Private vs. Permissioned
There are two general flavors of blockchain: public and permissioned. A public blockchain is available to anyone with a computer and internet connection. Anyone can download a copy of the Ethereum software and start processing incoming transactions on the network from their computer. Participant anonymity incurs the computational burden of POW, where network participants must effectively "pay" electricity to solve a math problem with a really large number.
A permissioned blockchain is a closed network available only to invited participants. Consensus is achieved deterministically, because we assume a legal consequence if anyone decides to cheat and submit double spent transactions onto the chain. Since only invited participants are eligible for inclusion into the network, network participants know the real-world identities of the other users on the network. As a result, mining, or validating new transactions and packaging them into blocks, is achieved deterministically. In contrast to a pseudonymous network like Bitcoin, permissioned networks achieve consensus by each participant taking turns proposing and voting on new blocks, and new blocks are finalized when unanimity is reached. In contrast, POW systems are only probabilistically final since there is always a chance that 51% of miners could agree on a different transaction state and decide to fork the chain.
So, if permissioned chains don't solve double spends, why use them at all? Blockchains are nice templates for building distributed networks. People often claim that state-bearing operations can be executed with the same assurance as a centralized database. While true, that claim ignores a tertiary benefit of blockchain: an out-of-the-box distributed network SDK. When faced with a non-double-spend-assured blockchain, it's helpful to think about the deployment process for a centralized database. It's just much easier to build a distributed network using a blockchain than it is to build each component separately. In orthodox IT, you'd probably start a network by spinning up the following components at a minimum: devops, cloudflare, VPNs, containerization, cloud hosting configuration, load balancers, and DNS configurations. This introduces new blockchain-specific issues into the fold: contract security, gas fee metering, block limits, replay attacks, etc.
What is Fabric?
I'll start with the main criticism first: it's not a blockchain. In POW blockchains, consensus is achieved by nodes competing to submit a valid block which fulfills the nonce requirement. If a node proposes a block which has a double spend, the block will be rejected the network. In comparison, Fabric only orders the transactions by timestamp and submits the block to all nodes that are then left to reconcile two transactions.
But it's not a real blockchain you say! No matter: Fabric gives users a timestamped record of transactions distributed across an arbitrarily large network of nodes running the Fabric software. This is a great use case for companies who seek to synchronize their data across an industry vertical, whether its supply chain or enterprise ETL. Decentralization isn’t necessary because each participant has a real-world incentive to behave honestly. In a business network, it doesn't make sense for a company to lie to its trading partners. Remember, the reason we need POW in public networks is because participants are anonymous and only have a financial incentive to behave honestly. In permissioned networks, participants identities are known and thus legally and reputationally discouraged from submitting double spends to the network. Trading partners on Fabric networks have a financial interest in maintaining allegiances with their partners and this assumption is built into the architecture of the network. Remember, different problems, different solutions. Fabric is solving a particular business problem.
Tokenization can refer to the creation of a digital representation of a physical asset for many markets: automobile titles, airline tickets, stocks, bonds, and mortgages. But the type of market being tokenized results in a bifurcation in the types of blockchains available. On the one hand there are (1) tokens designed for the public cryptocurrency markets. These tokens exist on the public Ethereum network and are highly liquid because of the retail demand in early-stage startup investing. On the other hand are (2) highly customized tokens designed for industry-specific business processes, that are generally not available in public markets.
Two Threads of Enterprise Blockchain
In tokenizing an asset, the benefit of a public chain is its accessibility: when an asset is traded on a public network, anyone who runs a blockchain node can access the "database" of assets. If the network is popular enough, exchanges will emerge to fulfill the needs of people who wish to have market makers fulfill their trading requests.
Ethereum is the most sensible option for creating a tokenized asset with the scale and accessibility of a public network. This has the downside of opening the Enterprise to regulatory or privacy issues: since the network is public, all nodes can view the state of account balances and contract states.
It's worth comparing this approach to the other variety of enterprise blockchains. For many enterprises investigating the blockchain space, their aim is to mitigate the costly process of reconciling data across different parties in a value chain. What this means is that companies who cooperate together in a vertical want to ensure that everyone has access to the same data at all times. If this seems trivial, consider the amount that is spent on ETL process products each year - the amount is staggering.
When a company hires a dev shop to build a Hyperledger Fabric blockchain, they're looking for someone to rebuild an entire slice of their supply chain. In my previous role, one of our major successes was migrating an automotive title management software company onto a blockchain network. Now that's a mouthful, but to accomplish this, we had to first station a solution architect and project manager on-site for two weeks while they learned about the automotive-title business. Then, we had to upload tens of millions of existing automotive titles onto the blockchain, which was the most difficult technical challenge we faced that quarter. And, we had to complete on-site training sessions for each counterparty involved in the automotive title lifecycle. Building this new business process was nothing short of a 10,000 foot dive into the deep sea of automotive title transfer, which is good if that's an industry you want to get into. But for those of us who want to build modular infrastructure for other people to build applications on top of its an astonishingly high upfront cost, and most of the billable hours are from the "solution architecture" (i.e consulting) time it took to understand the business process well enough to design a blockchain network around.
This is a generalized business process use-case that is best served by industry experts or Park Avenue consulting firms (industry experts).
Building an open architecture for the "internet of finance" is a much stronger and defensible position to be in, for two reasons.
In the blockchain space, there's a constant debate between protocols, namely the utility of digital money versus self-sovereign applications. On the digital money side, Bitcoin maximalists rejoice at the prospect of an asset that's agnostic to the inflationary-whims of a central bank, while relishing the freedom of being able to transport a billion dollars of value with nothing else but a 32-character private key. Meanwhile, Ethereum fans participate in hackathons, build business networks, and co-opt plans to return liberty to the users of the internet whose power has been usurped by our monopoly overlords: Facebook, Amazon, Apple, and Google.
The second killer app of cryptocurrency (the first is a permissionless and deflationary currency called Bitcoin) was the equity crowdfunding app, also known as an ICO. The ICO was the second killer app of crypto, and the first killer app of Ethereum. When Ethereum was founded in 2014, the organization raised $20mm by pre-selling units of digital currency, called Ether, which could be used to pay transaction costs on a new distributed network of computers, called Ethereum. In 2017, the amount of money raised in ICOs topped the amount of money raised in Seed, A, and B venture rounds combined.
Target and Non-Target Markets
The interesting thing about tokens is that instead of relying on the creation of a marketplace to aggregate demand for the commodity itself (Stubhub for tickets, Uber for cabs, Airbnb for apartments), the tokenization model aggregates demand around the speculative market for that commodity.
By creating markets for services in which the speculative asset is used for the underlying product, tokens bridge the gap between investors and users, which compounds the combined value and utility of the product.
So far, this model has proved effective at drawing increased demand for subpar products. This is why ICOs have become notorious for scams. Many projects drew early interest from speculators but lacked the functionality to sustain meaningful adoption of the service itself.
What remains to be seen is whether a genuinely useful digital service with an underlying token will actually cause it to grow faster. Would Airbnb have experienced faster growth when it was launched in 2008 if users had to rent apartments using tokens?
Additionally, speculative markets which affect the price of a commodity-based market can input erroneous entrepreneurial signals into the market. This is the Austrian argument for shunning Central Bank interference into an economy. An invisible hand which tries to predict the future may inevitably skew the genuine behaviors of economic actors and snarl the non-verbal coordination that naturally occurs in well-functioning markets.
In spite of any utility, tokens do offer a 250x increase in retail owners, 30x greater pool of US buyers, and 1000x shorter time to liquidity with no dilution in equity. This is a huge improvement on having capital monopolized by venture capitalists, and it poses an existential threat to venture investors who lack any specialized advisory knowledge that would warrant diluting equity for their capital.
The main challenge with building applications atop private networks is aligning the network incentives among network participants. If a user was to tokenize an asset on a private network, we'd need to sit down with our customer and decide on a counterparty who will honor the value of that asset at the point of sale. Public networks are existing, high-volume marketplaces, which customers can plug into using NPM libraries. Private networks are specialized marketplaces for a small number of participants.
Bitcoin was released in 2008 -- Ethereum in 2014. But in the span of a decade, cryptocurrencies have posed a meaningful threat to the status quo in central banking and venture investing. The promise of blockchain technology is broad, and it can be difficult to untangle the different threads manifest in the ideology of its supporters. It's possible to imagine a future where our payroll is processed on a Hyperledger-run blockchain, and it's also possible to imagine a future where investors in Korea own fractional shares in our Brooklyn condominiums. Many unknowns remain.
But if there is one thing certain, it’s the changing landscape of our financial ecosystem.