What are distributed ledgers (actually) good for?
- 1. Why “blockchains” are hard What are they (actually) good for? Sebastien Meunier
- 2. Fake news, fake news everywhere…
- 3. Those headlines feel like 2018 – Little happened in 3 years Sept. 24, 2014 CBW Bank and Cross River Bank announce Ripple Protocol Integration May 11, 2015 Nasdaq OMX embraces the blockchain June 9, 2015 Westpac, ANZ trial Ripple payments Oct. 25, 2015 ASX considers blockchain for clearing and settlement Dec. 31, 2015 Nasdaq Linq claims to have issued first securities over a blockchain Sept. 6, 2016 Ornua Completes World First Blockchain Trade Through Collaboration with Barclays Sept. 27, 2016 Credit Suisse, Ipreo, Symbiont, R3 Convene Proof of Concept for Syndicated Loans Oct. 24, 2016 Major banks mark first-ever international trade using blockchain tech https://ripple.com/insights/ripple-labs-signs-first-two-us-banks/ https://www.finextra.com/newsarticle/27342/nasdaq-omx-embraces-the-blockchain https://www.afr.com/business/banking-and-finance/westpac-anz-trial-ripple-payments-but-big-four-reluctant-on-bitcoin-20150605-ghhmsq https://www.smh.com.au/business/banking-and-finance/asx-considers-blockchain-for-clearing-and-settlement-20151025-gkhs46.html https://bravenewcoin.com/insights/nasdaq-linq-claims-to-have-issued-first-securities-over-a-blockchain https://www.ornua.com/ornua-completes-world-first-blockchain-trade-through-collaboration-with-barclays/ https://www.credit-suisse.com/corporate/en/articles/media-releases/blockchain-demonstration-shows-potential-loan-market-improvements-201609.html https://www.reuters.com/article/us-australia-tech-banks-idUSKCN12O0DX
- 4. McKinsey: from a “revolution” to “it’s very complex, a niche” 2015 “Blockchains have the potential to dramatically reshape the capital markets industry, with significant impact on business models, reductions in risk and savings of cost and capital” 2016 How blockchains could change the world: “why blockchains, the technology underpinning the cryptocurrency, have the potential to revolutionize the world economy” (interview) 2019 Blockchain’s Occam problem: “Blockchain has yet to become the game-changer some expected. A key to finding the value is to apply the technology only when it is the simplest solution available” 2018 “Blockchain beyond the hype: What is the strategic business value? Companies can determine whether they should invest in blockchain by focusing on specific use cases and their market position” 2017 The promise of blockchain: “Many have likened the revolutionary possibilities of blockchain technology to those of the internet, such is its perceived capacity to transform the ways in which people and businesses cooperate” ? ?
- 5. The shift of the narrative about “blockchains” 2009-2014 Bitcoin is internet money for drug dealers 2015 Blockchain not Bitcoin 2016 Blockchain will cure cancer and save the world 2017 ICO frenzy “Let’s raise $50M with some fancy words in a whitepaper” 2019 let’s institutionalize crypto (STO not ICO) Let’s leverage POW public blockchains (internet not intranet) 2018 Coins are failing (either scams or deliver nothing) Enterprise Blockchains are not interesting ? ? 2017-2018 Blockchain as a Service (IBM, MS, HP, AWS, Baidu, SAP…)
- 6. What is really new in Bitcoin? Peer to peer protocol Distributed databases Hash trees, verifiable timestamps Proof of Work Immutability of the data Champine, 1977 Johnson & Thomas 1975 Merkle, 1980 Dwork & Naor, 1992 Driscoll, Sarnak, Sleator & Tarjan, 1986 Those features are not *properties* of Bitcoin They are *design requirements* of the system, they serve a purpose Byzantine Fault Tolerant protocol Lamport, 1989 Cryptographically enabled digital cash Chaum, 1983 ARPANet 1969, Web 1990, Bitcoin 2009
- 7. Which data structure is simpler & more efficient? You must have a very good reason to use a blockchain data structure! 1 2
- 8. What is a distributed ledger? • A distributed ledger is a distributed system of records that leverages cryptography to provide a decentralized control of the read/write access to independent participants. Consensus between participants on the state of records is reached algorithmically, without requiring an external third party. • Distributed Ledgers generally use a blockchain data structure as part of their design, though other data structures such a Directed Acyclic Graphs may be used. • A blockchain is a data structure – literally a chain of blocks
- 9. What is a distributed ledger? (red pill version)
- 10. By definition, a DL is not controlled by a single entity If a distributed system is centrally controlled, it is a traditional distributed database system, whether it uses a blockchain data structure or not (Rauchs, M et al. 2018)
- 11. 5 truths about Distributed Ledgers Distributed Ledgers do not get rid of intermediaries You’re only replacing a software by another software, a fee by a fee… Question: how decentralized is the governance of intermediaries? Distributed Ledgers do not remove the need for trust Using Ethereum or Amazon.com is the same for the average user Core is secure; clients (wallets) and exchanges are hacked The physical world destroys the properties of DLs Third parties are needed to certify the data link physical<->digital Forget about supply chains where the problem is cheating humans! Smart contracts aren’t contracts and aren’t smart Not contracts in the legal sense, but simply computer programs They execute their program as written (even the bugs) Distributed Ledgers do not reduce friction & costs Data written once Vs n times; Data verified once Vs n times; Data communicated once Vs n times, etc.
- 12. It’s always a question of trade-offs There is no cheap scalable risk-free general purpose technology to decentralize all the things (sorry) Abadi, J., & Brunnermeier, M. (2018) Slepak, G., & Petrova, A. (2018)
- 13. What are distributed ledgers good for? (conceptually) Distributed Ledgers enable the decentralized management of scarce digital assets in adversarial environments More generally, they enable the algorithmic enforcement of software programs without possibility of interference by a third party They bring two main benefits compared to traditional distributed systems: resistance to censorship and empowerment of users, who literally own the key to their data They come with disadvantages: lower performance, higher costs, and a transfer of risks to the users
- 14. Practical problems Distributed Ledgers lose their unique properties when they are linked to external systems (esp. the physical world), or when their governance is re-centralized We are still looking for some real world use cases: is there a demand? – In controlled environments (private / permissioned) there are simpler ways to verify transactions ordering – When exactly do you need to circumvent trusted third parties? – When can’t you form a third party operator? – Who needs decentralized applications, anyway?
- 15. So what are distributed ledgers good for? (practically, in an institutional context) For virtue signaling: everybody has to pretend doing some “blockchain” For education: people re-discover decades old concepts such as money, game theory, cryptography, distributed systems and governance, like they were invented yesterday To generate discussions about digitalization in general and the need for cross-industry standards They allow to strengthen accountability along business processes (who did what and when). Some people call that “blockchain” or “internet of value”, when it’s really about cryptographic assurance (digital timestamping & signatures)
- 16. In a nutshell Censorship resistance Subsidiarity of responsibilities and risks Limited throughput High overhead costs Vulnerability to censorship Concentration of responsibilities and risks High throughput Cost efficiency Public POW Distributed ledgers Centralized Ledgers Enterprise/Permissioned /Private Distributed ledgers Protection against censorship/expropriation Revolutionary design but complex & costly Only narrow use cases, when participants don’t trust central operators Cryptographic assurance (accountability) Incremental impact affordable but complex Only narrow use cases, when participants can’t form central operators