“2019 will be a continued reiteration of how people are using blockchain”
Hyperledger Global Forum is behind us but now’s the perfect time for an overview of the most important announcements, plans, and takeaways. In this interview, we speak with Brian Behlendorf about what the future holds for blockchain, how certifications are changing the enterprise and the goals of Hyperledger in 2019.
JAXenter: Hyperledger really got its foot in the cloud door now that Fabric has the attention of all Cloud providers. What does this say about the maturity of Hyperledger and its commitment to the Cloud?
Brian Behlendorf: The project is just barely three years old – it first launched in December 2015. When it was launched, the code was still very research and development oriented. It was still very prototypical, very early stage. There was a clear message that this code would take some time to settle. We thought, let’s try to deploy it for some things and see what happens, but it was tagged with a “not safe for production use” at the very beginning.
We were still figuring out a lot of questions like, “How do we grow this community? How large of a community should we try to be? What standards do we want to follow? A lot of that congealed over the first year. Then, Fabric was our first production 1.0 release in the middle of 2017 – a year and a half in.
Now, most people don’t use a 1.0 product. You wouldn’t want to use Windows 1.0 or Linux 1.0. But for us, 1.0 says, “Here’s a big software that now developers are comfortable with people they’ve never met and they feel in control using this software in production. That doesn’t mean it’s done or perfect or bug-free, but it means there’s a level of confidence in that. There’s other vetting that we do. We have a third party security firm that comes in and scans the code for security vulnerabilities. We know you can’t catch them all, but we catch the low-hanging fruit.
There’s more room inside the Hyperledger greenhouse for many more projects. We want diversity at each level too. That’s why we have competing frameworks. It creates a bit of rivalry, but also a lot of collaboration.
With the 1.0 release, from that point forward, a lot of people started to use it and get it into production. Then you start to realize the next order of challenges, which is: “How do you make this easier to adopt or for somebody to turn on?” Especially since rarely in a blockchain application do you have only one company involved. Usually, you’ve got three companies, ten companies, a hundred companies. So, getting everybody to adopt a new technology is difficult. Ordinarily, it’s hard to get even one company to adopt one new technology. But to get an entire industry, or even a bootstrap number, you have to meet companies where they are. If it’s three banks starting a network, one might be very leading edge, but the other two might be more trailing edge.
And so, cloud adoption and cloud support for these technologies is incredibly important. It has to be easy for overworked, underpaid IT staff who don’t have the capacity to become experts in a new emerging technology to be able to stand up for their company’s interests. Many of these clouds including Amazon and Google, they’re hearing from their customers: “Do you support running Fabric?” Because they sell their access to virtual hardware essentially, customers could always run versions of Fabric on AWS. But now all those companies and all of the cloud providers now have branded, ported, first-class services around Fabric or Sawtooth. That’s a sign that our technologies are not just ready for early adopters, they’re ready for the broader market as well.
When somebody goes and builds a network for a banking sector, or a trading sector, or a supply chain, everybody can get on board, not just the early adopters.
Brian Behlendorf: We will look at that over time. Our goal is not to have a thousand projects, our goal is to have a portfolio that’s well-curated and that our projects work well with each other. There are upper limits to how many people you can have in a community before it gets too unwieldy to manage. But I could see a few more wings of the greenhouse. Maybe a wing dedicated to libraries or templates. We are probably not going to be building any end-user apps. I don’t think we really want to get into being a home for applications that only barely touch blockchain as a way to store and retrieve data. I would love to see projects outside of Hyperledger.
What’s in store I think are libraries, templates, and frameworks that make it easier for repeating patterns. For supply chain, traceability is a pattern that applies to lots of different sectors. Marketplaces and directories might be another pattern. Put all of these things together, and there’s more room inside the Hyperledger greenhouse for many more projects. We want diversity at each level too. That’s why we have competing frameworks. It creates a bit of rivalry, but also a lot of collaboration.
JAXenter: You also mentioned certifications in your keynote. What are the benefits of passing the exam? What can these certifications do for Hyperledger?
Brian Behlendorf: Certifications provide a professionalization of the space. We are helping people to understand that there is enough substance here to what we are building and enough that’s different. If you’re serious about building this into your core information system, you want to know that if you’re thinking of hiring someone and they claim to know how Hyperledger Fabric works that their claim is trustworthy. The training and test are intended to help convey that trust.
This is all about the mid-to-late adopters on board. The early adopters have their own process for vetting developers or are perhaps willing to pay for someone to receive training for a few months. Mid-to-late adopters need someone off the shelf who someone else has said that they are trustworthy and knowledgeable. In some cases, those late adopters will simply use cloud or cloud offerings and not have any staff. But there will be some that realize that they need a few staff with a direct connection who can audit and validate what’s going on.
My hope is that by being able to point to the thousands of developers who have that certification, it says that this is now a safer technology. Now there is a resiliency to the system. The certification for administrator of these technologies is what we are launching first. We will later look at certifications for other things, such as developer or for businesses offering blockchain as a service. How do you know that your instances or use of Fabric on one cloud provider is going to be the same as another, or that you can move it or add nodes for extra resiliency? All of that is important. The broader theme is professionalization.
By being able to point to the thousands of developers who have that certification, it says that this is now a safer technology. Now there is a resiliency to the system.
We’re not doing this as a revenue generator or to add new members. This is in our mission. It’s what we are here to do. This is what The Linux Foundation does in other places. Here, the Cloud Native Computing Foundation (home of Kubernetes) does exactly the same thing. They have a Kubernetes certified administrator course and certification exam, as well as Kubernetes certified providers and solutions providers. This is something that modern IT enterprises expect from a mature technology base.
So, how do we help blockchain technology go from this early Cambrian explosion of a lot of different ideas into something that’s a bit more predictable? I’ve been accused of making blockchain boring, and I’ll take that!
JAXenter: There is still a talent shortage, will the certifications help?
Brian Behlendorf: It certainly will help. We need more of that, for sure. Here is where you will learn how to go to the community to get help. It will teach you how to support yourself. Certificates will also teach you if you find a bug and how to fix it, how to create a pull request. It is equally important to help the broader marketplace as well as specialists.
We want to show people that our technology is usable and that there are people out there who can help get it set up.
Don’t use a blockchain because it’s a big data pool or because it’s fast. Only use a blockchain if you have an issue of trust. Then it is worth paying the penalty.
JAXenter: Are you planning on reaching out to developers beyond the blockchain world? If so, how?
Brian Behlendorf: I’ve been going to the Ethereum conferences each year without a marketing presence just to build personal connections and bring some of them closer. With other blockchain communities, I am letting them know how we work together. But I also spend a lot of time at other developer conferences, such as some of the Cloud and open source conferences by Linux. I give talks at those and help people who are all in very different places to see how our code works.
We want to do more outreach. Our marketing department wants to explain what this code is to other people. We do things such as have developer profiles on our website, as well as an active blog that we ask developers to write for. All of this hopes to encapsulate what is going on here into more digestible chunks that find their way well beyond the blockchain.
JAXenter: There are still some concerns surrounding blockchain. Scalability, for example, is one of them. There are a lot of questions about blockchain that haven’t been answered yet. What does it take for people to be confident in blockchain?
Brian Behlendorf: Normalization. We need more people working with this code and more stories about how people use it. That’s why it was great to have Aaron Symanski from Change Healthcare put a number down: 50 million transactions a day. That’s a signal to everybody else that if they design it right and do the right things they can build a system that can meet that number. Those numbers aren’t everything. There are a lot of different questions to answer about what the scalability actually needs in a blockchain network. It’s not just how many transactions you have, it’s also your resilience to security issues and the number of nodes on the network. It’s a little counter-intuitive, but the more nodes you have the slower the network gets.
There’s something to the art and science of blockchain design and application design at that layer that is probably going to be as unique as database design emerged after the SQL programming language came out. Once people started to work with relational databases, it became clear that there was quite a lot of architectural thinking about the app you want to build and how you model it in the database. Early databases were slow and you had to think about every byte and how they align. You had to think about the physical format of the hard drive to get the performance you needed when CPUs were much slower. If you’re a big database architect, you are constantly thinking about this. We are going to see a similar field of blockchain architecture. How do you squeeze the most performance while meeting all the business level objectives?
In the early days of the web, and even in the middle of the last decade, it was still the case that if you ran a website with a healthy amount of traffic you’d fall over. Sometimes you would go visit a site to see something really interesting, and it would often be down or not responding. It still happens here or there, but now we have things like content delivery networks and businesses to answer the question of, “What happens if I have too much of a good thing?” In the blockchain space, part of that architecture is how do we adapt what we’re doing when suddenly there’s ten times the amount of interest and traffic. I don’t think there’s ever a singular answer to the scalability problem.
How do we build skill sets that will have us answer the question: “How do I answer the next higher threshold of traffic?” It’s a mix of technology and business to answer that question.
JAXenter: Some people say that blockchain is the world’s slowest database right now.
Brian Behlendorf: Yes, and it’s not a singular slow database. But this is where we really need to get crisp about public blockchains and permissions. If you’re talking about Bitcoin or Ethereum, they can only do three or ten transactions a second. Even if they did three or ten million transactions a second, there would still be a really good reason not to run most of your transactions on the public ledger. Why make yourself subject to the activity of people you have no relationship with? It is definitely true, but you can always use a centralized database to solve any problem. It will always be faster and cheaper to do it that way.
Don’t use a blockchain because it’s a big data pool or because it’s fast. Only use a blockchain if you have an issue of trust. Then it is worth paying the penalty. There’s always a good question of bringing down the cost of that penalty. The more we bring it down, the more places it’ll be worth using. But there will always be a penalty over centralized and that’s because there’s value in decentralization.
JAXenter: What are your predictions for 2019 and what’s your New Year’s resolution for Hyperledger?
Brian Behlendorf: I think it will become clear to people that blockchains aren’t just about cryptocurrencies and that you can solve many of these problems that people are hoping to solve. We’ve been consistent on the message of where and how to build blockchains. Our signal has been growing and growing. The hype cycle of cryptocurrencies is arguably on a bit of decline and people are asking harder questions. I don’t think cryptocurrencies will ever go away, but I do think that this is the time that people see how they can use these technologies without having to worry about the market prices. 2019 will be a continued reiteration of how people are using blockchain and asking the harder questions about the value that they are getting from it.
Within Hyperledger, our resolution is about professionalization or normalization. It is a balancing act between enterprise software that doesn’t change a lot that people can trust, that gets better and smarter but still has room for disruptive new thinking.
My only New Year’s resolution is travel a lot less! I was on the road more than half of every day this year and circled a world a couple of times. I’d like to only circle it once or twice.