An article to understand the new Layer 3 concept proposed by StarkNet

The main purpose of Layer 2 is to reduce fees and expand capacity. The limited functions of the Ethereum public chain make it unable to carry the increasing amount of execution and calculation brought about by the rapid development of the ecosystem. To complete the transaction on the chain, it has to pay high transaction costs. The essence of public chain congestion is that too many nodes on the chain compete for too little block space. Decentralized redundant design is inherently opposed to efficiency. The transaction on the chain needs to go through a network-wide broadcast from reaching to final confirmation. , the nodes need to confirm the authenticity of the broadcast information through consensus. The more nodes there are, the more troublesome this process will be.

People are gradually realizing that the inherent inefficiency has become the biggest obstacle to the development and promotion of decentralized technology. In order to solve this dilemma, people began to strip and layer the functions of the public chain, leaving the most core consensus mechanism that provides security in a decentralized system on Layer 1, while delegating calculation and execution to Layer 2. Promoting the public chain to become the settlement layer of Layer 2. Layer 2 solutions have gradually been recognized by the market because they are easier to implement and operate than cross-chain and sharding. With the continuous application of the respective Layer 2 of the protocol, people can clearly feel the significant reduction in transaction costs. From the results, Layer 2 can be said to be a very successful industry change.

Layer 2 has developed from the early state channel, side chain, Plasma and other solutions to the current Optimistic rollups, Validium, ZKRollup, etc. There are different degrees of trade-offs in the performance of each solution. Generally speaking, the market currently recognizes Rollup. But the market still knows that Layer2 is not a near-perfect solution, and its flaws are also very obvious. Typical problems include its limited scalability, backward composability, and fragmented liquidity. However, the technical support for the interoperability and liquidity of Layer2 is constantly being explored and increased, and the proportion of users' activities on L2 is also increasing.

Layer 2 retains the advantages of the decentralized model as much as possible on the premise of reducing costs and improving efficiency. But for some future application scenarios that may require a high degree of customization, it is best to provide services by a new independent layer, which is the origin of the Layer3 concept. This concept was first proposed by StarkNet. They said that the relationship between Layer3 and Layer2 is the same as Layer2 and Layer1. This is actually a nested and recursive structure. As long as Layer2 can support the Verifier smart contract, Layer3 can use the validity proof to accomplish. The whole process becomes that Layer3 compresses the transaction to Layer2, and Layer2 compresses the compressed transaction to Layer1, which is a geometric expansion concept. If the cost of each layer is reduced by 1000 times, then the cost of Layer 3 can be reduced by 1,000,000 times compared to Layer 1.

The main advantage of Layer3 proposed by StarkNet, in addition to its hyper-scalability with the multiplicative effect of recursive proofs, is that it can be better controlled by the application designer of the technology stack, with more deterministic performance and cost , custom data availability models (e.g. based on Validium or application-specific on-chain data compression), faster functionality and technology turnaround (e.g. the ability to continuously introduce new features for testing that are not yet ready for general availability).