Side 2: Block production centralization (EOS, Cardano, NEO, etc.)

Centralization is accepted in the Proof of Mandate (DPoS) scheme because it is believed that the economies of scale in mining will naturally lead to centralization in the PoW scheme. Given this fact, Dan Larimer invented DPoS. DPoS accepts the fact that blockchains will be naturally centralized and takes advantage of this to achieve scalability.

Larimer, for example, recently built EOS, which publicly claims to be limited to 21-bit block producers at a time. It is expected that only data centers will eventually be able to run EOS nodes.

By limiting the number of block producers, it can be inferred that each block producer has more resources. In addition, Byzantine fault tolerance (BFT) algorithms typically require n² communications to reach a consensus, and reducing the number of block producers reduces network latency. Reducing the decentralization of block production will clearly improve scalability.

DPoS based blockchains not only bring scalability, but also shorten TTF, which in turn reduces latency. EOS explicitly proposes to limit the block generation time to 0.5 seconds, which is impossible to achieve with other consensus schemes. This is one of the great advantages of DPoS. Many applications need to reduce latency and increase throughput.

Take decentralized exchanges. As the 0X ecosystem matures on Ethereum, this need becomes more apparent. One of the biggest problems with 0X today is the increasing number of order conflicts, directly due to the long block time and high latency of Ethereum. The 0X team has come up with a solution. It is unclear how effective these solutions will be, given the inherent limitations of the underlying Ethereum blockchain. However, this is not a problem, because DPoS has the advantages of short block generation time, high throughput and low latency.

In general, betting on DPoS is a bet on two things:

  1. Some applications on neutral databases require high throughput and low latency.

  2. Not all distributed systems need to be able to resist central authority.

In trilemma, the DPoS system focuses on the base of the triangle. Note that the first two triangles represent systems with longer TTFS, whereas DPoS systems have shorter TTFS and are therefore dark.

Side 3: Multi-chain set (Cosmos, AION, ICON, Ark)

The Cosmos, AION, ICON, and Ark teams believe there will be hundreds, thousands, and even millions of blockchains in the future. Instead of building huge blockchains like Ethereum and EOS, these teams will build systems based on triangular side 3 trade-offs. They believe that different applications do not necessarily share a set of validators. Instead, they should have their own set of validators.

Under the extensibility trilemma, each chain in a system built on edge 3 contains less value. For both PoW and PoS solutions, security increases with value. However, at the cost of increased risk comes sovereignty (which may not be necessary or worthwhile), extensibility and a shorter TTF for each chain.

What is the value of sovereignty? It’s easy to understand with some examples. If some countries use blockchain for administrative purposes, they will want to create their own set of validators that implement their own social values, not those of other countries or the global community. Ethereum and Ethereum Classic, for example, cannot share verifiers.

Clearly, Ethereum itself runs counter to the concept of sovereignty. Vitalik set out to create Ethereum in 2013 after discovering that all cryptographic app developers had to deal with the same mining and consensus algorithm challenges. He has an eye for opportunities and uses underlying platforms to solve these complex problems so that developers can focus on the application layer.

Looking at the history of cloud computing, there are some interesting parallels. In the early 21st century, many web hosts hosted multiple websites on a single server. As long as one site traffic is too large, the server will crash, involving other sites.

In many ways, Ethereum resembles the early web hosts. It ties too many things together and creates an overall unstable system. If one app breaks the system, the whole system crashes.

The solution to the network host problem is the Virtual Machine (VM). With each application isolated by a separate virtual machine, a single server can run multiple applications, maximizing hardware utilization while reducing costs while maintaining integrity. If traffic floods in, only one virtual machine will crash, not all the virtual machines on the server. VMS can be migrated between physical servers to further improve redundancy and security. Combined with systems that enable great horizontal scalability, virtual machines become a critical part of cloud computing. While Docker Containers have recently replaced virtual machines, it’s just old wine in a new bottle.

Similarly, if distributed applications must be built on separate blockchains, systems built on edge 3 will have a huge opportunity.

To achieve this vision, an edge-based system would need to greatly reduce the difficulty of creating new chains and interoperating between chains. Cosmos is making that vision a reality with Ethermint. Ethermint is 100 percent open source and free. It is an empty template chain based on the Tendermint that runs the Ethereum Virtual Machine (EVM). Tendermint is a semi-centralized consensus algorithm with high throughput and short TTF developed by the Cosmos team. By making it much easier for developers to create new chains, Cosmos wants developers to focus on building apps. AION, ICON and Ark all share the same vision and are working to provide template-based ways for developers to create new chains easily and efficiently. (Note that while Wanchain is generally considered a solution that enables interchain interoperability, it is quite different from the other solutions mentioned in this section and offers different capabilities.)

Note that these systems can achieve shorter TTF. This is because each chain is relatively centralized, enabling low latency performance. However, the overall system is highly decentralized, as many chains have separate sets of validators, each of which can use the new consensus mechanism.

Visual renderings of many small interoperable chains in the trilemma are as follows:

Sharding (Ethereum 2.0, Polkadot)

Vitalik and Gavin Wood have been talking openly about Ethereum sharding for years (see here and here for details).

Each shard is essentially an independent chain. While the chains in the Cosmos ecosystem are also independent, the difference is that each chain on the Cosmos must manage its own consensus (to ensure security), whereas ethereum shards do not. In the case of Ethereum sharding, the consensus of all shards and the security they bring together is managed by the Validator Manager Contract in the master shard. In most cases, it is better to centralize security than to spread it across multiple chains.

No one has yet put a shard blockchain into production (Zilliqa claims to be shard, but not fully shard).

Sharding solves the extensibility trilemma, albeit at the expense of chain sovereignty. In addition, cross-chip communication is delayed. Basically, latency is a function of the consensus algorithm on each shard.

Ethereum intends to use Casper to finalize each shard. Although no specific parameters have been set, Casper is likely to generate TTFS measured in minutes, making sharded Ethereum a high-latency system.

Dfinity and Algorand use a new consensus algorithm called Threshold Relay, which aims to achieve decentralized block production, security, better scalability than individual Ethereum shards (though far less than what DPoS offers), and shorter TTF. If threshold relay works as well in practice as it did in the test environment, we can expect to see Dfinity and Alogrand sharding their chains to provide efficient cross-chip communication with shorter TTF and low latency.

Polkadot will use the TBA BFT consensus algorithm to create a sharding network that reduces TTF at the expense of DBP (similar to the tradeoff followed by Tendermint). However, Polkadot is likely to launch sharding networks much earlier than Ethereum or Dfinity. Unlike Ethereum and Dfinity, Polkadot does not require sharding to use a specific state machine, such as the Ethereum Virtual Machine (EVM) or WebAssembly (WASM). Instead, Polkadot allows each chain to define its own state machine. That said, you can move your consensus system to Polkadot through protocols such as Stellar’s ability to distribute arbitrary digital assets without smart contract capabilities, or Zcash’s ability to handle SNARK proof with highly optimized state machines.

In conclusion:

agreement Consensus algorithm Virtual machines (per shard Open Sharding plan
Ethereum Casper The EVM entry – eWASM Defined, in progress
Dfinity Threshold relay WASM undefined
Polkadot TBA undefined The defined

Full state sharding is an unsolved computer science problem. No one has been able to put it into mass production. However, if possible, this may be the best solution to the scalability trilemma while providing a shorter TTF and low latency.

In the trilemma, the visual renderings of sharded PoS are as follows:

Verifiable down – chain calculations

Could a user achieve this by simply asking a computer to perform a single calculation without compromising efficiency? Other than having dozens or hundreds of computers perform the same calculation, is there any other way to prove that one computer did the correct calculation? Can we guarantee accuracy without fundamentally inefficient blockchain technology?

Truebit is trying to solve this problem with a prover-verifier Game. The game will be implemented using proverse-response interactive protocols. Each computation is performed locally, with only one solver and several verifier, with no network cost, without the expectation that the results will be challenged. In the rare case of a challenge, solvers and validators need to re-run a wASM-based computation-intensive virtual machine to determine who is the perpetrator.

The interactive authentication protocol combines transparency, security, and the immutability of the underlying chain with the efficiency of off-chain computing. Truebit runs in the middle of an interaction and is probabilistic, so it does not work well in environments where low TTF is required. Truebit runs on the first “unanimous” consensus system, requiring at least one reasonable verifier per task to ensure security.

In the future, Truebit, or its competitors, may use SNARKs and STARKs to verify the accuracy of arbitrary calculations with non-interactive proofs rather than interactive proofs. If so, Truebit will shorten TTF and increase the design space for trust-free chain computing. However, when used with SNARKs and STARKs for general purpose computing, the zero-knowledge proof of arbitrary computation remains largely uncertain, unproven, and technically inefficient, and may ultimately fail to fulfill its potential.

Open source, replication and politics

In the long run, mainstream chains may absorb the best technology from niche chains. Ethereum, for example, is deploying zkSNARKs, the primary technology that allows Zcash to make private transactions. Furthermore, Vitalik has announced that he wants to port Ethereum to WASM, a technology already adopted by EOS and Dfinity. Ethereum theologians propose Plasma as a solution to all of Ethereum’s flaws because it allows alternative consensus algorithms and state machines to be selected in the Ethereum ecosystem.

Given the above issues and the network effect fallacy of smart contracts, how can chains be distinguished from one another?

Politics says, ideology and beliefs.

Assuming that no one can solve the scalability trilemma without compromise, different people and businesses need special chains suited to the task at hand.

Finally, the media will turn this into a religious debate. Blockchains will become religions because their followers are preachers who are willing to spread the Gospel and attract more people.

Will users choose the most decentralized and censorship-resistant chains? Even at the expense of performance and high network costs? Or will they accept a lower bar for censorship? If so, where is the threshold?

Digital gold

Based on the framework proposed in this paper, we can objectively evaluate how and why digital gold and programmable currency may become independent in the foreseeable future.

The Bitcoin core team prioritized DBP maximization and transaction verification at all costs, even at the expense of scalability and other forms of usability. Reducing bitcoin’s efficiency could increase its resilience. The core Bitcoin team doesn’t seem constrained by the need to sacrifice technical efficiency for social scalability.

You could argue that even the core bitcoin team of believers is misguided. For example, PoS advocates argue that PoS is more secure than PoW because once a malicious attacker is identified, she can be immediately kicked out. In contrast, in a PoW scenario, the perpetrator can always continue to attack the network.

In the long run, we do not know whether PoS schemes are more resistant to government interference than PoW schemes. The two schemes have their own merits. However, we know that PoW is really proven. Given that the priority of the Bitcoin core team is to achieve the highest level of censorship protection, in our experience, the bitcoin core developers made the right decision.

Some argue that the core bitcoin team’s views are too extreme, and that by sacrificing practicality for superfluity, bitcoin will lose meaning.

The question now is, how far is DBP enough? The design space for trust-free computing is much wider if you consider the design threshold to be lower than “able to withstand a combined frontal attack from ××, ×× and ××”.

In the 1990s, many thought the Internet would eventually become a democratizing force, connecting people in unprecedented ways and dismantling traditional media companies and authoritarian governments. As it turns out, it is mainly large corporations and governments that use the Internet to gather power and control.

Because all trust-free computing systems are permissionless, that means governments can make money from them. In addition to a frontal attack, governments may use the crypto industry in unforeseen new ways to increase rather than loosen their grip on society.

A hodgepodge of trust-free computing models?

In the foreseeable future, all of these trust-free computing scale-up models will be flourishing after a Cambrian explosion in the crypto ecosystem.

The road to a stable future is bound to be bumpy. Things inevitably go through a hodgepodge of complex interwoven stages before achieving homogeneity.

For example, a developer might create a new chain on Polkadot, only to discover that a relay chain on Polkadot has reached its limit. This chain may be connected to other relay chains, which in turn are connected via Cosmos Hub to chains on EOS, Ethereum and Kadena. Each chain can be sliced. Various Ethereum shards may contain Plasma chains that use DPoS and PoA consensus to ensure security.

In the future, contracts in these major systems will be able to call Truebit for off-chain calculations, making this uncertainty even greater. How much work can Truebit do? How will Truebit integrate the main chain and the profile?

It is not clear where or how the value of these systems will accrue. The argument goes that if Cosmos’s ATOM token did not follow Menger’s marginal utility theory, an interoperable chain like Cosmos would not accumulate so much value (which could be the difference between billions and trillions). Polkadot’s DOT token faces the same risk.

I hope to see substantial, if not complete, integration in the future. Given the fundamental value of sovereignty for certain chains (those controlled by governments, for example), I think a system like Cosmos will always have a place, if not the biggest winner.

It is unknown who will win the deer, the world’s heroes can be chased.

Thanks to Trent McConaghy (Ocean), Peter Czaban (Web3/Polkadot), Jesse Walden (A16Z), Will Martino (Kadena), Matt Luongo (Keep), James Prestwich (Integral), Robbie Bent (Truebit), and Zaki Manian (Cosmos) contributed to this article.

Note: After writing the first draft of this article, I discovered that Trent McConaghy had proposed a similar framework 20 months earlier. Thanks to Trent for his contribution to this article.

Original link: medium.com/multicoin-c…

By Kyle Samani

Translation & Proofreading: Min Min & A Jian

Source: Ethereum enthusiast (ethfans.org/posts/model…)