Starting from consensus mechanisms, exploring the changes in blockchain application paradigms

This article was published on ChainNews, author: outprog, founder of everFinance.

In 2009, Bitcoin was born, marking the beginning of the blockchain era. In a distributed network, all nodes compute a problem, and the first to solve it receives a Bitcoin reward. Through the infinite calculations of numerous problems, an immutable data chain is created, and the transfer information recorded on the chain, marked by objective signs, generates immense value driven by people's consensus. This immutable ledger later became known as blockchain, which revolutionized the ledgers that had been used for thousands of years.

The following text will explore the changes in consensus in the blockchain field in recent years, the evolution of blockchain applications, and why we can build a new trustworthy application through the storage computing paradigm.

Absolute Decentralization

Bitcoin purists firmly believe in absolute decentralization. The Bitcoin network is like Skynet from the famous sci-fi movie "Terminator," a network operating globally, with ledgers on countless servers. Like Skynet, shutting it down would likely require destroying the entire network. This P2P node network possesses decentralized characteristics and is not controlled by any country or institution.

Absolute decentralization requires users to verify all transactions themselves. In the early days, installing a Bitcoin wallet on a PC could take hours or even days to synchronize the entire network, and transfers could only be made after synchronization was complete. When our personal computers went offline or were shut down, causing the wallet to be offline for an extended period, we had to continue syncing data before making any new transactions, waiting until all data was fully updated. Such an experience is unacceptable for the vast majority of users.

People always need to sync and verify all transactions to gain absolute decentralized trust.

Trusting Wallet Service Providers

With the ICO boom in 2017, Ethereum brought blockchain to a new height. New users flooded in, and the official Ethereum Mist wallet (which, like Bitcoin, required days to download and verify all transactions) struggled to meet user demand. imToken emerged, creating a more user-friendly API wallet.

Users no longer needed to perform long data synchronization; downloading the app was enough to get started. Blockchain nodes were provided by wallet service providers, while users' private keys remained stored on their personal phones. Users accessed the service provider's nodes via API, signed transactions on their phones, and sent the transactions to the service provider's chain nodes.

Another example of an API wallet is MetaMask, which played a crucial role during the rise of DeFi from 2018 to 2020. The browser plugin wallet was also plug-and-play, allowing users to skip ledger verification and directly use DeFi with the new Ethereum infrastructure, Infura. From exchanging on Uniswap, lending on Compound, to the liquidity mining craze that began in mid-2020, MetaMask became an essential tool for DeFi users.

However, in the exhilarating experience of DeFi, some fundamental views on consensus may have been lost. During the use of API wallets, users did not verify each blockchain transaction but trusted the nodes deployed by the service providers. Absolute decentralization was abandoned, and user experience took precedence. The official Ethereum wallet, Mist, even announced it would stop updates before the DeFi explosion.

Although absolute decentralization (users independently verifying the ledger) gradually faded from the historical stage, decentralized consensus did not weaken. With the lowering of user thresholds, the prosperity of DeFi emerged, and the entire blockchain narrative became more deeply ingrained. Essentially, whether we use imToken or MetaMask, even if internal nodes are behind the scenes and users do not verify the ledger, once user transactions are packed onto the chain, they can be guaranteed through thousands of other blockchain nodes. This objective fact cannot be altered or forged by any single service provider or centralized cloud service. A centralized service provider can deceive users temporarily, but not indefinitely. The objective fact of user transactions is already packed in the vast network of blockchain nodes, making it immutable and unforgeable.

Objective Facts and Data Presentation

Compound is a financial lending application running on Ethereum, where each transaction is computed by thousands of Ethereum nodes to ensure the objective facts of Compound's lending business. However, neither users nor service providers can directly obtain the lending business's yield rates from the chain. The facts of lending are continuously occurring, but the annualized yield generated requires additional statistics on the on-chain data to be derived. The facts exist objectively, while users' decision-making relies on the representational data derived from those facts. At this point, objective facts and data presentation are separated from trust.

Returning to the perspective of Bitcoin purism, only when users use full nodes or SPV wallets to verify each transaction can the data presentation they obtain possess complete decentralized trust. When using API wallets like imToken or MetaMask, users request balance information from Infura or other service providers, which is essentially just a representation of data, not fundamentally different from Compound's yield rates.

Compound's yield rates are calculated on centralized servers using on-chain data (usually calculated using Ethereum event logs). At the same time, the balance requested by users from Infura or other service providers is also a state generated by the calculations of centralized servers' Ethereum programs. It is particularly noteworthy that Ethereum nodes on centralized servers are also ordinary applications that can be modified and adjusted. Therefore, whether the data obtained from Ethereum node APIs or the new data processed from node data cannot provide completely decentralized trust. Service providers' servers can entirely modify Ethereum nodes and any custom statistical programs. It is evident that when users do not independently verify all transactions, objective facts and data presentation inevitably create a binary separation. The data obtained by users through service providers is merely a secondary statement of objective facts.

To obtain absolutely objective facts, users must independently verify all transactions and calculate data themselves. This model of absolute decentralized trust is difficult to apply in real-world scenarios due to the high user threshold, becoming an idealistic notion.

The Evolution of Blockchain Application Paradigms

With the failure of the official Ethereum wallet Mist, the entire consensus has subtly changed: users no longer verify transactions themselves but instead use Infura or application service providers' chain nodes for transactions. At this point, trust has been completely handed over to service providers.

Users' understanding of consensus is gradually changing. One of the roots of this change is the shift in the blockchain user base, as purists continuously welcome new blockchain users. These new users have never independently deployed nodes or SPV wallets; their needs are not for decentralized trust but for simple and easy-to-use token transfers and high-yield DeFi applications. The vision of decentralization has shifted from requiring independent access to absolutely objective facts to relying on service providers to provide secondary statements of facts. The evolution process is illustrated in the following diagram, transitioning from a single-layer topological model to a dual-layer model, which may evolve into a three-layer model in the near future.

Figure 1 shows the absolute decentralization model, where the user's client and nodes are at the same level, and users sync nodes and verify all transactions. In this model, user experience is extremely poor.

After the ICO boom in 2017, imToken and MetaMask rose to meet the rapidly expanding user demand, leading us into the dual-layer model era depicted in Figure 2. Users no longer verify transactions; nodes are entirely maintained and deployed by service providers. Users simply send signed data to service providers, who process the transactions through their nodes. At this point, user experience has improved, as users no longer endure painful transaction synchronization and verification.

Figure 3's three-layer model completely decouples users from the blockchain; in this model, users no longer interact with the chain at all, including not sending signed transactions to chain nodes. The trustworthiness of service providers is provided by various blockchains, and users' transactions are entirely processed by application parties. The three-layer model offers the greatest benefit of providing users with an internet-level experience, including internet-level TPS and trustworthy internet accounts.

The entire evolution process is an enhancement of user experience, moving from difficult to usable, and finally to easy to use.

Evolving to a Three-Layer Model

In fact, our current DeFi applications are not purely a dual-layer model but rather a 2.5-layer model.

During the use of DeFi, most data is not natively output from nodes but has been statistically presented through centralized servers. The blockchain merely provides a settlement function, while the nodes deployed by service providers can also modify programs and forge data.

We can hypothesize two types of attacks:

1. Misleading data presentation + real nodes. The service provider offers real nodes, but the displayed interest rates, exchange rates, or other information have been modified, enticing users to perform on-chain operations. Transactions still occur on-chain, but the factual results and data presentation differ; for example, the displayed annualized yield is 10%, but the actual yield is only 9%; the exchange rate shown is 256, but the actual transaction is 250.
2. Completely false nodes, directly leading to users' transfers and DeFi results being fraudulent.

To avoid these two types of attacks, users must run nodes and perform data statistics themselves. However, this is very difficult for users, who can only trust this data presentation or verify data through other third parties. A real-world example is that when using DeFi through imToken or MetaMask, we always need to check transactions on Etherscan.

When service providers develop applications, absolute decentralized trust will be conveyed to the application. Users trust the service providers and can also verify transactions through other third-party service providers for higher trustworthiness. Regardless, users' trust has already been severed from the decentralized trust of the blockchain. In summary, we can now categorize trust into two types:

1. Absolutely trustworthy: distrustful of anyone, independently verifying all blockchain transactions to obtain a completely decentralized ledger.
2. Relatively trustworthy: trusting service providers, obtaining data presentations of the ledger through service providers, and using data from multiple other service providers for transaction comparisons.

The left image is an extension of the dual-layer model, where users interact with service providers' nodes while using the data presented by service providers. For users, this is a relatively trustworthy environment; the nodes provided by service providers may not necessarily be real nodes, and users need to verify through third parties or independently.

In the dual-layer model, users still directly interact with chain nodes, initiating transactions to the nodes. The right image depicts the three-layer model, where users, service providers, and chain nodes are three separate layers, and users no longer interact with chain nodes at all. Service providers offer trustworthy computing based on storage, processing users' transactions through off-chain programs. Although transactions are processed by off-chain programs, all processing steps are guaranteed through the blockchain (based on the storage computing paradigm of Arweave), so off-chain computations can also be verified by third parties or independently.

Storage computing paradigm: Open-sourcing deterministic programs or uploading them to the chain, with programs running entirely off-chain, and all inputs sent to a storage-based blockchain for preservation. Once data is packed, it possesses order (temporal sequence). Any third party can run the program, which will load factual data from the chain to generate factual states.

Whether in dual-layer, 2.5-layer, or three-layer models, users can only obtain relative trustworthiness; absolute trustworthiness is solely related to service providers. As long as users do not independently verify all transactions, complete trustlessness cannot be achieved.

In the three-layer model, we can construct a brand-new Compound that does not require using the EVM on-chain computing model; it only needs to ensure program trustworthiness and data trustworthiness, so the objective facts of application operation absolutely exist. In the architecture of the three-layer model, not only is the annualized yield calculated off-chain, but the balance status generated by each transaction is also computed and presented off-chain. Improving the dual-layer model to a three-layer model does not change the objective facts, and users' relatively trustworthy environment remains unchanged. The objective facts of transactions on-chain are supported by blockchain technology.

Advantages of the Three-Layer Model

Why establish a three-layer computing model that completely isolates users from the blockchain?

The chain has too many limitations, making it difficult to meet the low-threshold needs of ordinary users. The TPS of blockchain has many restrictions, expensive miner fees, and the management of private keys and mnemonic phrases make it hard for ordinary users to accept. After achieving the goal of absolute decentralization, blockchain has sacrificed a lot of usability.

Applications built using the three-layer model will empower applications with trustworthy blockchain nodes, allowing application development to gain tremendous flexibility and provide users with extreme usability.

• Composability: Using the storage computing paradigm does not sacrifice composability; any application can directly interact via API. Transactions generated by applications, once packed onto the blockchain, become objective facts that are immutable and verifiable. If two mutually trusting applications encounter false transactions during real-time API interactions, all subsequent transactions generated from those false transactions can be detected and rolled back, which is a form of off-chain rollback, or correcting erroneous accounts to keep off-chain programs consistent with blockchain facts.

• Low development threshold: By simply open-sourcing traditional applications and packaging transactions onto the blockchain, applications can achieve transparency and trustworthiness, which can be verified by third parties. At this point, there are no language restrictions on development; only the paradigm of thought is needed.

• No performance limitations: The performance of off-chain programs depends solely on the machines deploying the programs and the capabilities of the development team.

• Extremely low consensus costs: Reducing the cost of consensus to the cost of blockchain storage, using Arweave for transaction storage, where $1 can facilitate nearly a million transactions.

Disadvantages

The main loss is the entry barrier: applications are implemented and deployed by a centralized service provider, who can filter and review users' transactions. Additionally, a single point of service failure may lead to service interruptions.

However, no one can damage the objective facts of the ledger; factual programs combined with each transaction signed by users will be recorded on the blockchain, becoming verifiable factual states.

Conclusion

The minimum cost of consensus is the storage cost. Through the blockchain's immutable and traceable characteristics, we can preserve factual programs and factual data, and their operational results must be real, objective, and immutable. As for computation, it is not within the realm of facts, because whether computation has occurred or not, the objective facts do not change. Verifiability is the key to the entire blockchain consensus and trustworthiness. Anyone can verify those immutable objective facts, which are the foundation of all trust.

The verifiable trust provided by the storage computing paradigm establishes a bridge of trust between multiple parties, and the verifiable transparent ledger fundamentally revolutionizes traditional ledger technology, resolving various contradictions and disputes arising from traditional ledgers (contracts) in the coordination of production relations.

Ledgers are a key element of human collaboration, and the combination of each person's and each institution's ledgers facilitates the cooperation and division of labor in today's society. As a revolutionary factual machine, blockchain will fundamentally change the trust system built on ledgers, making trust easier to achieve and allowing value to flow without boundaries.