BT

Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ

Topics

Choose your language

InfoQ Homepage Articles The Current State of Blockchain - Panel Discussion (Part 2)

The Current State of Blockchain - Panel Discussion (Part 2)

Leia em Português

Key Takeaways

  • Enterprises are wary of vendor lock-in to particular solutions
  • This fear is fuelling the adoption of open-source blockchain platforms
  • Enterprise blockchains can be thought of as deploying middleware at a whole-market level rather than in individual companies
  • There is still a lot of hype in this space, and many projects are really only using a blockchain as window dressing
  • Caution and critical thinking is essential when navigating this space

UPDATE: This article was updated to correct a misattribution in the originally published version.

This is a second part of a panel discussion on the The current state of blockchain. Here you can read the first piece.

The final two panelists introduce themselves and share their views of the current state of the Blockchain world. We're joined by Richard Brown, CTO at R3 and David Gerard, journalist and author of "Attack of the 50 Foot Blockchain".

InfoQ: Please introduce yourself, and explain your connection to the space.

Brown: Hi. I'm Richard Brown, Chief Technology Officer at R3. We support and maintain Corda, an open source blockchain platform first designed for the financial sector but now finding usage way beyond that. Its design is a bit different to some other blockchains aimed at business in that it is not a fork of anything else and has a specific focus on privacy, ease of use (it runs on the JVM, for example), strong identity, deployability by real-world businesses and compatibility with the legal system.

Gerard: I'm David Gerard, author of the book "Attack of the 50 Foot Blockchain" and the news blog of the same name. I have an unrelated day job as a Unix system administrator, but I have what has turned into a second job as a finance journalist, specialising in this space.

InfoQ: How would you characterize the current state of the distributed ledger ecosystem? Is it, as some have suggested, "the end of the beginning", or is it something else?

Brown: On the "enterprise" side, I think we're at the end of the beginning. I've long predicted that the vast number of competing platforms will shake out very rapidly to a small number of winners - perhaps three or four - and that it will happen quicker than anybody expects. I was interviewed about this here: https://www.coindesk.com/r3s-hearn-and-brown-say-enterprise-blockchains-day-of-reckoning-is-here/ 

Fear of proprietary vendor lock-in given blockchains have such strong network effects, the reality that businesses will simply not put up with having to install ten incompatible platforms, and the reality that architecture _does_ matter makes this inevitable. It's why Corda is not only open source, but liberally so - Apache 2 licence, design and development done in public, all our design reviews published to YouTube, etc. 

There is still a lot of froth, however, and there still doesn't pass a day without a press release about an enterprise blockchain platform that, when you dig in, is entirely run by a single firm! Critical thinking is strongly advised in this sector.

Gerard: It's coming up to ten years of hype and unrealised promises, if not promises that have spectacularly failed. That's a rather long time for something of technical substance to be "coming soon".

There's still interest, and hence consulting income, to be made - even with a glaring lack of production systems, particularly ones realising any of the claimed advantages of "blockchain".

Most of the action is in the legislative and financial side - as cryptocurrency companies use claims of the technical advantages of "blockchain" to lobby for exceptions in law for their financial activities.

Technical claims are mostly unsubstantiated, and used as an excuse for questionable financial aims.

These systems are not about technology - they're about marketing hype. This is the overwhelmingly most important fact about them - the elephant in the room for any discussion of these things that claims to be technical, particularly from advocates. The hype, and the incentives for the hype, are the actual story; the technology is an excuse to send up some hype.

You can't really talk about this space in purely technical terms, because the technology is not new, and all the action is in the marketing.

InfoQ: What are the actual use cases for distributed ledgers? Are there specific markets and sectors that can make especially good use of them? How widely applicable are the use cases?

Brown: We spent SO MUCH time on this in the early days of R3. We began in 2015 as a consortium of banks. My mission when hired from IBM was effectively to lead a consulting project to answer some simple questions: "So this blockchain stuff - people _say_ it could be transformational to banks beyond its initial use-case of cryptocurrency. Is this really the case? If so, what is the relevance and what does it mean?"

I documented the essence of our findings here.

Aside from the obvious use-case of censorship-resistant digital currencies, the *other* really interesting thing about these new platforms is that they effectively create systems that are deployed _between_ multiple entities that don't trust each other yet enable them to form and maintain consensus about 'facts' they jointly care about.  In Bitcoin's case, the users of full nodes care about how many bitcoins there are and who owns them.  But you can _massively_ generalise that problem to _any_ facts that _any_ parties care about. 

When you look at it like that, you get this kind of dawning sense of excitement.  "Hang on... isn't this the problem that every business in the world struggles to solve? If you and I enter into a contract, you record it in your systems, I record it in mine... and then we then spend the rest of our lives dealing with the problems that arise because our computers are running different software and we don't both quite have the same view of the world?"

So what we concluded was: Yes... this idea of deploying software at the level of a _market_ rather than at the level of a _firm_ to automate or manage business processes that operate between the participants in that market could be world-changing. Aside from the cost savings from removal of duplicated systems, the increase in data quality and certainty could unleash opportunities we can't even imagine.

Now, of course, you *could* also solve this problem by running a centralised database. And, indeed, that's what some markets have done.  But it creates a massive source of risk, a monopolistic provider that then needs to be regulated. 

The lesson of history is that, outside places like the financial markets, people are highly resistant to creating these centralised sources of power. The end result is that you end up in this perverse situation where every _company_ has optimised the hell out its own operations but the processes that operate *between* firms are a joke. I wrote some more about it here.

THIS is the sweetspot for blockchain platforms like Corda.

Gerard: It may sound cynical - but the only "actual use case" that it has consistently achieved is marketing, and selling tokens.

The terms "blockchain" or "distributed ledger" in this sense originated as euphemisms for Bitcoin - they are marketing buzzwords, designed to sell a particular package of hype.

There is no agreed-upon technical definition of what a "blockchain" is. It usually has a Merkle tree in it somewhere, but that's about it.

Merkle trees have had all sorts of use cases since 1979. If you consider a "blockchain" or "distributed ledger" to be a Merkle tree with a consensus mechanism added - though there are quite successful products marketed as blockchain that just have the Merkle tree, e.g. Guardtime's KSI Blockchain - then I have seen no technical use case where a "blockchain" is the best fit for a job, except Bitcoin-style cryptocurrencies, insofar as they're a use case.

There are endless claims to this effect, which are almost always supported by citing other claims to this effect. You will see plenty of nontechnical white papers citing other nontechnical white papers, and hypothetical "could" statements transmuting into present-tense
"is" statements - whether or not software exists that fulfills the promise at all.

There are also projects which claim to solve the "blockchain trilemma" of decentralization, security and scalability, without using a Bitcoin-style proof-of-work mechanism - distributed hash trees, IOTA's "Tangle", HashGraph, etc. None of these have so far produced code demonstrating success; many appear just to use several pages of abstruse LaTeX as an excuse for aggressive marketing of ICO tokens.

In all cases, hypothetical future claims should be assumed worthless until a production system that stands up to testing shows itself.

InfoQ: What sense do you have of real projects making use of distributed ledgers and the ideas behind them? How many actual production systems are there out there based upon distributed ledgers?

Brown: As the CTO of R3, I have the privilege - and stress - of seeing a huge number of real-world projects in close-up. Here are just some examples from the top of my head:

  • Syndicated lending is a highly decentralised market (I wrote about it in the link above ). Finastra, one of the world's largest financial software firms, is using Corda to automate the sharing of data between participants with no central party, no centralised database, not change to the structure of the market. They're in live pilot and will go into full production soon.
  • Another good example is "Marco Polo", an initiative to make the financing of world trade far more efficient. Again, it's another example of a multi-firm, international, borderline intractable problem - where business could be so much more efficient if only everybody had a common view of the "truth", but where it is _never_ going to be acceptable to set up some centralised body to run it. So using a platform like Corda, where data is only shared with those who need it but those participants are nevertheless assured that "what I see is what you see" is a genuine step forward.
  • At the other end of the spectrum and outside banking, the world's insurance industry appears to be standardising on Corda and we are seeing examples in healthcare, oil and gas, even land registries.  Tradewind Markets (gold), GuildOne (oil and gas), Insurwave (hull insurance) are live. And there are lots of others.

To get an idea of the range of problems being solved with Corda, take a look at the videos from the recent CordaCon event, where members of the community presented on what they were doing. (Technical videosBusiness videos or browse marketplace.r3.com

Gerard: The very closest to a successful real-world use case for something that's a bit "distributed ledger"-like is git - where you have a Merkle tree with a chain of transactions (commits), identified by the hash of the last commit, where developers routinely sling around entire repositories, and the "mining" is done by a human maintainer, picking diffs to commit to their tree.

Many real-world systems branded "blockchain" are a simplified git at best.

There are real-world systems which loudly advertise their "blockchain" - but it turns out the blockchain is just being used as a centrally-administered database. Examples include the Walmart / IBM supply chain proposal (all nodes administered by Walmart and living on the IBM Cloud), the Maersk / IBM TradeLens system (similarly, with Maersk administering) and the World Food Programme's refugee funding disbursement programme (which uses a private single-user Ethereum instance as the back-end).

All of these claim the benefits of the "blockchain", but in no case does a blockchain bring anything to the party technically.

InfoQ: Thinking specifically about the distinction between public systems (such as Ethereum or Bitcoin) vs private, permissioned systems what do you think the balance is between them? Are projects preferring to use public or private ledgers? What is driving that distinction, if it exists?

Brown: There's a strange phenomenon unfolding.  Public blockchains, through initiatives like the behind-closed-door Enterprise Ethereum Alliance, are trying to adapt themselves to solve business problems.

Identity / privacy-focused blockchain platforms such as Corda are gaining capabilities to manage tokens and facilitate networks of huge scale. We recently released a sample on how to use Corda to issue real asset-backed tokens for example.

So it's as if we're in a race to the same centre ground. 

However, my bet (perhaps contrarian) is that the winning platforms will be those that started by solving problems of business. Corda's foundation of privacy by design, settlement finality, strong identity layer, support for Java, etc, is the perfect basis for the wave of regulatory compliant 'enterprise tokens' now being issued, for example. Whereas, on the other side, platforms that have only probabilistic finality and rely on arcane programming languages (eg solidity) have a much harder sell on their hands.

Gerard: Projects that are fundamentally about selling ICO tokens are using the public Ethereum chain. Projects that are intended as marketing for the concept of "blockchain" in business are using a private ledger.

InfoQ: Some technologies represent a minor enhancement to the state of computing and some represent a true sea change. For example, we might characterize Complex Event Processing (CEP) as an example of the former, and the arrival of Hadoop heralding general purpose Big Data processing as an example of the latter. Thinking about the technology in those terms, where do you see distributed ledgers as sitting along that spectrum? Can you point to any indicators in the market that support your position?

Brown: I guess time will tell. When I was at IBM I spent a lot of time on both CEP and Big Data and they both seemed like a huge deal up-front and yet, as you say, only one of them changed the world.

My money is that platforms like Corda *will* have lasting impact, however. You'd expect me to say that, of course. Part of the reason is that, despite the hype of blockchain, when you scratch under the surface, what I sometimes think we're building at R3 is industry-level middleware.

In other words, imagine you were tasked with writing an application that could be deployed by all participants in an industry such that each had their own copy of data that mattered to them (no centralised point of control required) and where they were guaranteed to be in sync (eventually).  What would you need?

  • App server capabilities. A nice environment to program against. Corda runs on the JVM, so if you know Java or Kotlin you can get going immediately. We get complimented on an almost daily basis about how much people enjoy developing for Corda. Developers love it.
  • The ability to route data based on real names - real legal identities - not queue names or JNDI names: "Send this transaction to Barclays head office!". Corda's identity layer and legal-entity-based-addressing gives you this out of the box.
  • Inter-firm workflows. a way to encode how data should flow and to whom.  The Corda Flow Framework gives you this.
  • A "trust but verify" transaction processing model. Don't take on trust what comes from counterparts but rerun the business logic you pre-agreed would govern your interactions and only accept the transaction if it checks out - the key to knowing that what you see truly is what your counterparts see.

And so on.

Looked at this way, it may be that Corda's lasting legacy is in defining a new era of market-level app development platforms, and it happens that techniques learned from blockchains is what allowed us to build it! I wrote about this here.

I should say that this isn't the official R3 view... just a belief that bounces around my head from time to time.

Gerard: A technological comparison would be software that does one trick well, from existing components.

I would compare the Bitcoin-style blockchain - a Merkle tree with a consensus mechanism - to BitTorrent, Tor, git and Freenet. BitTorrent and git are hugely popular and successful in their niches; Tor is successful with its smaller user base; Freenet largely failed.

Bitcoin had extensive aspirations, but its only present use case is as a tradeable commodity of no other utility.

Distributed ledgers have so far not supplied a clear technological advance. What they have supplied is an excuse for extensive hype, and remarkable claims about what software might do in the unspecified future.

About the Panelists

Richard Brown is Chief Technology Officer and Managing Director at R3. Richard's team builds Corda, the world's most advanced enterprise blockchain platform. R3 is an enterprise software company supported by a consortium of hundreds of banks, technology firms, regulators, trade associations and professional services firms. Richard was formerly Executive Architect for Industry Innovation and Business Development for IBM's Banking and Financial Markets business in the UK.

 

David Gerard is a Unix system administrator, an award-winning music journalist, and has blogged about music at Rocknerd.co.uk since 2001. He is a volunteer spokesman for Wikipedia, and is on the board of the RationalMedia Foundation, host of skeptical wiki RationalWiki.org. His website is davidgerard.co.uk. He lives in east London with his spouse Arkady and their daughter.

Rate this Article

Adoption
Style

BT