data reuse, which you can’t really do on a public ledger in the same way.
But then, a lot of things settling out or netting out to one or more public networks. And I think this is the other key thing is I don’t see these as a hierarchy, so much as I see them as islands connected by many bridges. And a lot of those might settle out to the Ethereum network, but there might be other networks in parallel that these processes either secondarily settle out to or settle out to instead of settling out to the Ethereum network. One of those could be a stable coin network run by central banks in a country whose tokens or coins are considered highly desirable like the U.S. dollar, or the euro, or the renminbi, so we’ll see.
Brian Fabian Crain: One of the things I would love to discuss with you briefly is, when you came on the last time, it’s 2016, if you think to back of the period between, I don’t know, maybe 2014, ‘15 to ’16, I guess, there was really a lot of interest in enterprise blockchain and enterprise blockchain companies were raising these massive funding rounds. The company is like a digital asset, or chain, or R3, or even Blockstream back then, I think, was basically talking about like enterprise blockchain use cases and they raised many cases, 50 million, close to 100 million, or even more in funding, that there seem to be so much interest in that. And now in the last two years or in the last, I would say, a year and a half, there’s been very, very little funding, I mean there’s been some, I think there was like Clearmatics and another one but very little. So what’s your take on that? Is it just that maybe these enterprise blockchain use cases are taking off but it’s hard to build like large businesses around it? Or like what do you think is going on?
Brian Behlendorf: You’re asking about investor interest and irrational exuberance and I mean I’ve been in the enterprise software space in one way or another for most of my life, sometimes as a consumer like when I was at the World Economic Forum. But most of the time I’m trying to sell this kind of concept whether I open source or not. And the enterprise sales cycles are very long, convincing companies to change how a core process works especially one that touches your core system of record and asking them actually to trust something other than the database they can put their hands around or virtually put their hands around in a cloud, is a really hard thing to do. And so I never saw this as a 18-month like, get in and have a big win and then go out and set off in the sunset. And I think to some degree the sense of . . . especially some of the early I.C.O.s set this expectation that this was more explosive in terms of returns than otherwise it was.
Enterprise sales, software sales is hard and sometimes it takes a new wave of investors to discover that all over again for like the world to reconcile. But I think this is kind of like that saying, things sometimes tend to get overestimated in the short term and underestimated in the long term. I don’t have any doubt that in the long term a lot of those bets will pay off. I was reading recently about how the operators of the Australian Stock Exchange who’ve brought in Blythe and Digital Asset to come in and reinvent how they do their core transaction systems that that market will see savings in the tens of billions of dollars a year from the deployment of a technology. Like this simply in cutting fees and then cutting overhead and cost in processing what they do. So I don’t have any doubt that in the long term the net of this will pay off for a lot of different parties.
But the other thing to keep in mind is that it very often in infrastructure investment waves the amount of total return to the infrastructure layer doesn’t pay off investment. So when the railroads were being invested in the latter half of the 19th century more money went in to building railroad startups, the Pittsburgh to Cincinnati railroad line and all these companies. There’s a Wikipedia article about this which is fascinating. More investment went in to railroads than ever came out in all the time since the 19th century as an investment category it has been a loose for most people who’ve gone in and tried. I’m sure some people made money in fact the railroad barons made sure that they capitalized quite a bit on it. But for the average investor it did not pay out. And yet no one would doubt that the railroads made a tremendous amount of money for people on top of that infrastructure, for people should have been goods across it.
It’s just that they managed to capture the bulk of the value created by this technology. And then that’s perhaps as it should be, maybe plumbing isn’t the right place to try to capture the majority of the innovation. Maybe plumbing should be commodity and maybe it should be something that evolves and grows and goes through revolutions and that’s what we’re seeing now. But the real money is in end user apps. And I think that always has been the case. Google didn’t make money by building a better search engine. They made money by taking that traffic that came to the search engine and sending it off to advertisers and others. And I think that’s something to keep in mind in these investment cycles.
And I know, I’ve seen the curve for I.C.O.s and kind of investment in I.C.O.s over the last year. I haven’t seen a similar curve for investment in enterprise startups that use blockchain in some way. It would not surprise me if fewer companies pitching for VC rounds use the V word in their one or first page a kind of intro or top-line explanation. Because I think people do realize this isn’t magic pixie dust that takes a bad business model and makes it better. We’re past that point and thankfully so.
Sebastien Couture: Now it was interesting how you phrase this regarding the plumbing not making the money because this is exactly the opposite of I think what was the kind of the going pieces that many VC’s had and nothing that drove many investments by VC’s in the I.C.O. space. And of course at Union Square Ventures, they were kind of famous at that point for this fat protocol thesis where it is exactly the idea that you put the token into the plumbing and then actually the plumbing is what will capture the most value. So I take it that’s not a view you share.
Brian Behlendorf: I can’t say I read that article and felt like I agreed with the main thesis. I didn’t like it because none of the previous technology movements ever had a situation where the hundredth person to pick up the technology made out with huge amounts of value over the millions of person to pick up the technology. It’s like really great infrastructure in my opinion, creates increasing returns, the more people use it rather than decreasing returns. If the hundredth person to buy a Bitcoin made out much better than the millions person to buy a Bitcoin and I mean that’s where I didn’t feel like protocols built that way would tend to lead to mass adoption in the same way that free protocols would.
Sebastien Couture: I’m of the opinion that these companies that went out and raised tens of millions and perhaps even over a hundred million dollars on the premise that they were going to build enterprise infrastructure with blockchain, I’m of the opinion that the reason why some of this has fizzled out is not so much because the sale site goes along, but that the idea that which was initially pitch to investors perhaps didn’t really pan out. And I speak also from experience because this was my personal experience with the company my previously was with Swift and that many of these companies went into the space with sort of a shotgun approach and tried to hit every single blockchain use case, which was floating around the time and tried to provide some sort of solution to that. And in most cases weren’t able to.
And as soon as this lack of a product, this lack of a clear vision and a product that has been the reason for a lot of these companies not really becoming this large infrastructure place that they were meant to be. I’d like you to maybe respond to that and get your other opinions about what do you think about this.
Brian Behlendorf: I think for a lot of people they saw the Ethereum I.C.O. and said, “I want to reproduce that.” And then more credibly they look at A.W.S. And A.W.S. is probably the strongest argument somebody could come to against the idea that nobody makes money in plumbing. And that A.W.S. makes $20 billion a year if I recall correctly in plumbing and being that kind of in pay for service infrastructure. And if arguably there’s a business in being the decentralized A.W.S. out there which I know is behind some of the folks who are building some of these systems. And I don’t think that’s necessarily an incorrect goal or desire. Or it was on its face incorrect. I think there is such a thing as well as timing.
There are times when I feel like one of the company I started in 1998 called CollabNet, which I built the subversion open-source tool but also really popularized the idea of making open-source software development a regular practice within enterprises and inside of enterprises, between enterprises as well as kind of public facing stuff. I look at things like GitHub selling for an obscene amount of money to a Microsoft which is great for them, great for Microsoft, great for GitHub, great for everybody, tons of hard work. But there’s no doubt I look at that as a little bit like, “Well, were we just two generations too early.” And frankly had I started CollabNet about six years later I would have definitely gone in a more a cloud friendly multi-tenant.
And there’s a whole bunch of like things I would have done differently. And I think there is such a thing as simply being too early not being wrong. And I’m willing to concede that a lot of the businesses that have raised funds or raise funds of the right premise, the right long-term idea. And if they have enough runway to survive, if they cashed out of their tokens at the peak and have some money to survive the next five or ten years, they may yet find that the technology based and the market comes around to decentralized A.W.S., to a decentralized answer to a lot of these things that does create some value for their tokens or for your enterprise value in the business that they build around it.
I just think that this wasn’t a magic money machine that allowed us all printer and money and therefore be infinitely wealthy without any accountability or without any need to actually create value in the ecosystem. That’s where I think this irrational exuberance came from in the market. And we’ve seen the appropriate correction for that.
Sebastien Couture: I mean I wasn’t speaking specifically about companies that went and raised I.C.O.s. I was actually talking about companies that raised with VCs and companies like the ones we mentioned previously and not specifically I.C.O.s or token models.
Brian Behlendorf: Well, I think it’s true no matter how you raise your money that it is possible to be cheerleading to a market. But I also think a lot of the companies you mentioned are not dead yet I mean Chain sold to Stellar and is focusing now on Stellar oriented applications for enterprises. Digital Asset is signing up new customers and moving forward with their project to the ASX. These companies don’t seem to be fly-by-night operations. These seem to be ones who are building a business. And it takes longer than perhaps some people thought it’s the enterprise sales cycle but squared because now you’re talking about N number of enterprises having to interlink their systems together before you see the real value come out of it.
So I’m not just trying to be diplomatic here. I genuinely feel, and I know you want something spicy but no. This is hard work. And what we focus on at Hyperledger is how do we at least avoid duplication of effort. How do we at least encourage those companies to build on common technologies, so that they just have to spend less time doing the same thing over and over? So they can move up the static as quickly as they can. And that’s resonating with a lot of startups in this space and big companies and others.
Brian Fabian Crain: I wanted to speak a little bit about what is your ultimate vision for blockchain. So what are the things that you hope 20 years from now that the impact blockchain has had on the world? And do you also see kind of possible trajectories that maybe end up being like not a good place? What are your thoughts on that?
Brian Behlendorf: Yeah, I think be a fun game to go around and ask people, “What’s your favorite dystopian movie?” And I’m 45, so it’s movies from the 80s. And my favorite dystopia isn’t 1984 its Brazil, the Terry Gilliam movie. I don’t know if any of you saw it. But I mean it’s a comedy. It’s kind of Monty Python asking humor but really that’s a society that is wrecked. It’s very technological but the technology is not very evenly distributed. It’s kind of retro futuristic in a way too off the typewriters everywhere. But one of the most dystopian things is about bugs in the system, in some cases literal bugs I don’t want to give anything away but bugs that cause the wrong person to be rendered to a prison somewhere rather than actually treated fairly.
And I worry, we’re definitely getting more digitized as a society and that’s going to create a lot of good things. But the risk is if you get your papers slightly out of whack, if there’s a typo somewhere or more importantly if there’s a process that’s not being followed that should be followed it’s either going to result in somebody unfairly getting maligned or someone unfairly going to prison or disappearing from the planet. Or most worryingly this creates even more opportunity for bad actors to get away with bad actions. And that we can debate for a long time about who’s a good actor, who’s a bad actor. We can debate for a long time about civil liberties.
My positive hope and the reason why I’m spending more time in this is I feel like decentralized systems even if we don’t have that pure idea and crypto anarchist kind of ideal of central assistance but instead we have more consortium networks, where the parties keep each other in check in a meaningful way, and where there’s enough diversity in the network to avoid large actors from being able to get away with that actions. Then we have a much more fair basis for society, much more auditable basis for society than we would otherwise have if current technology trends lead us to where it seemed to be leaving us otherwise. My dystopian fear is that we never get past the bugs that as complicated as these systems get we introduce more and more error and we simply drown in the complexity of it all.
And complexity, it ultimately serves the interests of parties that can afford to manage that complexity, who can afford the armies of programmers and lawyers and others to buffer them from the reality on the ground. It doesn’t speak well to startups. It doesn’t speak well to innovators. And then my worries if you counter that with governance models that are too automated and too driven around a model of consensus that says majority rule or super majority rule, then we forget that the appropriate role for civil rights in a society is to protect the minority, to protect the individual against the ravages of the mob, the ravages of the horde. And so that’s what I wrestle with especially with a lot of the desire to automate functions in governance that come from a good place, I think saying that there is wisdom in crowds.
But there’s also madness in crowds. And I don’t know we have enough protection against that. That would be my fear is that blockchain technology gets weaponized as it seems like almost every other internet technology has. But I think if we plant the right ideas out there, get the right initial project. That’s why we’re working a lot with nonprofits and government agencies to try to hopefully get the right kinds of technologies used. And at the end of the day this is why open-source still matters with access to the machinery of these tools we can figure out what’s wrong with them, we can fix them, we can subvert them, we can build new ones over and over again. And that access to open-source code. And my book is far more important than access to a network.
Sebastien Couture: So before we wrap up I do want to talk about the community aspect. You’ve mentioned in the last podcast about the importance of community and how important that was to Apache early days, talk about the Hyperledger community and how it’s grown? And who are some of the unexpected actors that are showing up there?
Brian Behlendorf: We have a real community of communities. There’s not only the 11 different software developer communities around each of those code bases. We have cross-cutting kind of working groups as well around performance and scalability and architecture and identity. And we have people participating in those cross-cutting working groups even if they’re not directly on the code base because they’re bringing a level of expertise and a level of willingness to create content around white papers that sort of thing that’s really important. And a lot of that activity bubbles up to this kind of weekly gathering that we have every Thursday morning at 7:00 AM Pacific got the technical steering committee call. And those meetings are public, anybody can join, the recordings are posted, minutes are posted.
And that’s where really the core governance of this ever growing community is managed, where new projects get approved, where we review the performance of existing communities, we try to answer questions and concerns about things. And the culture there it’s really strong. And I’m really happy to see how strong the commitment is to transparency in that community. But we have unevenness out there in terms of projects and how active they are and such and something that my staff continues to focus on. And then secondly we’ve now got this growing end user community, which were growing in a number of ways. But the most important is setting up these sector-specific working groups. We started with healthcare pretty early on and now we’ve added working groups in the public sector and the social impact space.
And we’re about to launch one in trade finance as well and looking at other areas. And this is basically a way for us to meet users in those communities with the technologies and the developers kind of working on it to kind of like make sure we’re speaking the same language to look for interesting use cases, sometimes it leads to new code. So we have this thing called Hyperledger labs I forgot to mention which is a place where experiments and templates and documentation and other kind of things can sit without having to be full-fledged projects. And there we’ve got a number of projects that have come out of these working groups.
The healthcare one for example spawned a demo application around breast milk supply chain tracing which I guess is a big issue is like a way to explore these questions about like the HIPAA for example, the healthcare information and privacy act basically ways to kind of see, how would somebody building a system like that walk all these different balancing acts. So those communities on those working groups are growing as well leading to participants like the ones I expect the businesses in those sectors, but also ones that we didn’t like a lot of nonprofits. I mentioned Kiva now announcing that they’re working with the government of Sierra Leone on those projects. The government of Bermuda recently joined and is looking at how do they take a lot of their business processes or we have the government of Dubai, government of Luxembourg, all of them are starting to have these kind of internal blockchain centers of excellence, blockchain expertise that they’re building.
And we’re bringing them in ways that I might not have expected two years ago to have that degree of interest and hands-on engagement. So that’s pretty fun. And a lot of these are coming together at the Hyperledger Global Forum. One thing I definitely want to make sure people know about thank you is we have an event coming up in Basel, Switzerland December 12th to the 15th called the Hyperledger Global Forum. We’re aiming for about 800 participants there. And we’ll have really a good mix of the software development side and kind of the technology what’s going on as well as the production and pilots and the end users and what they’re doing with the technology. Basel is really close to Zurich. Switzerland is lovely in December.
Lots of good skiing nearby and I would really encourage folks to make it out. It’s also surprisingly affordable especially if you’re based in Europe it’s easy to get to. We’d love to see more people there. So please if you have a chance Hyperledger Global Forum check it out.
Sebastien Couture: And I certainly look forward to being there. I’m going to be there doing some content on the show floor. So I look forward to seeing you there and hopefully reconnecting there again and seeing everybody from that community at the event. Thanks so much for coming on Brian. It was really fascinating and interesting to see how Hyperledger has grown and how that product and that community is evolving and looking forward to seeing what comes out of it in the future.
Brian Behlendorf: Thank you. Thank you for having me here.