TradFi's Cultural Bias to Permissioned Chains and the Slow Move to a Decentralized Financial Network

Whenever an innovative technology is introduced, it faces steep barriers to adoption.?There can be technical barriers, but often the toughest one to overcome is a cultural bias by the industry group or users of the old technology. ?Barbara Oakley writes brilliantly on the complex relationship between social behavior, neuroscience, and organizational design.?My favorite article by Dr. Oakley is entitled Concepts and implications of altruism bias and pathological altruism – I strongly recommend it (it is a quick read).?Oakley defines pathological altruism as acts by “a person who sincerely engages in what he or she intends to be altruistic acts” but the acts result “in objectively foreseeable and unreasonable harm to the self, to the target of the altruism, or to others beyond the target.”?She points out that “altruism bias may be one of the most pernicious, hard-to-eradicate biases in science, because it involves an even-handed examination of what groups of seemingly objective rational scientists subliminally have come to regard as sacred.”

Oakley cites examples in neuroscience and PTSD treatment, where scientists were ostracized from their scientific fields for questioning strongly held beliefs about the benefits of certain approaches.?I suggest that one of the pushes towards permissioned chains is in fact an instance of altruism bias in the field of computer science and, more specifically, in institutions seeking the benefit of a decentralized and networked market solution but centralizing control and neutering the sought-after benefit.?Oakley warns that “few wish to run the gauntlet of criticizing poorly conducted, highly subjective ‘science’ which is purported to help, or indeed, of daring to question the basis of problematic scientific paradigms that arise in part from good intentions,” but I think we need to relook at how we utilize permissioned chains, understand their value, and identify their shortcomings.?In particular, we need to facilitate permissioned environments to have on-ramps and off-ramps to permissionless, decentralized network solutions for finance.?Many large financial corporations seeking to move to a blockchain solution thought they were buying the Internet when they went to a permissioned chain, but then realized they had bought a Local Area Network.

Permissioned Blockchains?

Two weeks ago, a number of Inveniam folks were in Singapore to attend Digital Assets Week and had an excellent time.?We learned more about the major work performed in Tokenized Securities and Digital Assets by companies like ADDX , BondbloX , InvestaX , 星展银行 汇丰 , and Hamilton Lane .?One aspect that struck me was the prevalence of IBM’s Hyperledger Fabric, a permissioned closed system.?Moving to a broader permissionless environment and enabling synchronization across chains would grow the market, would aggregate liquidity across silos globally to clear trades, and would enable the 24/7/365 trading of private market assets as we see in #BTC.

If the benefits of permissionless, fully decentralized environments are so significant, I wonder why companies like some of the ones mentioned above forego these benefits in favor of a centralized version of decentralized technology, i.e., permissioned chains.?I used to think that these companies were “kicking the can down the road” to cover up the difficult problem of custody and wallets, because there are still no great institutional custody or wallet solutions for tokenized securities ( Fireblocks , Anchorage Digital , Ledger , HUB , and @Tangeny are few of the folks working on delivering a world-class solution).?This explanation would be in line with the technical barriers I mentioned at the very beginning.?However, I am gradually realizing that the very core concepts that computer scientists have been taught may have created an altruism bias against the concept of a decentralized solution.

To replace a traditional Central Securities Depository (CSD) with a digital Central Securities Depository based on blockchain, one would have to accept two premises:

1.????A distributed ledger as a data structure has the ability to statically represent all data aspects of dematerialized (book-entry) securities that a traditional database can represent; and

2.????A distributed ledger can enable the dynamic execution of ACID transactions (more on this term later) against the CSD just as a traditional database controlled centrally by a single organization does.

In what follows, I delve into these two requirements and the associated beliefs.

Ledger as a Data Structure for a CSD

Many people think that the concept of representing data as an ever-growing, append-only ledger of transaction entries is somehow novel.?Nothing could be further away from the truth, as the following examples illustrate.

·??????We all know people who use their email as a ledger / log-based database: they keep all their email messages forever (messages can be considered transactions and the mailbox can be considered a serial ledger), and when they want to find a document, they say, “let me search for all emails from Rick and find the latest version he had sent me as an attachment.” The latter is essentially a scan/search operation against a ledger.

·??????The concept of log-structured file systems has existed for decades. Mendel Rosenblum, now a professor at Stanford, who was also a co-founder of VMWare with his wife Diane Greene, did his Ph.D. dissertation on The Design and Implementation of a Log-structured File System in 1992 (30 years ago) under the supervision of John Ousterhout, a legendary Computer Science professor then at Berkeley and now at Stanford.

·??????Many IT systems today are built around event-sourcing databases, which are essentially serial ledgers of activity, from which ledgers the current values of various data objects can be derived by replaying history.

·??????Most database systems rely critically on serial logs.?For example, to ensure consistency and durability in the presence of failures (e.g., system crashes), transaction updates need to be written to non-volatile memory. To expedite the transaction commit process, most database/transaction processing systems write the changes caused by a transaction in a (serial) log and save this log. A separate background process applies the changes to the individual database records. As long as the log is safe on non-volatile storage, no transactions are lost, even in the event of a system crash, because the log can be replayed after the system reboots.?This technical approach has been in use for over 40 years.

Based on the above, it is no surprise that a ledger structure was readily acceptable to the technical community, as evidenced by the fact that even permissioned chains use ledgers.?A ledger with a scanner or search engine is a perfectly acceptable data structure for a CSD.

ACID Transactions Against a Blockchain-Based CSD

Traditional transaction processing systems, including those used by banking, airlines, et al., have been developed around the concept of ACID transactions.?To refresh your memory, ACID stands for:

·??????Atomic: all parts of a transaction complete successfully or none does.

·??????Consistent: transactions observe the consistency constraints – they take the database from one consistent state to another consistent state.

·??????Isolated: the effect of a transaction must be the same as if the transaction was operating alone on the system (typically there are hundreds or thousands of transactions running in parallel).

·??????Durable: once a transaction commits, its effects are permanent, i.e., the effects cannot be undone for technical reasons (e.g., if the system crashes after completing the transaction but before the results are moved from memory to disk).

The ACID properties are not entirely independent of each other.?For example, a violation of atomicity or isolation often leads to a violation of consistency.

To ensure that I do not fall victim to potential biases of my own, I am constantly on the lookout for possible challenges to the ACID concept, but I have not seen anybody disagree with this concept.?In fact, it is the atomicity property that is typically violated by hacks related to blockchains, like the recent hack on the Binance Bridge that took place last Thursday night (October 6, 2022): the hacker fooled the bridge into minting BNB tokens without a corresponding deposit into the bridge.?For a detailed explanation of the hack, I recommend reading the analysis by @samczsun, research partner and head of security at crypto research firm Paradigm.

Thus, nobody appears to question the value of ACID transactions.?Furthermore, nobody appears to question the difficulty of achieving atomicity in a distributed environment, even when that environment is homogeneous and under the control of a single centralized authority (e.g., consider a bank that operates two transaction processing systems of its own, one serving the eastern part of the US and one serving the western part).?Ensuring atomicity becomes harder when one deals with a decentralized system that is not under the control of a single organizational authority (e.g., a public blockchain).?The challenge becomes even harder when one deals with multiple such decentralized systems (i.e., multiple public blockchains) that allow transactions to span such decentralized systems (i.e., cross-chain transactions).?The nuances in the concept of transaction finality across different blockchains further complicate the issue.?If you are interested in learning more about such nuances, I recommend the excellent blog entry Practical Guide to Finality by Scott Sunarto.

The technical literature is replete with examples of distributed commit protocols in which flaws were discovered years after the protocol had been put into operation.?And the hacks that occur in cross-chain bridges serve as a recurring reminder of the complexity of ensuring cross-chain atomicity.?In light of all this, transaction processing systems designers (and IT operators) have been steeped in the mentality that homogeneity and central authority are the best ways to protect the interests of their customers, which protection holds extremely high moral value.?Furthermore, they often believe that any violation of ACID properties would be totally unacceptable to customers. (It is important to note, we are discussing the technical finality of a trade.?The legal finality may very well be determined by the venue, regulatory environment, etc. that a trade is settled within.?For example, in some jurisdictions venues prevent bankruptcy court clawback due to legal and regulatory authority to clear or settle trades.?This is a core value proposition of exchanges like Euronext as an example.)

Typical Two-Phase Protocol

Source: https://www.alibabacloud.com/blog/tech-insights---two-phase-commit-protocol-for-distributed-transactions_597326

These sacred beliefs may be at the core of an altruism bias.?There may also be a familiarity bias, as many computer systems engineers prefer to stick to the (centralized) technical approaches they have used in the past.?According to Oakley, “it appears that altruism bias, like many such biases, is a Jamesian fringe phenomenon of consciousness, much like the feeling of familiarity.”?This in itself is a rejection of the fullness of the third ledger value in a triple ledger entry system.?Here is some evidence of bias I offer:

·??????Consistency. One way to preserve consistency is to ensure that inconsistencies never happen.?Another way to preserve consistency is to ensure that violations of consistency are detected, and consistency is restored quickly through remedial action.?Blockchain enables the detection of inconsistencies, which is how hacks are detected.?For cryptocurrencies, which are bearer instruments, restoration of consistency may not always be feasible.?However, securities are non-bearer instruments: if a hacker steals securities from a blockchain-based CSD, the securities can be returned to their legitimate owner by burning and reminting the corresponding tokens; this is technically possible using most tokenization platforms so that regulators and the judiciary can exercise their legal rights.?Furthermore, the regulatory framework around securities typically prevents anonymity, which will further discourage illegal activity.?Thus, token theft is not as much of an issue for digital securities tokens, since consistency can be restored.

If you think that the approach of allowing temporary deviations from consistency and restoring consistency subsequently is anathema, you can take comfort in the fact that there are applications where this has been happening for years.?If a cellular call ends abruptly (either because you went into a dead spot or because you dropped your phone into the bathtub), the billing database of the phone carrier becomes inconsistent: there is a call with a begin_call entry but without a matching end_call entry.?However, your bill at the end of the month does not show a call that lasted for days or months or years; the phone company detects the inconsistency and creates a synthetic end_call entry to restore consistency.

Thus, the moral value of avoiding even the slightest deviation from consistency may not be as high as originally thought.

·??????Cross-chain atomicity.?There are multiple ways to address the problems related to cross-chain atomicity.

o??First of all, bugs in current solutions (like bridges) will gradually be removed.?As the technical community learns from past mistakes, robust design principles will emerge.

o??Second, new cross-chain protocols are emerging, in which protocols avoid off-chain actions (bridges, off-chain side consensus mechanisms, etc.) and rely solely on actions that are recorded on the public blockchain.?Thus, these protocols are transparent and auditable.?An academic paper by Zakhary et al. at UCSB shows one such possible solution, even if not the most efficient one; it is noteworthy that the authors are recognized experts in the traditional world of transaction processing.?Also, protocols like @Accumulate can offer the opportunity to make the multi-signature consensus process for cross-chain transactions directly auditable on the blockchain itself.?Furthermore, protocols like Accumulate can meet the scalability requirements of applications like SWIFT.?On the Accumulate topic, I defer to @PaulSnow.?

Thus, the value of adhering to centralized solutions to avoid risks may not be as high as originally thought.

Conclusion

The concept of a digital CSD based on a public blockchain ledger with a scanner is viable and is what new forms of DTCC, Euronext , or NYSE will look like in 20 years in addition to the network of ATSs like Oasis Pro .?The technology to link the worlds of centralized finance and DeFi and effect a successful transition is available today.?The first areas of adoption will be interval funds, closed-end funds, and then private markets broadly.?The benefits to the industry will be the increased liquidity from aggregating liquidity silos across the globe and the 24/7/365 trading of assets.?Adoption of cross-chain transactions for securities will lead to (and may even accelerate) the adoption of cross-chain transactions for cryptocurrencies.

Inveniam is working with its partners to facilitate the successful transition from centralized finance to DeFi in terms of both (i) business model, market/industry structure, new services, new delivery models, etc., and (ii) technological approach.?From our perspective, the right solution for migrating out of a permissioned-only environment such as Hyperledger Fabric is to use Accumulate as a meta chain (Inveniam is a significant investor in Accumulate ), Hub as a multi-chain wallet solution to be deployed in conjunction with a custody solution (Inveniam owns the Hub in partnership with Oasis Pro ), Tokeny to create the security token, and then cross-list the token on Oasis Pro , ADDX , Nasdaq , Euronext , et al.?(Inveniam is an investor in both Tokeny and Oasis Pro , and I personally sit on the board of both of these companies).

Christos Polyzois and Richard Bunker are my collaborators in this new DLT area and are my teachers in the history and functioning of institutional-grade software for financial transactions, and institutions.?As we progress, we must call out technical, and cultural problems that drive certain folks to centralize, as we seek a networked and decentralized financial future and as a store of value evolves. Better data, price discovery, secondary markets, and global trading of assets will deliver liquidity for private market assets and funds.?

Apex Group Ltd 戴德梁行 Oasis Pro Accumulate @DeFiDevs Avalanche OpenZeppelin Tokeny Houlihan Lokey @Zand

Inveniam is an investor in the following companies mentioned in this article: Oasis Pro – (I sit on the board of Oasis Pro as well), Tokeny – (as is Euronext mentioned in the article) – Dynamic Mining , Rialto Markets , and DeFi Devs the creators of Accumulate .

Inveniam has commercial relationships with the following companies mentioned in this article: 戴德梁行 , Apex Group Ltd , ValuStrat , and 德勤

Be sure to?follow Inveniam on social media, listen to?our podcast, Sound DeFi, and subscribe to this newsletter,?The Missing Link.